Hours of Service of Railroad Employees; Substantive Regulations for Train Employees Providing Commuter and Intercity Rail Passenger Transportation; Conforming Amendments to Recordkeeping Requirements

Federal RegisterAug 12, 2011

Ask Donna

What actually matters in this document.

Text

DEPARTMENT OF TRANSPORTATION

Federal Railroad Administration

49 CFR Part 228

[Docket No. FRA-2009-0043, Notice No. 2]

RIN 2130-AC15

Hours of Service of Railroad Employees; Substantive Regulations for Train Employees Providing Commuter and Intercity Rail Passenger Transportation; Conforming Amendments to Recordkeeping Requirements

AGENCY:

Federal Railroad Administration (FRA), Department of Transportation (DOT).

ACTION:

Final rule.

SUMMARY:

FRA is amending its hours of service recordkeeping regulations, to add substantive hours of service regulations, including maximum on-duty periods, minimum off-duty periods, and other limitations, for train employees (

e.g.,

locomotive engineers and conductors) providing commuter and intercity rail passenger transportation. The new substantive regulations require that railroads employing such train employees analyze and mitigate the risks for fatigue in the schedules worked by these train employees, and that the railroads submit to FRA for its approval the relevant schedules and fatigue mitigation plans. This final rule also makes corresponding changes to FRA's hours of service recordkeeping regulation, to require railroads to keep hours of service records and report excess service to FRA in a manner consistent with the new substantive requirements. This regulation is authorized by the Rail Safety Improvement Act of 2008.

DATES:

Effective Date:

This final rule is effective October 15, 2011. Petitions for reconsideration must be received on or before October 5, 2011.

ADDRESSES:

Petitions for reconsideration:

Any petitions for reconsideration related to Docket No. FRA-2009-0043, Notice No. 2, may be submitted by any of the following methods:

•

Web site:

The Federal eRulemaking Portal,

http://www.regulations.gov.

Follow the Web site's online instructions for submitting comments.

•

Fax:

202-493-2251.

•

Mail:

Docket Management Facility, U.S. Department of Transportation, 1200 New Jersey Avenue, SE., W12-140, Washington, DC 20590.

•

Hand Delivery:

Room W12-140 on the Ground level of the West Building, 1200 New Jersey Avenue, SE., Washington, DC, between 9 a.m. and 5 p.m. Monday through Friday, except Federal holidays.

Instructions:

All submissions must include the agency name and docket number or Regulatory Identification Number (RIN) for this rulemaking. Note that all petitions received will be posted without change to

http://www.regulations.gov

including any personal information. Please see the Privacy Act heading in the

SUPPLEMENTARY INFORMATION

section of this document for Privacy Act information related to any submitted petitions, comments, or materials.

Docket:

For access to the docket to read background documents or comments received, go to

http://www.regulations.gov

or to Room W12-140 on the Ground level of the West Building, 1200 New Jersey Avenue, SE., Washington, DC, between 9 a.m. and 5 p.m. Monday through Friday, except Federal holidays.

FOR FURTHER INFORMATION CONTACT:

Mark H. McKeon, Special Assistant to the Associate Administrator for Railroad Safety/Chief Safety Officer, FRA, 1200 New Jersey Avenue, SE., RRS-1, Mail Stop 25, Washington, DC 20590 (telephone: 202-493-6350); Dr. Thomas G. Raslear, Staff Director, Human Factors Research Program, Office of Research and Development, FRA, 1200 New Jersey Avenue, SE., RPD-321, Mail Stop 20, Washington, DC 20590 (telephone 202-493-6356); Colleen A. Brennan, Trial Attorney, Office of Chief Counsel, FRA, 1200 New Jersey Avenue, SE., RCC-12, Mail Stop 10, Washington, DC 20590 (telephone 202-493-6028 or 202-493-6052); or Matthew T. Prince, Trial Attorney, Office of Chief Counsel, FRA, 1200 New Jersey Avenue, SE., RCC-12, Mail Stop 10, Washington, DC 20590 (telephone 202-493-6146 or 202-493-6052).

SUPPLEMENTARY INFORMATION:

Table of Contents for Supplementary Information

I. Executive Summary

II. Statutory Background and History

III. Scientific Background

A. Validated and Calibrated Fatigue Models

1. Fatigue Avoidance Scheduling Tool

Model

2. Fatigue Audit InterDyne

Model

B. Diary Study of Train Employees on Commuter and Intercity Passenger Railroads

IV. Railroad Safety Advisory Committee Process

A. Overview of the RSAC

B. RSAC Proceedings in this Rulemaking

C. Significant Task Force Contributions to the Development of the NPRM

1. Schedule Analysis

2. Fatigue Mitigation Tool Box

D. Areas of Working Group and Task Force Concern During Development of the NPRM

1. Proposed Definitions of “Type 1 Assignment” and “Type 2 Assignment”

2. Proposed Limitations on Number of Consecutive Days

3. Precision of Fatigue Models and Threshold

4. Freight Railroad Employees Acting as Pilots for Commuter or Intercity Passenger Trains

V. Response to Public Comments on the NPRM

VI. Section-by-Section Analysis

VII. Regulatory Impact and Notices

A. Executive Order 12866 and 13563 and DOT Regulatory Policies and Procedures

B. Executive Order 13132

C. Executive Order 13175

D. Regulatory Flexibility Act and Executive Order 13272

1. Description of Regulated Entities and Impacts

2. Certification

E. Paperwork Reduction Act

F. Unfunded Mandates Reform Act

G. Environmental Assessment

H. Privacy Act

I. Executive Summary

Having considered public comments in response to FRA's March 22, 2011 proposed rule in this rulemaking (76 FR 16200), FRA issues this final rule establishing substantive hours of service regulations for train employees who provide commuter or intercity rail passenger transportation (passenger train employees).

Federal laws governing railroad employees' hours of service date back to 1907 with the enactment of the Hours of Service Act (Pub. L. 59-274, 34 Stat. 1415), and FRA, under delegations from the Secretary of Transportation (Secretary), has long administered statutory hours of service requirements for the three groups of employees now covered under the statute, namely employees performing the functions of train employees, signal employees, and dispatching service employees, as those terms are defined at 49 U.S.C. 21101. See 49 CFR 1.49; 49 U.S.C. 21101-21109, 21303.

These requirements have been amended several times over the years, most recently in the Rail Safety Improvement Act of 2008 (Pub. L. 110-432, Div. A) (RSIA). The RSIA substantially amended the requirements of 49 U.S.C. 21103, applicable to train employees, defined as “individual[s] engaged in or connected with the movement of a train, including a hostler.” 49 U.S.C. 21101(5). However, the RSIA also granted the Secretary

authority to prescribe regulations governing the hours of service of passenger train employees. 49 U.S.C. 21109(b)-(c). As will be discussed below, FRA interprets commuter or intercity rail passenger transportation to include rail passenger transportation by tourist, scenic, excursion, and historic railroads. The RSIA provided that this particular subset of train employees (

i.e.,

passenger train employees) would continue to be governed by 49 U.S.C. 21103 as it existed prior to the enactment of the RSIA (old Section 21103), until the earlier of, the effective date of final regulations prescribed by the Secretary, or the date that is three years from the date of enactment of the RSIA. 49 U.S.C. 21102(c). In the absence of a final rule in effect governing this group of train employees, the requirements of the RSIA currently in effect for other train employees (new Section 21103) would go into effect for passenger train employees on October 16, 2011. 49 U.S.C. 21102(c).

As will be discussed further below, FRA reviewed the applicable fatigue science, and sought input from FRA's Railroad Safety Advisory Committee (RSAC). Based on FRA's understanding of current fatigue science, and information received through RSAC, FRA determined that the requirements imposed on train employees by the RSIA were not appropriate for passenger train employees. The chart below compares and contrasts (1) the hours of service requirements in 49 U.S.C. 21103 as amended by the RSIA, (2) the statutory hours of service requirements applicable to all train employees immediately prior to the RSIA, which are currently still applicable to passenger train employees until the effective date of this final rule, and (3) the requirements of this final rule that applies to passenger train employees from the effective date of the rule, with the compliance date of some provisions delayed for a period of 180 or 545 days from the effective date.

Freight train employee statute

Train employee statutory

provisions immediately prior

to the RSIA and currently

applicable only to passenger

train employees

FRA passenger train employee final rule

Citation

49 U.S.C. 21103 (as amended by the RSIA effective July 16, 2009) (new section 21103) (Applies to train employees on freight railroads. Will apply to train employees on commuter and intercity passenger railroads if no regulations are in effect by October 16, 2011)

49 U.S.C. 21103 as it existed prior to the October 16, 2008, enactment of the RSIA (old section 21103) (Train employees providing commuter and intercity rail passenger transportation are currently covered by these provisions pursuant to 49 U.S.C. 21102(c).)

49 CFR part 228, subpart F.

Use of Fatigue Science

None

This final rule requires passenger train employees' work schedules to be analyzed under an FRA-approved validated biomathematical fatigue model such as the specified version of the Fatigue Avoidance Scheduling Tool

or Fatigue Audit InterDyne

, with the exception of certain schedules (completely within the hours of 4 a.m. and 8 p.m., or nested within other schedules that have been previously modeled and shown to present an acceptable level of risk for fatigue, and otherwise in compliance with the limitations in the regulation) deemed as categorically presenting an acceptable level of risk for fatigue that does not violate the defined fatigue threshold. Analysis must be complete 180 days from the effective date of the final rule, except that tourist, scenic, historic and excursion railroads have 545 days from the effective date of the final rule to complete their analysis.

Limitations on Time on Duty in a Single Tour

12 consecutive hours of time on duty or 12 nonconsecutive hours on duty if broken by an interim release of at least 4 consecutive hours uninterrupted by communication from the railroad likely to disturb rest, in a 24-hour period that begins at the beginning of the duty tour

12 consecutive hours of time on duty or 12 nonconsecutive hours on duty if broken by an interim release of at least 4 consecutive hours, in a 24-hour period that begins at the beginning of the duty tour

Limitations on Consecutive Duty Tours or Total Duty Tours in Set Period

May not be on duty as a train employee after initiating an on-duty period on six consecutive days without receiving 48 consecutive hours off duty free from any service for any railroad carrier at the employee's home terminal. Employees are permitted to initiate a seventh consecutive day when the employee ends the sixth consecutive day at the away-from-home terminal, as part of a pilot project, or as part of a grandfathered collectively bargained arrangement. Employees performing service on this additional day must receive 72 consecutive hours free from any service for any railroad carrier at their home terminal before going on duty again as a train employee

None

If employee initiates an on-duty period each day for six consecutive calendar days including at least one “Type 2” assignment (generally, those including time on duty between 8 p.m. and 4 a.m.) employee must have 24 consecutive hours off duty at the employee's home terminal. If an employee initiates an on-duty period on 13 or more calendar days of a period of 14 consecutive days then must have 2 consecutive calendar days without initiating an on-duty period at the employee's home terminal. Employees may be permitted to perform service on an additional day to facilitate their return to their home terminal. These limitations are effective 180 days from the effective date of the final rule, except that they become effective for tourist, scenic, historic and excursion railroads 545 days from the effective date of the final rule.

Cumulative Limits on Time on Duty

Limited to 276 hours of time on duty, in deadhead transportation to a point of final release, or any other mandatory activity for the railroad carrier

None

None.

Limited to 30 hours of time spent on duty and waiting for or in deadhead transportation to a point of final release after reaching 12 hours of time on duty and waiting for or in deadhead transportation to a point of final release

Mandatory Off-Duty Periods

10 consecutive hours of time off duty free from any communication from the railroad likely to disturb rest, with additional time off duty if on-duty time plus time in or awaiting deadhead transportation to final release exceeds 12 hours

8 consecutive hours (10 consecutive hours if time on duty reaches 12 consecutive hours)

8 consecutive hours (10 consecutive hours if time on duty reaches 12 consecutive hours). This is effective on the effective date of the final rule.

48 consecutive hours off duty, free from any service for any railroad carrier, after initiating an on-duty period for 6 consecutive days. If 7 consecutive days are permitted, mandatory off-duty period extended to 72 consecutive hours

Specific Rules for Nighttime Operations

None

Schedules that include any time on duty between 8 p.m. and 4 a.m. must be analyzed using a validated biomathematical model of human performance and fatigue approved by FRA. Schedules with excess risk of fatigue must be mitigated or supported by a determination that mitigation is not possible and the schedule is operationally necessary and approved by FRA. Analysis must be complete and required submissions must be made 180 days from the effective date of the final rule, except that tourist, scenic, historic and excursion railroads have 545 days from the effective date of the final rule to complete their analysis and any required submission.

Specific Rules for Unscheduled Assignments

None

The potential for fatigue presented by unscheduled work assignments must be mitigated as part of a railroad's FRA-approved fatigue mitigation plan. Plans must be submitted for FRA review and approval along with the associated schedules requiring mitigation, 180 days from the effective date of the final rule, except that tourist, scenic, historic and excursion railroads have 545 days from the effective date of the final rule to complete their analysis and any required submission.

Recordkeeping requirements

Record for each duty tour must contain 15 elements specified in 49 CFR 228.11(b)

Record for each duty tour must contain the first 12 elements specified in 49 CFR 228.11(b), as items 13 through 16 relate to RSIA requirements not applicable to train employees providing commuter or intercity rail passenger transportation

Record for each duty tour must contain the first 12 elements specified in 49 CFR 228.11(b). Record must also indicate the date on which the series of at most 14 consecutive calendar days begins, as well as the date of any calendar day on which the employee did not initiate an on-duty period during the series. These recordkeeping requirements go into effect at the same time as the substantive provisions being tracked by them, which is 180 days from the effective date of the final rule, except that those provisions go into effect for tourist, scenic, historic and excursion railroads 545 days from the effective date of the final rule, as would the associated recordkeeping requirements.

Excess Service Reporting Requirements

Requires reporting of any of 10 different ways in which hours of service limitations may be exceeded

Requires reporting of any of 4 different ways in which hours of service limitations may be exceeded

Requires reporting of any of 8 different ways in which hours of service limitations may be exceeded (reflecting various ways of violating new consecutive-days requirements). These recordkeeping requirements go into effect at the same time as the substantive provisions being tracked by them, which is 180 days from the effective date of the final rule, except that those provisions go into effect for tourist, scenic, historic and excursion railroads 545 days from the effective date of the final rule, as would the associated recordkeeping requirements.

This rule mirrors the existing limitations set by old section 21103 on the maximum number of hours in a duty tour and minimum number of hours in a statutory off-duty period. Additional limitations are added on the number of consecutive days or total days within a prescribed period that a passenger train employee may work, depending on the time of day of the assignment. This differentiation takes into account the fact that work during nighttime hours may present a greater risk for fatigue. (For ease of reference, these provisions of this regulation are collectively referred to as “consecutive-days limitations”). Conforming changes are also made to the recordkeeping and reporting requirements to accommodate the consecutive-days limitations.

The limitations on maximum hours worked, minimum hours of rest, and consecutive days or total days within a prescribed period provide a “floor,” a minimum set of limitations, within which the rule requires railroads subject to this rule to analyze the work schedules of their passenger train employees using a validated and calibrated biomathematical model of human performance and fatigue, and to mitigate any fatigue identified that violates the fatigue threshold for the model.

The fatigue threshold is a level of fatigue at which safety may be compromised. As will be discussed below, especially under Section III.A, there are two models that currently have been validated and calibrated using data from freight railroads, that have been approved by FRA to be used for the analysis required by this rule. The rule also allows for the development of new models. It discusses procedures for validating and calibrating a model, and provides that evidence of a new model's validation and calibration may be submitted to FRA for approval.

In the NPRM, FRA referred to “exceeding” the fatigue threshold. The two currently approved models differ in how their thresholds are expressed, with FAST requiring an effectiveness score greater than or equal to its threshold, and FAID requiring a score less than or equal to its threshold, so FRA realized there could be confusion as to what it meant to “exceed” the threshold depending which model is being used, while it is equally applicable to say the threshold is violated, however that threshold is expressed.

The rule defines as a “Type 1 assignment” any assignment that requires an employee to report for duty no earlier than 4 a.m. and be released from duty no later than 8 p.m. Based on analysis conducted during the formulation of this rule, such assignments are subjected to a less restrictive consecutive-days limitation, and such schedules are deemed to present an acceptable level of fatigue when otherwise in compliance with the limitations established in this rule. Thus, these schedules are not required to be submitted to FRA for approval, nor is the application of fatigue mitigation tools to these schedules required.

A “Type 2 assignment” is any assignment having any period of time during a calendar day before 4 a.m. or after 8 p.m. Within 180 days of the effective date of this regulation, railroads are required to analyze the fatigue risk of assignments that they make to their passenger train employees using an FRA-approved fatigue model. If the analysis shows that a schedule does not violate the fatigue threshold, and the schedule is otherwise in compliance with the limitations of the rule and does not require the employee to be on duty for any period of time between midnight and 4 a.m., the rule allows that schedule to be treated as a Type 1 assignment for the purposes of the consecutive-days limitation, and there is no requirement to submit the schedule to FRA or to mitigate fatigue in that schedule. However, for those schedules that the analysis indicates have a level of risk for fatigue violating the fatigue threshold, the railroad is required to mitigate the fatigue. Railroads are also required to complete their analysis and submit any schedules with a risk violating the fatigue threshold, and the mitigation tools the railroad applied to mitigate the fatigue risk in those schedules to FRA for approval. In addition, any schedule, the fatigue risk of which could not be sufficiently mitigated so that it no longer violates the fatigue threshold, but which the railroad deems operationally necessary, must also be submitted for FRA approval, along with a declaration of operational necessity for FRA approval.

The rule also requires railroads to submit any schedule changes that result in a schedule that would have been required to be submitted if it were an original schedule, unless the new schedule is the same as another schedule that has previously been analyzed and approved.

Within 120 days of any railroad submission, FRA will notify the railroad of any exceptions taken to its submission and the time frame within which the railroad must correct the exceptions. While the rule requires FRA approval of the schedules and fatigue mitigation tools, FRA expects that it will work with a railroad to make necessary modifications to schedules or mitigation tools to minimize fatigue to the greatest extent possible.

Railroads are required to consult with affected employees and applicable labor organizations regarding the analysis of work schedules, fatigue mitigation tools, and submissions to FRA. Should the employees or labor organizations disagree with the railroad, they have the opportunity to file a statement for FRA's consideration in reviewing the submission and determining whether to approve it.

Finally, the rule requires initial fatigue training, addressing a list of subjects, and refresher training every three years. This training may be combined with other training that the railroads are providing to their employees.

FRA analyzed the economic impacts of this rule against two baselines. One is a “status quo” baseline that reflects present conditions (

i.e.,

primarily, the statutory hours of service provisions (specifically, old section 21103) and, secondarily, the hours of service recordkeeping and reporting regulations) that have applied, and will continue to apply to passenger railroads, with respect to their train employees, until this final rule becomes effective. The other baseline is a “no regulatory action” baseline that reflects what would have happened in the absence of this rulemaking (

i.e.,

the freight hours of service laws would have been applied to passenger railroads with respect to their train employees).

With respect to the “no regulatory action” baseline, this rule represents a substantially more cost-effective alternative for achieving the goal of identifying and mitigating unacceptable fatigue risk levels and thus ensuring the safety of passenger train operations. The following table presents the costs of the final rule compared to the “no regulatory action” alternative.

Cost description

No regulatory action alternative— freight HSL

Undiscounted

PV@7%

PV@3%

Final rule

Undiscounted

PV@7%

PV@3%

New Engineer Training, Initial (20% New Hires)

$31,237,549

$26,299,825

$28,705,081

New Engineer Training, Refresher (20% New Hires)

$4,599,050

$2,278,431

$3,327,802

New Conductor Training, Initial (20% New Hires)

$30,847,974

$25,942,971

$28,330,908

New Conductor Training, Refresher (20% New Hires)

$8,636,745

$4,278,146

$6,249,071

Work Schedule Analysis (No-Reg Action)/Initial Analysis of Work Schedules + Follow-up Analysis and Fatigue Mitigation Plan Review (FRA rule)

$189,723

$177,312

$184,198

($126,482 + $240,316) = $366,799

($118,208 + $122,175) = $240,382

($122,798 + $175,894) = $298,692.

Indirect Determination that Type 2 Schedules are Acceptable (“Nested” Schedules Reduction)

−$91,700

−$60,096

−$74,673.

Biomathematical Model of Fatigue Software (Training on model use)

$417,500 (includes $192,500 for training on model use)

$268,723 (includes $119,175 for training on model use)

$337,240 (includes $152,843 for training on model use).

Use of Rest Facilities

$30,988

$28,961

$30,086.

Fatigue Training

$1,312,920

$782,634

$1,025,158.

Fatigue Training (Tourist & Excursion)

$20,000

$12,000

$16,000.

Total (rounded)

$75,511,041

$58,976,685

$66,797,059

$2,056,507

$1,272,605

$1,632,502.

FRA estimates that the recordkeeping and reporting costs per employee record under the no-regulatory action alternative and this rule will be practically the same.

The estimated accident reduction benefits of the rule relative to the statutory hours of service requirements currently in place include prevented accident damages, injuries, and fatalities. The table below presents the estimates for the 20-year period of analysis for the benefits of this rule, which FRA estimates to be the same as the benefits of the no-regulatory action alternative.

Intercity Passenger, Commuter, Tourist and Excursion Railroads

[All track types]

Accident reduction benefits

VSL = $6 M undiscounted benefits

VSL = $6 M discounted PV@7%

VSL = $6 M discounted PV@3%

Property Damage

$685,915

$348,713

$502,039

Injuries

94,861

48,227

69,431

Fatalities

407,634

207,237

298,358

Total (rounded)

1,188,410

604,177

869,828

FRA does not expect that the overall number of casualties and property damages attributable to the rule will differ from those that would be prevented under the statutory freight hours of service requirements. However, as noted above, there are significant additional potential safety enhancement benefits that may result from the FRA approach. FRA believes that the safety of passenger train operations will be enhanced under this rule as a result of subjecting every crew assignment to a biomathematical analysis either via the analyses conducted while developing the RSAC recommendation or the analyses that will be performed by railroads in the years ahead. The information that railroads will have as a result of this rule regarding fatigue, its causes and symptoms, and its impact on

safety, will allow them to make crew assignments that take this into consideration and minimize fatigue beyond the requirements of this rule. Based on its literature review, FRA is confident that, overall, fatigue awareness training will positively contribute to a stronger safety culture that will extend beyond railroad operations, which is a benefit that extends beyond what would result under the freight hours of service law. For instance, safety and health benefits may accrue from the transfer of knowledge to employees, their families, friends, and others with whom they may share the fatigue knowledge that they acquire from the required fatigue awareness training programs. This fatigue awareness may result in more optimal decisions regarding rest and sleep, leading to less fatigue and improved safety outside of passenger train operations during the course of daily activities that may also include the operation of motor vehicles or other heavy machinery. This fatigue awareness may also result in proper identification and treatment, if necessary, of fatigue symptoms. Although FRA has not identified research on the effectiveness of the specific types of fatigue training programs required under this rule, many studies have indicated health training programs in general produce meaningful behavioral performance improvements.

With respect to the “status-quo” baseline, this rule imposes costs that are higher than the safety benefits FRA was able to quantify. Costs compared to the “status quo” baseline total $2.1 million (undiscounted), $1.3 million (PV, 7 percent), and $1.6 million (PV, 3 percent). Quantified benefits compared to the “status quo” baseline total $1.2 million (undiscounted), $0.6 million (PV, 7 percent), and $0.9 million (PV, 3 percent). However, there are additional benefits that have not been quantified, but should be considered when comparing the overall costs and benefits. As when compared to the “no-regulatory action” baseline, FRA believes that the safety of passenger train operations will be enhanced under this rule as a result of a stronger safety culture that may extend beyond railroad operations, which would be a benefit that extends beyond what would result under the freight hours of service law. Separately, accident avoidance will result in fewer unplanned delays to passengers and freight commodities impacted by passenger train accidents and incidents that result in blocking one or more tracks for prolonged periods. These costs can be very substantial given the need to investigate accidents and often clear wreckage. It is not unreasonable to expect that the unquantified benefits will raise the benefits to a level quite comparable to the costs. FRA also believes that the unquantified benefits coupled with the quantified safety benefits are comparable to the costs associated with meeting the intent of the statutory mandate.

After careful consideration of comments received in response to the NPRM, FRA has made modifications to its proposal in the final rule that reduce the overall burden by approximately $100,000 due in equal part to flexibilities added by extending the deadline for fatigue awareness training and the expanded ability to rely on the findings of analyses conducted for other assignments.

II. Statutory Background and History

Federal laws governing railroad employees' hours of service date back to 1907 with the enactment of the Hours of Service Act. These laws, codified as amended primarily at 49 U.S.C. 21101-21109, are intended to promote safe railroad operations by limiting the hours of service of certain railroad employees and ensuring that they receive adequate opportunities for rest in the course of performing their duties. Public Law 103-272 (1994). The Secretary is charged with the administration of those laws, collectively referred to in this document as the hours of service laws (HSL). This function has been delegated to the FRA Administrator. 49 U.S.C. 103(c); 49 CFR 1.49(d).

Congress substantially amended the HSL on three occasions. The first significant amendments occurred in 1969. Public Law 91-169, 83 Stat. 463. The 1969 amendments reduced the maximum time on duty for train employees

from 16 hours to 14 hours effective immediately, with a further reduction to 12 hours automatically taking effect two years later. Congress also established provisions for determining, in the case of a train employee, whether a period of time is to be counted as time on duty. 49 U.S.C. 21103(b). In so doing, Congress also addressed the issue of deadhead transportation

time, providing that “[t]ime spent in deadhead transportation to a duty assignment” is counted as time on duty. Although time spent in deadhead transportation from a duty assignment to the point of final release is not included within any of the categories of time on duty, Congress further provided that it shall be counted as neither time on duty nor time off duty. 49 U.S.C. 21103(b)(4). This provision effectively created a third category of time, known commonly as “limbo time.”

A “train employee” is defined at 49 U.S.C. 21101(5) and 49 CFR 228.5 as an individual engaged in or connected with the movement of a train, including a hostler. FRA also interpreted this statutory term in published interpretations in 49 CFR part 228, appendix A, providing: “Train or engine service refers to the actual assembling or operation of trains. Employees who perform this type of service commonly include locomotive engineers, firemen, conductors, trainmen, switchmen, switchtenders (unless their duties come under the provisions of section 3 [49 U.S.C. 21105]) and hostlers.” Other employees, such as food service providers or sleeping car attendants, who may work on passenger trains, but have no responsibility for assembling or operating the train, are not within the definition of a train employee, and are, as such, not generally covered by this rule, or any other hours of service limitations, but they would be covered if they performed functions related to assembling or operating the train, regardless of the employee's job title.

“Deadheading” is defined at 49 CFR 228.5 as the physical relocation of a train employee from one point to another as a result of a railroad-issued verbal or written directive.

In 1976, Congress again amended the HSL in several important respects. Most significantly, Congress expanded the coverage of the laws, by including hostlers within the definition of employees now termed “train employees,” and adding the section providing hours of service requirements for “signal employees,” now codified at 49 U.S.C. 21104. Congress also added a provision that prohibited a railroad from providing sleeping quarters that are not free from interruptions of rest caused by noise under the control of the railroad, and that are not clean, safe, and sanitary, and prohibited beginning construction or reconstruction of sleeping quarters in an area or in the immediate vicinity of an area in which humping or switching operations are performed after July 7, 1996. See Public Law 94-348, 90 Stat. 818 (1976).

Section 108 of the RSIA also amended the HSL in a number of significant ways, most of which became effective July 16, 2009. See Section 108 of Public Law 110-432, Div. A, and FRA Interim Statement of Agency Policy and Interpretation at 74 FR 30665 (June 26, 2009). The RSIA established a limit of 276 hours per calendar month for train employees on service performed for a railroad and on time spent in or waiting for deadhead transportation to a point of final release, increased the quantity of the statutory minimum off-duty period after being on duty for 12 hours in broken service from 8 hours of rest to 10 hours of rest, prohibited communication with train or signal employees during certain minimum statutory rest periods, and established mandatory time off duty

for train employees of 48 hours after initiating an on-duty period on six consecutive days, or 72 hours after initiating an on-duty period on seven consecutive days. 49 U.S.C. 21103-21104. The RSIA also revised the definition of “signal employee” to include contractors who perform the work of a signal employee within the scope of the statute. 49 U.S.C. 21101(4).

However, Section 108(d) of the RSIA, which became effective on October 16, 2008, provided that the requirements described above for train employees would not go into effect on July 16, 2009, for train employees when providing commuter or intercity rail passenger transportation. 49 U.S.C. 21102(c). Section 108(d) further provided that these train employees, who provide commuter or intercity passenger rail service, would continue to be governed by the old HSL (as they existed immediately prior to the enactment of the RSIA, at 49 U.S.C. 21103 prior to its 2008 amendment), until the effective date of regulations promulgated by the Secretary. 49 U.S.C. 21102(c). However, if no new regulations are in effect before October 16, 2011, the provisions of Section 108(b), which applied to train employees, would be extended to these employees at that time.

Id.

Section 108(e) of the RSIA specifically provides the Secretary with the authority to issue hours of service rules and orders applicable to train employees engaged in commuter rail passenger transportation and intercity rail passenger transportation (as defined in 49 U.S.C. 24102), that may be different from the statute applied to other train employees. 49 U.S.C. 21109(b). Section 108(e) of the RSIA further provides that such regulations and orders may address railroad operating and scheduling practices, including unscheduled duty calls, communications during time off duty, and time spent waiting for deadhead transportation or in deadhead transportation from a duty assignment to the place of final release, that could affect employee fatigue and railroad safety.

Id.

Section 108(e) of the RSIA also provides—

[i]n issuing regulations under subsection (a) the Secretary shall consider scientific and medical research related to fatigue and fatigue abatement, railroad scheduling and operating practices that improve safety or reduce employee fatigue, a railroad's use of new or novel technology intended to reduce or eliminate human error, the variations in freight and passenger railroad scheduling practices and operating conditions, the variations in duties and operating conditions for employees subject to this chapter, a railroad's required or voluntary use of fatigue management plans covering employees subject to this chapter, and any other relevant factors.

49 U.S.C. 21109(c). Section 21109(a) of title 49 of the U.S. Code refers to other regulatory authority granted to FRA, as the Secretary's delegate related to the HSL, which is not relevant to this rule. One of the goals of the present rulemaking is to identify and reduce unacceptable fatigue for the employees who will be covered by the final rule. Therefore, as will be described below, FRA has based these regulations on scientific research related to fatigue and fatigue abatement, as applied to railroad scheduling practices and operating conditions for train employees providing commuter and intercity rail passenger transportation. Section III below will describe the primary scientific foundation and support for the requirements contained in this rule. In addition, scientific considerations will also be addressed in discussion of various elements of the rule, including in the discussion of specific provisions in Section VI, Section-by-Section Analysis, below.

III. Scientific Background

Most mammals, including human beings, have an approximately 24-hour sleep-wake cycle known as a “circadian rhythm.” Rapid changes in the circadian pattern of sleep and wakefulness disrupt many physiological functions such as hormone releases, digestion, and temperature regulation. Physiological functions can be affected, performance may be impaired, and a general feeling of fatigue and debility may occur until realignment is achieved. Jet lag, a commonly experienced syndrome when flying across several time zones, especially when flying east, is similar to the experience of individuals working schedules with abrupt changes in the timing of work and subsequent sleep.

Fatigue risk in an industry that operates 24 hours a day and 7 days a week is not just dependent on how many hours per day a person is permitted to work, or the amount of time that a person is required to be off duty between periods of work. Other significant factors in the level of fatigue risk include the time of day that an employee works, the number of consecutive hours worked, direction and rate of schedule rotation, and the number of consecutive days that an employee works. In addition, the quantity and quality of sleep vary with the time of day and environmental conditions in which sleep occurs. Furthermore, there are significant individual factors such as sleep disorders, age and time of day (

e.g.,

morning or evening that may affect one's fatigue and alertness. Because of natural circadian rhythms and environmental and social factors, most people are able to achieve the best quality and most restful sleep at night.

The railroad industry by necessity is a continuous operation, 24 hours a day, seven days a week, 365 days a year, including both day and night work. Consequently, fatigue risk mitigation is a very important strategy of a railroad safety management system. In fact, the design and operation of the work schedule system are perhaps the most essential elements of that fatigue risk management strategy. While the purpose of any work schedule system is to provide the organization with a methodical means of organizing the timing and structure of work to maximize efficiency and productivity, seldom are these schedules designed to minimize the safety risks associated with work schedules that are incompatible with human biological limitations, such as our circadian system. Because the railroad industry is a continuous service industry, and because both employees and the general public are exposed to the safety risks associated with railroad operations, researchers have long called for validated fatigue models to better identify and mitigate fatigue-related risks associated with work scheduling.

See Pilcher and Coplen, Ergonomics, 2000, Vol. 43, No. 5, 573-588.

The general purpose for a regulation requiring an industry to use a valid fatigue model is to impose a minimum standard for identifying and mitigating fatigue risk that otherwise might not occur without such a standard. These models take into account the complex interaction between human physiology and work times, something that would be very difficult to specify through other means. Use of fatigue modeling tools to evaluate work schedules, however, is just one aspect of mitigating fatigue risk in a larger system. While FRA intends to enforce the minimum standards in the regulatory text, including those related to fatigue models, it also hopes that the industry will go beyond compliance with this standard by using the models and other tools to assess and address fatigue risk across the system.

FRA notes that other provisions of the RSIA mandate issuance of regulations requiring certain railroads to implement railroad safety risk reduction programs and plans; one component of each plan is a fatigue management plan. See 49 U.S.C. 20156.

For example, if a fatigue model were to identify a particular type of work

schedule that violates the model's fatigue threshold, and thus requires fatigue mitigation, the carrier may discover underlying systems issues and factors (

e.g.,

inadequate rest facilities,

etc.

) that contribute to fatigue-related risks on not only that work schedule, but also on other less fatiguing schedules that do not violate the fatigue threshold. The use of fatigue modeling in this way, then, provides the organization with a method for systematically identifying and addressing the underlying system risks, as opposed to those risks only for a given work schedule. In going beyond compliance with the minimum standard, the organization also builds its organizational capacity for mitigating fatigue as a major safety risk factor across the system.

As previously mentioned, the statutory hours of service requirements currently in effect for train employees providing commuter and intercity rail passenger transportation establish a maximum on-duty time of 12 hours in a 24-hour period, and a minimum off-duty time of 8 hours in a 24-hour period, or 10 hours after a period of 12 consecutive hours on duty. Statutory requirements applicable to train employees on freight railroads, as revised by the RSIA, include a limitation on the number of consecutive days on which a train employee may initiate an on-duty period. However, the HSL for the railroad industry have never, up to the present day, differentiated in their requirements based on the time of day in which service is performed, or the time of day that a period is available for rest.

As will be discussed further below, FRA conducted two work/rest diary studies with train employees in freight and passenger operations. Data from these studies indicate that train employees get more sleep than the average U.S. adult. While 46 percent of U.S. adults get less than seven hours of sleep, only 35 percent of freight train employees and 41 percent of passenger train employees get less than seven hours of sleep. This amount of sleep results in a level of fatigue that increases accident risk by 21 to 39 percent.

Moreover, certain operational characteristics of commuter and intercity passenger service mitigate the fatigue associated with this amount of sleep loss relative to freight service. For example, many train employees on commuter and intercity passenger railroads work scheduled assignments, in which they begin and end their work day at approximately the same time each day. These employees also usually begin and end their duty tour at the same location, meaning that they can go home at the end of their work day and sleep in their own beds. In addition, very few scheduled assignments on most railroads operate during late night hours, and many of them result in duty tours significantly shorter than the maximum hours that the employee would be allowed to remain or go on duty under the existing law or this regulation. Because these characteristics are more likely to allow for periods of rest that are consistent with normal circadian rhythms, they will provide better opportunities for rest, and less risk for fatigue. In addition, as will be discussed further below, two FRA work/rest diary studies demonstrate that levels of fatigue are not equivalent in freight and passenger operations (Work Schedules and Sleep Patterns of Railroad Train and Engine Service Workers

http://www.fra.dot.gov/downloads/Research/ord0922.pdf

) (which included data from a small number of train employees in passenger operations); Work Schedules and Sleep Patterns of Railroad Train and Engine Employees in Passenger Operations

http://www.fra.dot.gov/rpd/downloads/TR_Work_Schedules_and_Sleep_Patterns_final.pdf

(the diary study conducted to support this rulemaking).

See Hursh,

et al. infra

at footnote 7.

For all of these reasons, FRA has determined that some of the specific limitations that Congress applied to train employees on freight railroads in the RSIA are not appropriate for train employees on commuter and intercity passenger railroads.

However, FRA also recognizes that some train employees covered by this rule experience a level of fatigue at which safety may be compromised. This is particularly true of those employees who do not work scheduled assignments and may not return home at the end of each duty tour, or who are required to perform service during late night hours, or to work duty tours of the maximum length allowed by existing requirements, with only the minimum required rest between duty tours. FRA has attempted, in this regulation, to specifically address those employees who are most at risk for fatigue, even when in compliance with specific hours of service limitations. As will be discussed below, research that resulted in the validation of fatigue models using data from freight railroads demonstrated that fatigue increases the risk of a human factors accident. In addition, as will be discussed below, diary data show the risk of fatigue in passenger operations. The risk must be measured in order to be managed, and fatigue models allow for that measurement.

An effective proactive fatigue risk management program needs to balance the amount of work performed against when the work is performed, how long a work schedule is in effect in terms of hours in a day, consecutive days, and other variables. This regulation addresses fatigue risk by going beyond establishing limitations on the amount of time that an employee may work, and the minimum amount of time that an employee must be off duty between duty tours. It additionally requires the analysis of the fatigue risk in employee work schedules using a biomathematical model of performance and fatigue, identification of those schedules that present an unacceptable level of fatigue risk, and mitigation of the identified fatigue risk. In addition, the regulation establishes different requirements for schedules of employees who operate trains during the late night hours in which the fatigue risk is greatest. Thus, the rule will specifically address those schedules the characteristics of which present a risk for fatigue, even when otherwise in compliance with required maximum on-duty and minimum off-duty periods and other limitations. These risks would not be addressed by a regulation that simply established maximum on-duty and minimum off-duty periods, just as they are not addressed by the existing statutory requirements.

A. Validated and Calibrated Fatigue Models

For a discussion of existing models and their application, see Dean II, D.A., Fletcher, A., Hursh, S.R. and Klerman, E.B.,

Developing Models of Neurobehavioral Performance for the “Real World,”

J. Biol. Rhythms 2007; 22; 246.

A biomathematical model of performance and fatigue that has been properly validated and calibrated predicts accident risk based on analysis of identified periods of wakefulness and periods available for sleep.

“Validation” of a biomathematical model of human performance and fatigue means determining that the output of the model actually measures human performance and fatigue. There are two dimensions to this validation. The first is that the model must be demonstrated to be consistent with currently established science in the area of human performance, sleep, and fatigue. The second part of the validation process involves determining that the model output has a statistically reliable relationship with the risk of a human factors accident caused by fatigue, and that the model output does

not have such a relationship with nonhuman factors accident risk.

In general, and for the purpose of compliance with this rule, a model will be validated if statistical analyses demonstrate the existence of a statistically significant relationship between the output of the model and the human factors accident risk ratio, and the absence of such a relationship between the output of the model and the non-human factors accident risk ratio. The presence of a statistically significant relationship is evaluated by way of the correlation coefficient (r) with statistical significance requiring a

-value of less than 0.05. The first step is the selection of bin

edges that correspond to varying levels of fatigue, (

e.g.,

the “not fatigued” bin and the “severely fatigued” bin). The “not fatigued” bin is determined by the output of the model when sleep occurs or can occur for 8 or more hours, without abrupt phase changes, between 10 p.m. and 10 a.m. This is similar to the amount of fatigue produced by the standard 9 a.m. to 5 p.m., Monday through Friday work week. The performance bin “severely fatigued” is determined by the output of the model when there is total sleep deprivation for 42.5 hours after waking at 7 a.m. This is similar to the amount of fatigue produced by a permanent night shift schedule with six consecutive 12-hour work periods followed by 1 day off. These two bins are the “anchor” bins for the validation procedure. Four additional bins, equally spaced between the anchor bins, accommodate the intermediate fatigue scores.

In statistics, a “bin” is a discrete, non-overlapping interval of a variable. Here, the variable is the level of fatigue.

Calibration is, in general, the assignment of numerical values to represent aspects of empirical observations. In the case of human fatigue and performance, the calibration of a fatigue scale would start with the assignment of values to “not fatigued,” and the most fatigued condition might be described as “severely fatigued.” The calibration process starts during the validation process with the assignment of model output values to anchor bins for “not fatigued” and “severely fatigued.” The next step consists of determining the fatigue threshold. Given a scale for human fatigue and performance and a relationship between that scale and human factors accident risk, a final calibration point would be to determine the fatigue value at which fatigue becomes unacceptable because the increase in accident risk at that level compromises safety. This is the fatigue threshold.

For the purposes of this regulation, the fatigue threshold is referred to as a level of fatigue at which safety may be compromised, in recognition of the fact that while it is possible to determine the level of fatigue expected to be produced by working a certain schedule, that is not necessarily the exact level of fatigue experienced by each individual employee working that schedule.

The procedure for determining the fatigue threshold consists of several computations. First, the cumulative risk for the six fatigue score bins is determined for human factor and non-human factor accidents. Next, a 95-percent confidence interval is calculated for the cumulative risk in each bin. Finally, the fatigue score bin in which human factor cumulative risk exceeds both human factors Accident Risk Ratio = 1 and the mean non-human factors risk is determined. This is the fatigue threshold for the model.

A model may also be calibrated by reference to a model that has been previously validated and calibrated, as discussed in Section III.A.2, below.

The accident risk is defined as an odds ratio, expressed as a percentage of accidents occurring when employees involved in the accident are within a given range of fatigue, divided by the percentage of time spent by the individual working in that given range of predicted fatigue. For example, if 20 percent of accidents occur when an employee is within a particular range of predicted fatigue, and 10 percent of an employee's time in a given duty tour is spent within that range of predicted fatigue, then that specific range of predicted fatigue has doubled the accident risk.

For more information on the proper procedures for validation and calibration of a biomathematical model of performance and fatigue, see Raslear, T.G.,

Criteria and Procedures for Validating Biomathematical Models of Human Performance and Fatigue; Procedures for Analysis of Work Schedules.

(A copy of this document has been placed in the docket for this rulemaking.)

1. Fatigue Avoidance Scheduling Tool

Model

FRA-sponsored research resulted in the development of a Sleep, Activity, Fatigue, and Task Effectiveness (SAFTE) model and Fatigue Avoidance Scheduling Tool

(FAST) that have been validated and calibrated using data from freight railroads. FAST is a biomathematical model of performance and fatigue that can be used to assess the risk of fatigue in work schedules and to plan schedules that ameliorate fatigue. The model takes into account the time of day when work occurs (circadian rhythm) and opportunities for sleep based on work schedules.

For a description of the FAST model, see Hursh, S. R., Redmond, D. P., Johnson, M. L., Thorne, D. R., Belenky, G., Balkin, T. J., Storm, W. F., Miller, J. C., and Eddy, D. R. (2004).

Fatigue models for applied research in warfighting. Aviation, Space, and Environmental Medicine,

75,

A44-53.

The model validation used work histories from 400 human factors accidents and 1,000 non-human factors accidents on freight railroads. The model has not specifically been validated using passenger railroad accidents, because there were not enough such accidents in the relevant time period to obtain statistically significant results, and had the period of analysis been extended sufficiently to capture enough passenger railroad accidents, much of the needed work schedule data for the employees involved in those accidents would no longer be available. However, FAST measures fatigue and effectiveness, based on laboratory analysis of cognitive and sensory motor functions during sleep deprivation, which are not job specific. Furthermore, the tasks associated with freight and passenger train operations are actually highly similar. In addition, there was no statistically significant difference between the proportion of accidents in categories associated with fatigue, between freight and passenger railroads. For all of these reasons, FRA has determined that the model is valid for use in evaluating fatigue levels in passenger railroad schedules for the purposes of this rule. Indeed, the FAST model has been used by other entities, including the military and the airline industry.

FAST was used to calculate cognitive effectiveness (the inverse of fatigue) on a scale from 0 (worst) to 100 (best) using the 30-day work histories of locomotive engineers prior to the accidents and at the time of the accidents.

Cognitive effectiveness is a metric that tracks speed of performance on a simple reaction time test and is strongly related to overall response speed, vigilance, and the probability of lapses.

Hursh, S. R., Raslear, T. G., Kaye, A. S., and Fanzone, J. F. (2006).

Validation and calibration of a fatigue assessment tool for railroad work schedules, summary report

(Report No. DOT/FRA/ORD-06/21). Washington, DC: U.S. Department of Transportation.

http://www.fra.dot.gov/downloads/Research/ord0621.pdf;

Hursh, S. R., Raslear, T. G., Kaye, A. S., and Fanzone, J. F. (2008).

Validation and calibration of a fatigue assessment tool for railroad work schedules, final report

(Report No. DOT/FRA/ORD-08/04). Washington, DC: U.S. Department of Transportation.

http://www.fra.dot.gov/downloads/Research/ord0804.pdf.

The analysis revealed a significant high correlation between reduced predicted crew effectiveness (as a result of increased fatigue) and the risk of a human factor accident for freight railroads. As was discussed above,

although FAST was validated using freight railroad accidents, the cognitive and sensory motor functions it measures are not job specific, so the resulting determinations of effectiveness and accident risk are equally applicable to passenger railroads. There was no significant relationship between increased fatigue and non-human factor accidents. In addition, the data showed that there is a reliable relationship between the time of day of human factor accidents and the expected, normal circadian rhythm. The circadian pattern was not reliably present for non-human factor accidents. The risk of a human factor accident is increased by 20 percent by working during the hours from midnight to 3 a.m.

Id.

The study showed that there is an elevated risk of human factors accidents at any effectiveness score below 90, and accident risk increased as effectiveness decreased. The risk of a human factors accident is increased by 21 percent at effectiveness scores at or below 70, which is a level of risk elevated beyond chance level, and greater than the mean risk of non-human factor accidents. Twenty-three percent of the freight accidents examined occurred when an employee involved was at or below an effectiveness score of 70. The study also found that cause codes associated with accidents that occurred at or below an effectiveness score of 70 showed an over-representation of the type of human factors accident that might be expected of a fatigued crew, such as passing a signal indicating stop, or exceeding the maximum authorized speed, which confirmed that the detected relationship between accident risk and predicted effectiveness is meaningful.

Other research, comparing the effects of alcohol and sleep deprivation on performance on a driving simulator, has also indicated that an effectiveness score of 70 is the rough equivalent of a 0.08 blood alcohol level, or the equivalent of being awake for 21 hours following an 8-hour sleep period the previous night.

However, direct comparisons between the performance effects of alcohol and fatigue must be made with caution. Some aspects of a complex task, such as driving an automobile simulator, show a high degree of congruence between the effects of alcohol and fatigue, while the effects of alcohol and fatigue on other aspects of the same task are highly dissimilar. For instance, Arnedt

et al.

(2001) found that tracking, tracking variability, and speed variability were all similarly affected by alcohol and fatigue in a driving simulator. However, Arnedt

et al.

found that, while subjects drove faster after consuming alcohol, fatigue did not affect driving speed. In addition, alcohol produced a more rapid deterioration in performance in off-road events (incidents in which the simulated vehicle was driven off the road) than did fatigue. Thus, while it is clear that alcohol and fatigue can both cause deterioration in task performance, the effect of alcohol is often more severe and extensive.

See

Arnedt, J.T., Wilde, G.J., Munt, P.W., and MacLean, A.W. (2001). How do prolonged wakefulness and alcohol compare in the decrements they produce on a simulated driving task?

Accident Analysis and Prevention, 33,

3, 337-44; Dawson, D., and Reid, K. (1997). “Fatigue, alcohol and performance impairment.”

Nature 388,

23.

See also Williamson, A., Feyer, A.M., Friswell, R., and Finlay-Brown, S. (2000).

Development of Measures of Fatigue: Using an Alcohol Comparison to Validate the Effects of Fatigue on Performance

(Road Safety Research Report CR 189). Canberra, Australia: Australian Transport Safety Bureau.

As a result of this analysis, a fatigue threshold (the fatigue level at which there is an unacceptable accident risk due to fatigue) of 70 was established for FAST.

Accordingly, an effectiveness score less than or equal to 70 violates that threshold for the purposes of this regulation.

See Hursh,

et al., supra

note 7.

A 21-day free trial of the FAST Model can be downloaded at

http://fatiguescience.com/products/fast

2. Fatigue Audit InterDyne

Model

For a description of FAID, see Roach, G. D., Fletcher, A., and Dawson, D. (2004). A model to predict work-related fatigue based on hours of work.

Aviation, Space, and Environmental Medicine, 75,

A61-9.

Another biomathematical model of performance and fatigue that has recently been validated and calibrated is the Fatigue Audit InterDyne

(FAID). FAID was validated and calibrated using the same accident data from freight railroads as FAST used.

For the same reasons described above with regard to FAST, FRA has determined that FAID is valid for use in evaluating fatigue levels in passenger railroad schedules for the purposes of this rule.

For details see Tabak, B., and Raslear, T. G. (2010).

Procedures for Validation and Calibration of Human Fatigue Models: The Fatigue Audit InterDyne (FAID) Tool

(Report No. DOT/FRA/ORD-10/14). Washington, DC: U.S. Department of Transportation. (

http://www.fra.dot.gov/rpd/downloads/TR_Procedures_or_Validation_and_Calibration_final.pdf

) (“FAID validation report”).

Analysis of the FAID scores resulted in a statistically significant correlation for human factor accidents and no statistically significant correlation for non-human factor accidents, which meant that FAID could be validated for freight railroads, and, as explained above, FRA has determined that it is equally applicable to passenger railroads. The FAID model was validated with scores of 40 and 120, corresponding to “not fatigued” and “extremely fatigued.” FAID scores showed a statistically reliable relationship (

-value below .05) with the risk of a human factors accident but did not show such a relationship with other accident risk.

Id.

at 9.

However, in analyzing the FAID data for the purpose of calibration, none of the confidence intervals demonstrated a statistically significant increase in cumulative risk. This was true for both human factors and non-human factors accidents. An alternative procedure, using FAST, which was already a validated and calibrated model, allowed for calibration of FAID. The alternative procedure required correlating FAST and FAID scores. The calibration of FAST is the equivalent of fundamental measurement in physics, while the calibration of FAID by reference to FAST is the equivalent of derived measurement, both of which are valid measurement methods.

Kranz, D.H., Luce, R.D., Suppes, P., and Tversky, A. (1971). Foundations of measurement. Volume 1. Additive and polynomial representations. New York: Academic Press.

Correlation of individual FAST and FAID scores found a high level of variation in the individual FAST scores within a FAID bin, so linking fatigue scores on an individual level was not feasible. An alternative method is to calculate confidence intervals for the population, or mean, score. Since biomathematical models are known to be more accurate at predicting population behavior rather than individual behavior, the confidence intervals of the bin means were compared. When analyzed at the population level, the regression line for FAID scores as a function of FAST scores, or FAST scores as a function of FAID scores, has an r of 0.909.

The calibration of FAID indicated that FAID scores above 80 indicate a severe level of fatigue, and that FAID scores between 70 and 80 indicate extreme fatigue. A fatigue threshold (as with FAST, the fatigue level at which there is an unacceptable accident risk due to fatigue) of 60 was established for FAID in its validation report, and an effectiveness score greater than or equal to 60 would violate that threshold.

A free trial of the FAID Model can be downloaded at

http://www.faidsafe.com/products-main.htm#faid330.

In the NPRM, FRA proposed to use the threshold of 60 to trigger the requirements to mitigate fatigue in work

schedules analyzed using FAID. However, following publication of the NPRM, further schedule analysis revealed that some schedules that had an acceptable level of risk for fatigue when analyzed using FAST, violated the proposed FAID threshold when analyzed using FAID, including schedules, to be discussed in detail below, that included work entirely between the hours of 4 a.m. and 8 p.m. 10.6 percent of these schedules violated the proposed FAID threshold rather than the 2.5 percent expected. Schedules wholly within these hours are defined as “Type 1 assignments,” that are deemed not to violate the fatigue threshold, are not required to be analyzed, mitigated or submitted to FRA for approval, and are subjected to a less restrictive consecutive-days limitation. Representatives of the Association of American Railroads (AAR) and the American Public Transportation Association (APTA) who actively participated in the development of this rulemaking, submitted to FRA data illustrating this issue, and suggestions for addressing it. These documents have been added to the docket for this rulemaking.

The calibration of FAID, as indicated in its validation report,

and described above, was not successful as a direct process using the fatigue accident validation database. Instead, an indirect process in which values for FAID were related to values for FAST was used. This process is similar to calibrating a measurement instrument by reference to a known standard. In this case, FAST is the known standard because it was directly calibrated using the fatigue accident validation database. There is inherent variability in both FAST and FAID values, so FRA used regression analysis, a statistical method, to determine the estimated mean values of FAID that correspond to mean values of FAST. Table 8 in the FAID validation report shows the corresponding approximate values for FAID and FAST using this procedure. The exact threshold for FAID, as noted in its validation report, is 63.18,

as calculated from the regression equation: FAID score = 149 − 1.227 × (FAST score). Taking into account the variability associated with predicting mean FAID scores from mean FAST scores, a range of FAID scores that is highly likely to include the true mean FAID score can be calculated. The upper 99-percent confidence limit

for estimating FAID at FAST = 70 is 72.16. This means that we can expect the true mean FAID score to be as high as 72.16.

Tabak and Raslear, infra note 19.

Id.

at 24.

The upper 99-percent confidence limit represents the highest value of a variable within a 99-percent confidence interval. The 99-percent confidence interval is a range of values with a 0.99 probability of including the true population value of a variable.

Allowing the FAID threshold for fatigue to be as high as 72 reduces the percentage of schedules that violate the FAID threshold from 10.6 percent to 2.11 percent in the data presented by AAR and APTA. The passenger train and engine diary study (Work Schedules and Sleep Patterns of Railroad Train and Engine Service Employees in Passenger Operations. DOT/FRA/ORD-11/05), which will be discussed in detail in Section III.B below, indicates that none of the employees subject to this regulation work more than 2.5 percent of the time at a FAST score of ≤ 70. Therefore, FRA concludes that allowing the FAID threshold to be placed at the upper 99 percent confidence limit of 72 is a reasonable solution to this issue. FRA expects that the percentage of schedules that violate a FAID threshold of 72 would be approximately 2.5 percent, which will allow the railroads to focus mitigation efforts on those schedules that are at greater risk for producing an unacceptable level of fatigue and thereby reduce fatigue-related accidents and injuries.

FRA believes that the prediction of the effectiveness of an employee's performance may be used to improve work schedules, to alter to the extent possible the timing of safety-critical tasks to coincide with periods of optimal performance, and to apply countermeasures to reduce the fatigue risk, and the corresponding risk of accidents or other errors associated with that fatigue. It is for this reason that FRA has concluded that it is appropriate to require analysis of employee work schedules using a validated and calibrated biomathematical model of performance and fatigue, as an essential component of these hours of service regulations.

As will be discussed in detail below, this rule requires railroads to mitigate the fatigue resulting from following a certain work schedule, and submit the schedules and fatigue mitigations to FRA for approval. These requirements will be triggered when analysis reveals that an employee working a given schedule will experience 20 percent or more of his or her working time during the schedule at an effectiveness score violating the fatigue threshold under the model used for analysis; that is to say, at an effectiveness score of 70 or less determined by FAST, or at an effectiveness score of 72 or greater as determined by FAID. The applicable effectiveness score could be different if a railroad were using another model that had been properly validated and calibrated. FRA encourages the development, validation, and calibration of alternative models, and their submission to FRA for approval under § 228.407(c), by any railroad desiring to use an alternative model for the analyses required by this rule.

FRA expects fatigue science to continue to develop, and also anticipates the implementation of the rule will assist the agency in better assessing the role of fatigue in accidents that may occur in the future. As a result, FRA will consider such developments and new evidence in its regularly-scheduled retrospective review of the rule, and will expedite that review of the rule should evidence suggest such review is appropriate.

See Raslear,

supra

note 11 for information on procedures for validating and calibrating a model.

B. Diary Study of Train Employees on Commuter and Intercity Passenger Railroads

To further support this rule, FRA conducted primary research specifically directed to train employees of commuter and intercity passenger railroads (Office of Management and Budget (OMB) Control Number 2130-0588).

The results of the study provided valuable evidence of the actual levels of fatigue experienced by train employees on commuter and intercity passenger railroads, because the study allowed analysis of the actual periods of time that an employee reports having worked, slept, or spent in other activities during the period analyzed, which may be different from the assigned schedule and presumed periods available for sleep.

http://www.fra.dot.gov/rpd/downloads/TR_Work_Schedules_and_Sleep_Patterns_final.pdf

FRA had previously conducted similar surveys for signal employees (OMB Control Number 2130-0558), maintenance of way employees (OMB Control Number 2130-0561), dispatching service employees (OMB Control Number 2130-0570), and train employees generally (OMB Control Number 2130-0577). The purpose of these studies was to characterize, using a consistent statistical survey methodology, the work schedules and sleep patterns of each unique group of railroad workers. Because each of these studies used a random sample of each worker population, they provide defensible and definitive data on work/rest cycle parameters and fatigue for the

respective group. The small number of train employees on commuter and intercity passenger railroads represented in the previous study of train employees generally did not allow for meaningful conclusions with regard to this subpopulation of train employees. As a result, the present study, specifically focused on this population, was necessary. The present study of train employees on commuter and intercity passenger railroads used the same methodology as the previous studies.

The primary objectives of this study were to design and conduct a survey to collect work schedule and sleep data from train and engine service (T&E) employees, and to analyze the data to characterize the work/sleep patterns and to identify work schedule-related fatigue issues. The goal was to characterize train employees on commuter and intercity passenger railroads as a group, not to characterize such employees on a specific railroad.

The research described in this report had three phases: preparation; field data collection; and data analysis. Since no existing source would provide answers to the study's research questions, a survey of train employees was the only means to obtain the necessary data. The preparation phase included securing approval from the OMB for the proposed data collection. Representatives from the Brotherhood of Locomotive Engineers and Trainmen (BLET) and the United Transportation Union (UTU) worked closely with the researchers throughout the study.

The study used two survey instruments—a background survey and a daily log. Survey participants used the background survey to provide demographic information, descriptive data for their type of work, type of position, work schedule, and a self-assessment of overall health. The daily log provided the means for survey participants to record their daily activities in terms of sleep, personal time, time spent commuting to and from work, work time, limbo time, and periods of interim release. Study participants also provided self-assessments of the quality of their sleep and their level of alertness at the start and end of each work period. This study used a 14-day data-collection period to accommodate those train employees who did not work a regular schedule.

Researchers drew a random sample of 1275 train employees on commuter and intercity passenger railroads. The size of the sample from each of the two unions was proportional to that organization's representation in the total number of eligible participants. Retirees, full-time union officials, and anyone currently holding a railroad management position were not eligible for the study. Determination of the sample size assumed a 95-percent confidence interval on the estimates for mean sleep time, an error tolerance of 15 percent, and a 33-percent response rate.

Mailing of the survey materials occurred on December 31, 2009. Ten days later, every potential survey participant received a postcard, signed by his or her union president, to encourage the employee to participate in the survey. Three weeks after distribution of the materials, a second postcard thanked those who had decided to participate and encouraged those who had not yet done so to participate.

The overall response rate for the survey was 21 percent. Of the 269 complete responses, 13 could not be part of the analysis because either there were problems with the respondents' log books, or the respondents were not in crafts covered by the survey. (It was not possible to identify these individuals from the information contained in union membership databases.) The nonresponse-bias study based on age found no difference between survey respondents and nonrespondents.

The results of the study support the approach that FRA has taken in this rule. For instance, the results are consistent with the separate analysis during the development of this rule of schedules provided by commuter and intercity passenger railroads, indicating that a fairly small percentage of employee work time (about 1.8 percent) violates the fatigue threshold. The rule focuses additional attention and effort specifically on those schedules presenting this fatigue risk by requiring the mitigation of that risk, while schedules not at risk for fatigue would not be subject to these additional requirements.

In addition, when compared to the results of the previous study that primarily considered train employees on freight railroads, the results of the study of train employees on commuter and intercity passenger railroads support a significantly different approach. Train employees on freight railroads were found to experience some level of fatigue (equivalent to an effectiveness score <90 using the FAST model) during 73 percent of their work time, while train employees on commuter and intercity passenger railroads experienced this level of fatigue during only 14 percent of their work time. The substantive limitations imposed on train employees on freight railroads in the RSIA would largely be unnecessary for the commuter and intercity passenger railroad industry, as well as ineffective to target the specific areas where there is a fatigue risk.

IV. Railroad Safety Advisory Committee (RSAC) Process

A. Overview of the RSAC

In March 1996, FRA established RSAC,

which provides a forum for developing consensus recommendations to FRA's Administrator on rulemakings and other safety program issues. The Committee includes representation from all of the agency's major stakeholder groups, including railroads, labor organizations, suppliers, and manufacturers, and other interested parties. A list of member groups follows:

For more information about RSAC activities, see

http://rsac.fra.dot.gov/

. Meetings of the full RSAC are also announced by publication in the

Federal Register

• American Association of Private Railroad Car Owners (AARPCO);

• American Association of State Highway and Transportation Officials (AASHTO);

• American Chemistry Council;

• American Petroleum Institute;

• APTA;

• American Short Line and Regional Railroad Association (ASLRRA);

• American Train Dispatchers' Association (ATDA);

• AAR;

• Association of Railway Museums;

• Association of State Rail Safety Managers (ASRSM);

• BLET;

• Brotherhood of Maintenance of Way Employees Division (BMWED);

• Brotherhood of Railroad Signalmen (BRS);

• The Chlorine Institute;

• FRA;

• Federal Transit Administration (FTA);*

• Fertilizer Institute;

• High Speed Ground Transportation Association (HSGTA);

• Institute of Makers of Explosives;

• International Association of Machinists and Aerospace Workers;

• International Brotherhood of Electrical Workers (IBEW);

• Labor Council for Latin American Advancement;*

• League of Railway Industry Women;*

• National Association of Railroad Passengers (NARP);

• National Association of Railway Business Women;*

• National Conference of Firemen & Oilers;

• National Railroad Construction and Maintenance Association (NRC);

• National Railroad Passenger Corporation (Amtrak);

• National Transportation Safety Board (NTSB);*

• Railway Supply Institute (RSI);

• Safe Travel America (STA);

• Secretaria de Comunicaciones y Transporte;*

• Sheet Metal Workers International Association (SMWIA);

• Tourist Railway Association, Inc.;

• Transport Canada;*

• Transport Workers Union of America (TWU);

• Transportation Communications International Union/BRC (TCIU/BRC);

• Transportation Security Administration (TSA);* and

• UTU.

*Indicates associate, non-voting membership.

When appropriate, FRA assigns a task to RSAC, and after consideration and debate, RSAC may accept or reject the task. If the task is accepted, RSAC establishes a working group that possesses the appropriate expertise and representation of interests to develop recommendations to FRA for action on the task. These recommendations are developed by consensus. A working group may establish one or more task forces to develop facts and options on a particular aspect of a given task. The individual task force then provides that information to the working group for consideration. When a working group comes to unanimous consensus on recommendations for action, the package is presented to the full RSAC for a vote. If the proposal is accepted by a simple majority of RSAC, the proposal is formally recommended to FRA. FRA then determines what action to take on the recommendation. Because FRA staff plays an active role at the working group level in discussing the issues and options and in drafting the language of the consensus proposal, FRA is often favorably inclined toward RSAC recommendations. However, FRA is in no way bound to follow the recommendation, and the agency exercises its independent judgment on whether the recommended rule achieves the agency's regulatory goal, is soundly supported, and is in accordance with policy and legal requirements. Often, FRA varies in some respects from the RSAC recommendation in developing the actual regulatory proposal or final rule. Any such variations would be noted and explained in the rulemaking document issued by FRA. However, to the maximum extent practicable, FRA utilizes RSAC to provide consensus recommendations with respect to both proposed and final agency action. If RSAC is unable to reach consensus on a recommendation for action, the task is withdrawn and FRA determines the best course of action. If the working group or RSAC is unable to reach consensus on a recommendation for action, FRA moves ahead to resolve the issue through traditional rulemaking proceedings.

B. RSAC Proceedings in This Rulemaking

FRA proposed Task No. 08-06 to the RSAC on April 2, 2009. The RSAC accepted the task, and formed the Passenger Hours of Service Working Group (Working Group) for the purpose of developing implementing regulations for the hours of service of train employees of commuter and intercity passenger railroads under the RSIA.

The Working Group is comprised of members from the following organizations:

• AASHTO;

• Amtrak;

• APTA;

• ASLRRA;

• ATDA;

• AAR, including members from BNSF Railway Company (BNSF), Canadian National Railway Company (CN), Canadian Pacific Railway, Limited (CP), CSX Transportation, Inc. (CSXT), Iowa Interstate Railroad, Ltd. (IAIS), Kansas City Southern (KCS) railroads, Metra Electric District, Norfolk Southern Corporation (NS) railroads, and Union Pacific Railroad Company (UP);

• BLET;

• BRS;

• FRA;

• FTA;

• IBEW;

• Long Island Rail Road (LIRR);

• Metro-North Commuter Railroad Company (Metro-North);

• National Association of Railroad Passengers (NARP);

• National Railroad Construction and Maintenance Association;

• National Transportation Safety Board (NTSB);

• Southeastern Pennsylvania Transportation Authority (SEPTA);

• Tourist Railway Association; and

• UTU.

The Working Group completed its work after six meetings and several conference calls. The first meeting of the Working Group took place on June 24, 2009, in Washington, DC. At that meeting the group heard several presentations on fatigue science, including a report on the diary study that was to be conducted as described above. The group discussed the general approach for the rulemaking, and it was agreed that analysis of the railroads' work schedules would support the rulemaking. Subsequent meetings were held on February 3, 2010; March 4, 2010; April 6, 2010; May 20, 2010; and June 29, 2010. In addition, a Task Force was formed that met on January 14-15, 2010, March 30-31, 2010, and April 28-29, 2010.

At the conclusion of the June 29, 2010 meeting, the Working Group voted to approve a draft of the proposed rule text, with the exception of two sections, to which the group had suggested numerous edits. It was agreed that FRA would address the remaining issues in those sections and circulate a revised draft, on which the group would vote electronically. After the revised draft was produced, the Task Force had several conference calls to discuss the revised provisions, and FRA also participated in several calls with task force members. Ultimately, on September 22, 2010, the Working Group voted unanimously to agree to the rule text presented in the proposed rule. The group's recommendation was presented to the full RSAC on September 23, 2010. The full RSAC agreed to vote electronically on the proposed rule text recommended by the Working Group, and ultimately accepted its recommendation. Although only a majority was required, the vote was unanimous.

The rule text voted on by the full RSAC and recommended to FRA is available on the RSAC Web site.

Following the vote of the Working Group and the full RSAC, FRA recognized the need to make two changes to the recordkeeping and reporting requirements in 49 CFR 228.11 and 228.19, to accommodate a new substantive limitation contained in the proposed rule as approved by the RSAC. While the RSAC voted in favor of the proposed substantive requirements in question, and all other elements of the proposed rule, the corresponding amendments to the recordkeeping and reporting provisions were not presented to them. After publication of the proposed rule on March 22, 2011, and consideration of public comments, FRA has made additional changes, as discussed in Section V of the preamble, below.

Earlier, at the February 3, 2010, meeting, FRA presented an initial draft of the rule text, identifying the basic concepts and direction of the rulemaking. Based on discussions at that meeting, a more complete draft was presented at the March 4, 2010 meeting, and the text was refined and supplemented at subsequent meetings. In addition, during the course of the

Working Group and Task Force meetings, a number of significant issues were discussed that resulted in changes in the rule text or common understanding of the intent of specific provisions that should be explained. Some such issues will be explained in this section, while other subjects of discussion by the Working Group and the Task Force will be discussed in the Section-by-Section Analysis at Section VI of the preamble.

In addition, as discussed below in the Regulatory Impact and Notices section of the preamble, Section VII, FRA has considered the costs and benefits of this rule. Implementation costs would be associated with analyzing work schedules, training, and rest facilities. However, relative to the “no regulatory action” alternative in which passenger railroad train employees would become subject to the new HSL in effect for freight train employees, the rule would result in a cost savings of $57.7 million (discounted at 7 percent) and $65.2 million (discounted at 3 percent) over a 20-year period. The quantified accident reduction benefits achieved under both the “no regulatory action” baseline and the rule total $1.2 million (undiscounted), $0.6 million (PV, 7 percent), and $0.9 million (PV, 3 percent). FRA does not expect that the overall number of casualties and property damages prevented will differ under either scenario. Implementation of the final rule will yield these benefits at lower cost. While the rule has lower monetized benefits than costs, when compared to the current HSL, FRA believes that there are unquantified benefits that could close the gap.

C. Significant Task Force Contributions to the Development of the NPRM

As was noted above, the Working Group created the Task Force, comprised of representatives from BLET, UTU, APTA, AAR, and FRA. The Task Force met between Working Group meetings to provide additional input and advice to the Working Group on the approach to the rule, specific concerns as to the rule text, and implementation of the regulatory requirements. Although the Task Force was extremely helpful throughout the development of the proposed rule in offering suggestions as to the rule text, its primary contributions were in the areas of schedule analysis and the creation of a fatigue mitigation tool box.

1. Schedule Analysis

The diary study discussed in Section III.B of the preamble provided valuable evidence of the actual levels of fatigue experienced by train employees on commuter and intercity passenger railroads. However, since many of these employees work scheduled assignments, it was also valuable to evaluate the schedules themselves, to get a sense of the parameters of those assignments that would result in fatigue violating the threshold, which informed some of the provisions of this rule. The Task Force assisted the Working Group by evaluating the schedules and presenting their results to the Working Group.

APTA hired a consultant to analyze the schedules provided by the railroads that were worked by their train employees. The railroads provided all of their schedules for the month of July 2009. The schedules were analyzed using the FAST model, including conservative assumptions about the sleep that would be obtained by an employee working that schedule. For example, the analyses assumed that employees did not sleep during periods of interim release.

The analyses that the Task Force presented to the Working Group demonstrated that most schedules did not result in an employee's violating the fatigue threshold. This was true even for schedules in which the employee reported for duty at 4 a.m. and was relieved from duty at 8 p.m., for a 16-hour duty tour that included a total of 12 hours on duty and a 4-hour interim release. Most of the problematic schedules identified through the analysis presented by the Task Force involved duty tours in which some time was spent working during late night hours. These analyses formed the parameters for FRA's definitions of “Type 1 assignment” and “Type 2 assignment” for which different requirements would apply in this rule.

2. Fatigue Mitigation Tool Box

Because a major aspect of this rule requires mitigation of the fatigue risks identified in those schedules that resulted in an employee's violating the applicable fatigue threshold, and experiencing a level of fatigue at which safety may be compromised, the Task Force assisted the Working Group by developing a fatigue mitigation tool box, a document that would illustrate the variety of ways in which a railroad might seek to address the fatigue risks in its schedules. (A copy of this document has been placed in the docket for this rulemaking.) The tool box itself is not intended to become a part of the regulatory text. Instead, it is intended to provide the variety of methods from which a railroad may propose, in its plans submitted to FRA for approval, to mitigate identified fatigue risks in its work schedules, to bring them into compliance with the regulation. It is expected that not every tool will be appropriate for each railroad, or for individual locations or schedules on a given railroad, and that the railroads, in consultation with their labor organizations, will choose the mitigation tools most appropriate to each circumstance, subject to FRA review and approval. In addition, the tool box is expected to be a living document, as the available fatigue mitigation tools will change over time as fatigue science continues to develop, or as railroad operations change, either generally or as related to specific properties or schedules. The tool box as a whole will not be approved by FRA, nor will it be maintained by FRA as it evolves. FRA will evaluate the appropriateness of specific fatigue mitigation tools as they are submitted to FRA as part of a railroad's plan to mitigate fatigue risks associated with particular schedules.

This section will describe a representative sample of the variety of the tools included in the tool box developed by the Task Force, which may be applied to mitigate fatigue risk. This discussion is not intended to provide an all-inclusive list of the possible fatigue mitigation tools. A railroad is free to use any fatigue mitigation tool that it believes is effective in reducing the fatigue risk found in its schedules, subject to FRA's review and approval when the tools are applied to mitigate fatigue in a particular work schedule.

Perhaps the easiest mitigation tool to understand that was identified by the Task Force is the adoption and implementation of a napping policy, and the provision of facilities for employees to take a nap during interim releases or other periods between assignments that may be available for rest during a duty tour. The addition of a period of sleep to the employee's schedule would have a clear impact on the employee's level of fatigue when working that schedule, and the level of fatigue that the employee would be expected to experience throughout the remainder of the duty tour after a nap, which might reduce the risk of fatigue sufficiently to bring the schedule and the employee's effectiveness score within the fatigue threshold.

To use this tool to mitigate fatigue, a railroad would be required to identify, in consultation with its labor organizations or employees, the facilities that would be available for the purpose of rest during the duty tour, that are appropriate to the schedule and location at issue. This would not always require a bunk or a quiet room, though

this might be available at some locations and in certain situations. However, the period available for rest would have to be at least 90 minutes for this mitigation tool to be applied, as this amount of time would provide sufficient opportunity for an employee to get to his or her napping location and fall asleep, having enough time for a nap of sufficient duration to be beneficial to the employee's level of fatigue, and then also allowing the employee time to become fully awake and ready to resume the duty tour.

Another mitigation tool, applicable to railroads and locations using employees from an extra board, would be the use of multiple extra boards that are temporally separated, so that employees would be scheduled to work morning assignments or evening assignments, rather than being subject to calls for assignments at all times of day. For example, employees assigned to a morning extra board might be subject to being called only for assignments requiring them to report for duty between 4 a.m. and 10 a.m., while employees assigned to an evening extra board might be subject to being called only for assignments requiring them to report for duty between 4 p.m. and 10 p.m. Employees on either extra board would know that they would not be called for an assignment requiring them to report for duty outside the times established for the employee's particular assigned extra board. This would lead to greater predictability of schedule and ability to plan rest, while also avoiding (1) circadian shifts between duty tours resulting from changes in the time of day that the employee is awake and (2) difficulties in adjusting to changing periods available for sleep.

Call windows (

i.e.

, limited periods of time during which an employee is subject to receiving calls from the railroad to report for duty) are another mitigation tool in the tool box, which may be combined with a temporally separated extra board, but could also be used even if the extra board were not so divided. For example, a railroad might decide to establish a call window that would reduce or eliminate calls to the employee during the time from 11 p.m. and 5 a.m. Open assignments that would need to be filled from an extra board of employees who would otherwise be called for the assignment during that time would instead be filled before 11 p.m., which would give the employees greater predictability and ability to plan rest, as well as allowing them more rest during the late night hours.

Another possible tool would be to allow employees a period of uninterrupted rest, similar to the requirement that applies to train employees on freight railroads, which is found at 49 U.S.C. 21103(e). The uninterrupted rest could be applied to an employee's statutory off-duty period before or after the employee is to work a schedule violating the fatigue threshold. It could also be applied to periods of interim release within the duty tour.

Education could also be part of the tools that a railroad will use to mitigate fatigue in certain circumstances, and is also a key component of the other mitigation tools. The mitigation tools will not be beneficial if the employees working the schedules to which they are applied do not understand the available tools, and how to properly use them to reduce their fatigue and increase their effectiveness. If employees do not take advantage of the mitigation tools, and use them properly to increase their rest, even those mitigation tools most likely to have the greatest and most tangible impact on reducing fatigue will not have the desired effect. FRA has also recognized the importance of education as a component of fatigue management by specifically requiring in this rule that employees and supervisors receive training on fatigue and strategies for reducing it.

Finally, one additional mitigation tool was discussed by the Task Force that was extremely well-received and supported by the Working Group, including FRA representatives. That suggestion was to develop software that would link the railroad's crew management resources to both the employee's electronic hours of service records (created and maintained in compliance with subpart D of 49 CFR part 228), and a valid biomathematical model of performance and fatigue.

The idea is that the fatigue model would be able to look back at previous duty tours and rest periods to determine which schedules might have sufficiently rested employees available to report for the assignment, not only under the limitations on time on duty and consecutive days and the requirement for minimum time off duty established by this rule, but also in terms of the fatigue threshold. The model would have the benefit of the data from the previous duty tours to take into account in determining whether these schedules would violate the fatigue threshold during the duty tour, as well as at the report-for-duty time. If the analysis revealed that the employees on these schedules would be too fatigued to report for the assignment, or would violate the fatigue threshold during the duty tour, crew management would be alerted that these employees could be at risk if they work this particular assignment. Employees would have to affirm their fitness for duty if asked to work such assignments and be empowered to reject the assignments, because the model is being used to predict group (average) fatigue from work schedules that could be worked by several individuals. Any individual could be more or less fatigued than the average or group. Employees have a responsibility to indicate if they feel fit to work or not, regardless of the effectiveness score that a model would predict. The employer's responsibility is to arrange schedules that minimize fatigue.

While all of the parties to the Working Group agreed that this idea showed great promise as an effective fatigue mitigation tool for the future, it is not something that the railroads will be able to apply immediately, for technological reasons. Most railroads subject to this rule do not yet create and maintain their hours of service records electronically in compliance with subpart D, although there is interest among those railroads in developing hours of service electronic recordkeeping programs. In addition, software would need to be developed that would allow the fatigue model to retrieve data from the electronic recordkeeping system, without any possibility of altering or otherwise affecting the integrity of the records maintained in the system. Likewise, software would be needed to connect the fatigue model to the crew management system, so that it could appropriately alert that system and prevent an employee being placed on an assignment for which he or she would be too fatigued. If the necessary systems and software can be developed, compliance with the fatigue threshold would become much easier, and there would be much less excessive fatigue to be mitigated.

Areas of Working Group and Task Force Concern During Development of the NPRM

During the course of the Task Force and Working Group meetings, a few issues resulted in significant discussion. Some issues were related to specific provisions in the rule text, while other concerns were about the broader implications of the rule, as well as its effects on aspects of railroad operations or existing collective bargaining agreements.

1. Proposed Definitions of “Type 1 Assignment” and “Type 2 Assignment”

Some members of the Working Group suggested that there should be a way to

determine a template for schedules that would be deemed not to violate the fatigue threshold. As was discussed above, the Task Force presented schedule analyses showing that a schedule in which an employee began work at 4 a.m. and was relieved at 8 p.m., resulting in a duty tour with a total time on duty of 12 hours, with a 4-hour period of interim release, did not violate the fatigue threshold.

Based on this analysis, FRA initially defined any assignment beginning no earlier than 4 a.m. and ending no later than 8 p.m., assuming at least a 4-hour period of interim release, as a Type 1 assignment, which would be deemed not to violate the fatigue threshold. Assignments that included any period of time outside the defined time parameters of a Type 1 assignment would be considered a Type 2 assignment, which would be subject to more stringent requirements, including analysis of the schedule using a scientifically valid biomathematical model, and a more restrictive limit on the number of consecutive days on which an employee working such an assignment would be allowed to initiate an on-duty period.

However, some Task Force members pointed out that there could be assignments that include time outside the time parameters of a Type 1 assignment that would not violate the fatigue threshold. In some cases these schedules would only have a small amount of their overall time outside of the Type 1 parameters. For example, an assignment might begin at 4:30 a.m. and end at 8:30 p.m. In addition, some assignments might not violate the threshold because of the short duration of the duty tour involved, such as, perhaps, an assignment from 5 p.m. until 9:30 p.m.

Based on these considerations, FRA amended the definition of “Type 2 assignment” to indicate that if an assignment does not include any time between midnight and 4 a.m., then the particular time of day or night that an assignment is to be performed is not the only determinant of whether an assignment is considered a Type 2 assignment. In particular, a Type 2 assignment that is analyzed using a scientifically valid biomathematical model and is determined not to violate the fatigue threshold, and that includes no period of time between midnight and 4 a.m., would be considered a Type 1 assignment.

FRA also added language to the definitions of both “Type 1 assignment” and “Type 2 assignment” to require compliance with the substantive limitations contained in § 228.405. FRA expects that railroads would not be operating schedules that violate these limitations; most schedules have long been in effect for the railroads subject to this rule, and this was an implicit assumption of the Working Group. For example, a schedule that requires an employee to report for duty at 4 a.m. and to be released from duty at 8 p.m. would have to include a period of interim release of at least 4 hours that is not time on duty, as defined by § 228.405(b). However, this language was added to the definitions to make clear that the schedule analysis and fatigue mitigation requirements of this rule supplement, but do not replace, the specific limitations, and any schedule that violated other provisions of this rule (for example, exceeded 12 hours total time on duty, or did not allow for at least 8 hours off duty, or 10 hours off duty after 12 consecutive hours) could not be deemed “approved” by FRA and subject to the less stringent requirements applicable to Type 1 assignments.

2. Proposed Limitations on Number of Consecutive Days

In the Working Group, both the railroads and labor contended that FAST and/or FAID analysis would suggest that an employee could work beyond the limitations in what became the proposed rule and this final rule without adversely affecting safety. One potential requirement about which this was specifically argued was the limitation on the number of consecutive days or days within a prescribed period that an employee would be permitted to initiate an on-duty period before the employee was required to have a 24-hour or two-consecutive calendar days off-duty period at the employee's home terminal under this regulation, which would differ depending on the time of day that the employee works. See § 228.405(a)(3) and (a)(4) of the proposed rule, and § 228.405(a)(3) of this final rule. In the Working Group, the railroads and labor unions presented fatigue analyses for theoretical schedules that would have an employee initiating on-duty periods for numbers of days that exceeded those permitted by the contemplated rule. The railroads and labor also indicated that the current agreements or practices on their properties allow for such schedules.

Research shows that work on successive days without a full day off exponentially increases the accident risk as the number of days worked increases. For instance, after working four consecutive day shifts, there is a 17-percent increase in risk, and after working four consecutive night shifts, there is a 36-percent increase in risk.

FRA research on train crew work schedules and sleep patterns

has shown that train crews average a 10.25-hour day (work period, limbo time, and commute time) and get 6.88 hours of primary sleep per day. A follow-up study on passenger train crews found that workers on split shift assignments average a 13.75-hour day (work period, interim release, and commute time) and get 6.18 hours of primary sleep. Laboratory studies of restricted sleep

show a 5-percent decrease in performance after 7 days with 7 hours of sleep per day and a 15-percent decrease after 7 days with 5 hours of sleep per day. These studies are consistent with the previously noted increase in accident risk with the number of days worked.

Folkard, S. and Akerstedt, T.,

Trends in the Risk of Accidents and Injuries and Their Implications for Models of Fatigue and Performance,

Aviat. Space Environ. Med. (2004).

Gertler, J., and DiFiore, A. (2009). Work schedules and sleep patterns of railroad train and engine service workers (Report No. DOT/FRA/ORD-09/22). Washington, DC: U.S. Department of Transportation.

Balkin, T., Thorne, D., Sing, H. (2000).

Effects of sleep schedules on commercial driver performance

(Report No. DOT-MC-00-133). Washington, DC: U.S. Department of Transportation.

Therefore, FRA reasoned that, even if an employee were working a schedule for which the employee's effectiveness score did not violate the fatigue threshold, even when the schedule was worked for more consecutive days or days in a 14-day period than the regulation would permit, at some point the employee would have to use some of the time between duty tours (time that a model would otherwise view as available for rest) to attend to other personal activities. This time spent in activities other than rest would decrease the time actually available to the employee for rest, and, therefore, the employee's actual effectiveness score. This circumstance would be particularly problematic for schedules featuring long duty tours, such as the maximum 12 hours on duty, including an interim release, for a total time of 16 hours in the duty tour, followed by the minimum of 8 consecutive hours off duty before reporting for the next duty tour. From this perspective, FRA believes that, although the available research does not identify the exact number of consecutive days or days in a prescribed period allowed under this rule as the maximum that can be safely worked, the limitations that FRA has established are reasonable.

FRA remains aware that the requirements of the final rule may have

an impact on the collective bargaining agreements affecting the railroads and employees covered by subpart F. For example, there may be some agreements that would allow employees to work a greater number of consecutive days or days in a 14-day period than would be allowed by this regulation. FRA also remains mindful that the law provides an option that enables the regulated community to seek waivers to implement pilot projects in accordance with the requirements of 49 U.S.C. 21108(a) and encourages members of the regulated community to consider this option. Pursuant to 49 CFR part 211, subpart C, the Railroad Safety Board will consider whether or not granting such waivers would be in the public interest and consistent with railroad safety. Where warranted, and upon the necessary showing, FRA may grant waivers of the requirements of this rule, including requirements concerning the maximum number of consecutive days or days in a 14-day period that an employee may work, to allow for the establishment of pilot projects to demonstrate the possible benefits of implementing alternatives to the strict application of the requirements contained in this rule.

3. Precision of Fatigue Models and Threshold

There was considerable discussion in the Working Group of the precision embodied in the FAST model or the FAID model, and the appropriateness of requiring compliance with a specific fatigue threshold. The railroads argued that models such as the FAST model and the FAID model are not scientifically precise enough to warrant the adoption of a specific threshold, and that different types of operations could safely function at different levels of fatigue. For example, the railroads contended that yard switching activities could safely operate at a different level of fatigue than passenger operations or through-freight activities.

The railroads conceded, however, that the regulatory structure contained in the proposed regulation, and in provisions of the final rule that mirror the proposal would not be problematic for passenger operations. The railroads' concern was that, in the future, someone might argue for adoption of the same regulatory structure for freight operations and, were that to occur, schedules might be prohibited from use that should, in fact, be acceptable from a fatigue perspective.

In FRA's view, a specific threshold is desirable because it provides regulatory certainty as to what railroads must do to be considered in compliance with the regulations. FRA has based its regulation on the best available fatigue science, including the FAST model and the FAID model, which are the only currently validated models, and the appropriate fatigue thresholds for the purpose of compliance with this regulation. As was discussed in Section III above, FRA has adjusted the FAID threshold from the level stated in the preamble of the proposed rule, to achieve a closer correlation between the FAST and FAID thresholds for the purposes of the analyses required by this regulation. FRA has also left open the possibility that other models may be validated, and other thresholds established in the future, which could be used for the purpose of compliance with this regulation.

In addition, as new scientific evidence comes to light, FRA will review this rule as discussed in Section III, above.

See Raslear,

supra

note 11.

As FRA has determined that use of these models and their established thresholds adequately protects safety, that this rule does not present significant implementation problems for passenger service, and that a specific threshold provides the desired regulatory certainty, FRA believes that it is appropriate to include in the regulations a requirement for a specific threshold. FRA based this belief on the understanding that the regulatory requirements will be satisfied based on a “70/20 threshold” using the FAST model (meaning that the fatigue threshold is violated if an employee's effectiveness score is less than 70 for 20 percent or more of the employee's time on duty,) or a “72/20 threshold” using FAID (meaning that the fatigue threshold is violated if an employee's effectiveness score is more than 72 for 20 percent or more of the employee's time on duty.)

See Hursh,

et al., supra

note 13, and Tabak and Raslear,

supra

note 19.

In establishing a substantive hours of service regulation with a specific threshold for train employees in passenger service, FRA is not drawing any conclusion about the suitability of such a regulatory scheme for freight operations. There may be substantial differences between freight railroad operating and crew schedules and passenger operating and crew schedules. Passenger railroads have analyzed the results of applying the regulations to their work schedules and concluded that this regulation is feasible. Freight railroads have not undertaken such analysis, nor would they be required to under the regulations, except to the extent that employees of freight railroads may work in passenger service.

4. Freight Railroad Employees Acting as Pilots for Commuter or Intercity Passenger Trains

The Working Group also discussed the application of requirements of proposed subpart F, which have now been adopted, to train employees of freight railroads who occasionally provide pilot service to a commuter railroad or intercity passenger railroad. FRA's locomotive engineer certification regulations require a pilot to assist an engineer who may not be sufficiently familiar with the territory over which he or she is called to operate. See 49 CFR 240.231(b). The railroads indicated that a request for a pilot may come without advance notice, so that it would be difficult to comply with the substantive hours of service limitations and recordkeeping requirements of this regulation, and even more difficult to adhere to the schedule analysis requirements, for an employee who did not otherwise regularly engage in commuter or intercity rail passenger transportation.

The Working Group also cited the safety benefits of having a pilot available on a route when necessary, and the potential risk if commuter or intercity passenger railroads were to become less likely to request a pilot, or freight railroads less likely to be able to make a pilot available when requested, because of concerns about the requirements of this regulation, which has been adopted. FRA acknowledges these benefits. Therefore, although a pilot is performing covered service under the HSL on the assignment on which the pilot service is provided, FRA will not consider a train employee employed by a freight railroad who serves as a pilot on a train operated by a commuter railroad or intercity passenger railroad to be a train employee who is engaged in commuter or intercity rail passenger transportation.

V. Response to Public Comments on the NPRM

FRA received 10 sets of comments on the proposed rule. Comments were received from the National Institute for Occupational Safety and Health (NIOSH); the American Academy of Sleep Medicine (AASM); Port Authority Trans-Hudson (PATH); Metropolitan Transportation Authority (MTA); SEPTA; Strasburg Rail Road Company (Strasburg); Transportation Trades Department (TTD), AFL-CIO (American Federation of Labor and Congress of Industrial Organizations); BLET and

UTU, which filed joint comments; AAR and APTA. Issues raised in the comments will be addressed in this section. Some issues arising out of the comments were also discussed in Section III, Scientific Background, and some will be further discussed in Section VI, Section-by-Section Analysis, below.

Comments Related to the FAST and FAID Fatigue Models

AAR and APTA indicate in their comments that their analysis shows that passenger train employees' work schedules that are acceptable when analyzed using FAST with a proposed fatigue threshold of 70, violate a proposed FAID fatigue threshold of 60. Consequently, MTA, SEPTA, AAR and APTA, each recommend using a FAID threshold of 90, rather than the threshold of 60 proposed in the NPRM. AAR and APTA each attach to their comments, an analysis performed by the same consultant who performed work schedule analysis for APTA during the development of the proposed rule, in support of their request. MTA, SEPTA, AAR and APTA also contend that FRA agreed with a threshold of 90 for FAID during the Working Group, prior to FAID's validation. FRA disagrees both with a FAID threshold of 90 and with the analysis submitted in support of it.

FRA did not agree during the Working Group process that 90 was the appropriate threshold for FAID, and indeed recalls little, if any, discussion of a FAID threshold, as FAID had not been validated or calibrated at that time. It is possible that the railroads internally discussed a threshold of 90, as some railroads had been using FAID for the purposes of their own analysis even before the commencement of this rulemaking.

The analysis attached to the AAR and APTA comments looked at 101 work schedules from “some of the largest railroads” involved in passenger service. It is not clear why that number of schedules was chosen, nor why the specific schedules were chosen for analysis. This suggests that the 101 work schedules are a convenience sample, rather than a random sample of work schedules, which means that these schedules may not be representative of the rail passenger service industry. In addition, the analysis looked at work schedules alone, rather than both work schedules and on-duty accidents in which those working the schedules were involved, as had the FAST and FAID validation studies. The threshold that FRA is seeking is the point at which the risk of a human factors accident involving the person working the schedule increases. That is the point, for the purpose of this regulation, at which “safety may be compromised” and the rule requires action to be taken to mitigate fatigue. See § 228.407(a). Looking at work schedule data only, the analysis provided by AAR and APTA has not identified that point. The analysis that they provided uses statistics, rather than fatigue science, to equate a FAST score of 70 with a FAID score of 90, based on where the effectiveness scores produced in the analyzed schedules were clustered. In validating and calibrating FAID, FRA used bins to analyze the data in light of the variation among FAID scores. Biomathematical models such as FAID are more accurate when used to predict population behavior rather than individual behavior, and the goal is establishing a fatigue threshold rather than establishing links between all FAST scores and FAID scores at an individual level. Accordingly, FRA does not believe that the statistical comparison of individual scores is an appropriate basis for establishing a FAID threshold for the purposes of this rulemaking.

FRA recognizes the concern with schedules that are acceptable using one model violating the threshold using another. In Section III, Scientific Background, FRA explained its basis for modifying the FAID threshold, not to 90, as urged by the railroads, but to 72. This change is achieved by basing the FAID threshold on the upper limit of the 99-percent confidence interval rather than the mean. A 99-percent confidence interval for a FAID threshold of 72 means that there is only a one-percent chance of a false positive (

i.e.,

a schedule that will violate the FAID threshold of 72 while not actually posing a risk for the level of fatigue indicated by the threshold). A confidence interval for the FAID threshold is appropriate, since it is calibrated in relation to FAST.

Finally, APTA suggests that FRA commit to further analysis, including analysis specifically of passenger data, which could form the basis for establishing a FAID threshold other than 90. As noted above, FRA does not believe that 90 is a scientifically valid fatigue threshold for FAID. In terms of APTA's recommendation that FRA agree to do further analysis, FRA is certainly willing to acknowledge that the area of fatigue science is still developing and that future developments or analyses may make it appropriate to revisit the models, their thresholds, or other aspects of this rulemaking, as discussed in Section III.

Comments Related to Costs of Compliance With the Proposed Rule

NIOSH questions whether the training costs included in the NPRM included costs to train staff on the use of the models. In the proposed rule, the cost of training staff to use the models was included in the cost of the biomathematical model, which also includes programming (for product enhancement) and technical support, and remains included in the model cost of the final rule. For purposes of clarification, FRA is presenting training related to the models separately.

APTA indicates that the licensing cost for FAST is approximately $500,000 for a single railroad, which is far in excess of the cost estimated by FRA at the NPRM stage, and that the licensing cost for FAID is about five percent of the cost of FAST, or $25,000. FRA clarifies that its cost estimate was used for conduct of the regulatory analysis and as such includes only the cost to “society,” which does not include distributional effects that may arise through transfer payments including the revenue collected through a fee, surcharge in excess of the cost of services provided. “Transfer payments are monetary payments from one group to another that do not affect total resources available to society.” OMB Circular A-4, p. 38.

Thus, the FRA cost estimate included some programming costs for the development of certain enhancements tailored to the passenger rail industry that included the license cost, training on use of the model, and system support. FRA did not include costs associated with the original model development or economic rent from the sale of licenses to passenger railroads. Administrative costs associated with using the model to analyze assignments for purposes of complying with this rule are included in the FRA cost estimate separately. The development costs of the models themselves are considered “sunk costs” incurred prior to the rulemaking and not attributable to this rule.

OMB Circular A-4 is available at:

http://www.whitehouse.gov/sites/default/files/omb/assets/omb/circulars/a004/a-4.pdf.

In addition, FRA assumed that railroads would select the lowest cost alternative for achieving compliance. FRA recognizes other factors may contribute to model selection. While FRA did not and does not endorse any particular model or method for use in complying with this rule, and railroads are certainly permitted to use more costly alternatives, for purposes of conducting regulatory analysis, only the

“opportunity cost”

is included. Any additional expense, however, would not be a cost attributable to this rule. APTA did not provide a basis for its cost estimate of $500,000 per railroad for the FAST model, and based on information available to FRA, a cost of $500,000 does not reflect the opportunity cost to society.

“The opportunity cost is equal to the net benefit the resource would have provided in absence of the requirement.” OMB Circular A-4, p. 19.

In this case the opportunity cost includes the programming and licensing cost estimated at $75,000, the training cost estimated at $50,000, and product support associated with analyzing assignments for purposes of complying with this rule estimated at $7,500 annually. As noted at the NPRM stage, FRA believes that a significantly lower-cost viable alternative for compliance would be for the railroads to enter into a cost sharing agreement via a trade organization, such as APTA and the Association of Railway Museums (ARM), to facilitate so that one or few licenses are purchased for the use of all member railroads.

On a related note, MTA points out that early in the Working Group process, as the NPRM was being developed, FRA indicated a willingness to explore funding access to the models. Unfortunately, FRA is not in a position to fund access to the models, but, as discussed above, FRA encourages relevant organizations to work together, as there may be ways to provide the model for a group of members that are more cost effective than for each member railroad to secure access individually.

APTA also contends that the cost of fatigue training will exceed $1.8 million for a sample of 5 commuter railroads subject to this regulation. APTA does not provide any background or details related to this stated cost, and it is not consistent with information provided to FRA during the development of the proposed rule. However, it is possible that these costs are based on providing formal, classroom training to all of the employees to be covered by this regulation. As was explained in the NPRM, FRA incorporated significant flexibility into the training requirement, so that each railroad would be allowed to tailor the level of complexity and formality to the needs of its employees. There are likely railroads, or locations on a particular railroad, where the nature of the operations and assignments do not warrant formal classroom training and such training would not be practical or cost-effective. In many cases, there will be lower cost alternatives that will be more appropriate and sufficient to comply with the training requirement.

APTA and MTA both claim costs related to the hiring of additional personnel. MTA says that it would have compliance costs of at least $5 million per year, including the cost of hiring additional train and engine employees. APTA contends that the cost of additional personnel will exceed $15 million for five sample commuter railroads, and $12 million for Amtrak. Neither MTA nor APTA provides any specific information regarding these costs, and FRA does not believe that additional personnel will be required by the regulation. The rule provides substantial flexibility in how railroads may mitigate fatigue in their schedules. Many of the available fatigue mitigation tools, such as allowing employees to take a nap during available periods within a schedule, would significantly reduce fatigue without requiring the railroad to hire additional employees. In addition, should a railroad be unable to sufficiently mitigate the risk of fatigue in one of its schedules, it would also have the option of submitting a declaration of operational necessity to FRA for approval. See § 228.407(b)(1)(ii). Although there may be some circumstances in which a railroad would choose to hire additional employees, the regulation does not require extra hiring, especially not to the extent of the costs alleged by MTA and APTA. Finally, addition of new train crews to perform the same train operations would result in a decrease in the hours of service performed by existing train crews, which in turn would result in a savings that would in large part offset the expense associated with the hours of service performed by new employees and must be taken into account. In other words, it would basically take the same number of total employee hours to operate trains if the train schedules are unchanged regardless of how many train crews participate in the operation, leaving the total wage expense largely unchanged and only impacting the fixed overhead costs resulting from a larger employee pool. The Regulatory Impact Analysis contains a more detailed discussion of such impacts.

Some of the personnel costs described by MTA and APTA may be a result of concerns about the FAID threshold, proposed as 60 in the NPRM, which resulted in a greater number of schedules than expected violating the fatigue threshold. FRA responded to comments about the fatigue models above, and also addressed the issue in Section III of this preamble, Scientific Background. In light of the modifications made by FRA, the impact of this issue will be significantly reduced. In addition, as noted above, schedules violating the threshold do not require the hiring of additional personnel, as there are a variety of ways to mitigate the fatigue that would not require the expense of additional hiring.

PATH also indicates that it would need to hire additional engineers and conductors “to mitigate the effects of a mandatory 48 to 72 consecutive-hour rest period” the cost of which it estimates at $4 million annually. This comment appears to refer to the statutory requirements at 49 U.S.C. 21103(a)(4), which do not apply to train employees subject to this regulation. The requirements of this regulation are imposed instead of, rather than in addition to, the requirements for train employees in freight service. If, as PATH contends, its schedules will pass any fatigue analysis, its costs resulting from this regulation should be minimal.

Finally, AAR objects to the cost of having some employees subject to two different sets of hours of service requirements, referring specifically to those employees working from an extra board that includes both freight and passenger assignments. For this reason, AAR suggests that train employees employed by freight railroads should be governed only by the freight hours of service provisions in 49 U.S.C. 21103. This comment will be more fully discussed below, with comments related to the scope of the rulemaking. From a cost perspective, however, the cost of compliance with two separate hours of service schemes is not a new cost, as freight railroads have already had to track their train employees who perform both freight and passenger service under the different statutory provisions currently applicable to both, as freight and passenger train employees have had different requirements since the effective date of the RSIA. In addition, AAR admits that very few employees would be affected by being subject to both freight and passenger requirements, so any cost would likely be minimal.

Comments on the Scope of the Proposed Rule (§ 228.401 and § 228.403)

AASM suggests that FRA should develop an additional subpart to establish comparable language for train employees engaged in transportation services outside of commuter or intercity rail passenger transportation. As was described in the Section II, Statutory Background and History, prior to the RSIA, the Secretary had no

authority to issue regulations governing the hours of service of train employees. In the RSIA, Congress amended the then-existing statutory hours of service requirements for train employees, but specifically excluded train employees providing commuter or intercity rail passenger transportation from the application of those provisions for a period of three years, during which FRA, as the Secretary's delegate, was granted authority to promulgate hours of service regulations for these train employees. Other train employees remain subject to the hours of service statutory provisions as amended by the RSIA.

AAR and APTA both suggest that train employees on freight service extra boards who occasionally are called to operate passenger trains should be subject exclusively to the freight hours of service statutory requirements, rather than this final rule, and they suggest amending § 228.403 to exclude such employees from the requirements of this rule. FRA does not believe this exception would be consistent with the Congressional authorization, which is to establish hours of service regulations for train employees providing commuter or intercity rail passenger transportation. Congress recognized that the transportation of passengers has different characteristics that make the requirements established for freight operations inappropriate, and that regulations based on fatigue science would be more appropriate to passenger operations, regardless of the entity that employs the train employee providing this service. In addition, the railroads would have to track freight and passenger service separately for business purposes, to bill the commuter operator for the employee's time, even if the employees were just under the freight provisions. Finally, if the fact that an employee could be called on to perform freight service on an as-needed basis is enough to exclude them from the coverage of this rule, this could result in excluding employees who perform predominantly passenger service just for the possibility of their performing occasional freight service.

AAR also suggests that train employees of freight railroads who operate non-scheduled passenger service such as “Santa trains” or steam trains should not be subject to this regulation. AAR contends that these employees are “akin to employees operating work trains” who were specifically proposed for exclusion from the application of the proposed rule and who are specifically excluded from the application of this final rule by a definition in § 228.403(c). FRA disagrees with this analogy, as train employees operating “Santa trains” or steam trains are transporting passengers, while train employees operating work trains are not. In the NPRM, FRA stated its belief that Congress intended that these regulations apply to all railroads providing rail passenger transportation, and therefore included tourist, scenic, excursion and historic railroads within the scope of this regulation. FRA likewise believes it was the intent of Congress to cover operations such as those described by AAR that also involve rail passenger transportation.

AAR also suggests that FRA remove the limit on the number of times a month that train employees employed by a freight railroad who may provide pilot service for a locomotive engineer of a passenger railroad without being subject to the schedule analysis and other requirements of this regulation. AAR acknowledges that it would be unlikely that an employee would provide pilot service more than four times in a month, but says it should be permitted if necessary. FRA agrees with this suggestion for the reasons discussed above in Section IV.D.4, and has eliminated the cap on the provision of pilot service. FRA has also added the exclusion of freight train employees providing pilot service from the coverage of this rule to the rule text, in § 228.403(c), rather than just including it in the preamble, as was done in the NPRM.

APTA recommends that mechanical breakdowns, signal failures, switch failures and similar conditions should come within the non-application provision of § 228.403. FRA does not believe this is appropriate, as these common operational issues do not justify a complete exemption from the provisions of this regulation. This position is consistent with FRA's longstanding interpretation of the comparable statutory nonapplication provision at 49 U.S.C. 21102. See 49 CFR part 228, Appendix A. However, as will be discussed below, to the extent that such issues delay schedules the fatigue implications of which a railroad had previously analyzed and mitigated as appropriate, FRA will allow flexibility as to the schedule analysis requirements and consecutive-days limitations of this rule, if the schedule as delayed does not extend past midnight.

Strasburg suggests that Class III tourist, scenic, historic, and excursion railroad operations should be excluded from the schedule-analysis requirements of this rule, and specifically excluded from the definition of “Type 2 assignment,” because of the nature of these operations. Strasburg contends that, even in their busiest periods, these operations generally operate shorter assignments than the duration permitted for a Type 1 assignment under this rule. In addition, employees rarely work more than five days in a row, and schedules begin and end at the same time and location each day. FRA acknowledges that the nature of these operations reduces the risk of cumulative fatigue experienced by employees of such railroads.

While FRA does not believe these operations should be categorically excluded from the requirements of this regulation, FRA will delay the compliance date for tourist, scenic, historic and excursion railroads until 18 months from the effective date of the final rule, or a year longer than other railroads will have to complete their work schedule analysis and make any required submission of schedules and fatigue mitigation tools to FRA.

This extra year to prepare to comply would allow additional time for such operations to obtain necessary resources, but may also allow many such operations to avoid the necessity of obtaining access to an approved biomathematical model and analyzing schedules, if their only Type 2 assignments had already been approved by FRA on the submission of another railroad, or had been modeled by another railroad and showed that they could be treated as Type 1. This deferral of the compliance date is also consistent with a suggestion in APTA's comments that FRA should allow a schedule approved for one railroad to be used by others without also having to analyze the same schedule. FRA will create a public docket of schedules that it has approved, but if such a listing is to be complete, railroads would have to submit to the docket established for that purpose those Type 2 schedules that they analyze and determine do not violate the fatigue threshold and do not need to be mitigated or submitted to FRA for approval and can be treated as Type 1.

Comments on Consecutive-Days Provisions (§ 228.405(a)(3) and (a)(4))

BLET/UTU and TTD contend that FRA has not made a sufficient case for imposing the limitation on employees working only Type 1 assignments included in the proposed rule, which would require that if an employee had not had at least two calendar days in which he or she had not initiated an on-duty period in a period of 14 consecutive calendar days, that employee must have two consecutive calendar days off duty at his or her

home terminal (unless the fourteenth day ended at his or her away-from-home terminal, in which case the employee would be permitted to work a fifteenth day to return to his or her home terminal and then would be required to have two consecutive calendar days off duty at his or her home terminal). BLET and UTU note that schedule analysis conducted during the RSAC process did not support a limitation on Type 1 assignments, and they argue that the proposed limitation was therefore not based on science but was a subjective requirement. FRA does not dispute the assertion that the work schedule analysis did not suggest the specific limitation proposed and adopted in the final rule. However, as FRA stated in the NPRM, even a Type 1 schedule that allowed the minimum rest required by this regulation would eventually result in an employee using time for other life activities (such as commuting, eating, grooming, personal errands,

etc.

) that the approved models assume to be available for sleep, if the employee is not at some point required to have a day off. FRA also notes that fatigue science indicates that individuals may require more than one recovery day to recover from sleep restriction.

See,

e.g.,

Balkin, T.J.

et al.

“Effects of Sleep Schedules on Commercial Motor Vehicle Driver Performance,” FMCSA Technical Report No. DOT-MC-00-133, U.S. Department of Transportation, (2000); Belenky

et al.,

“Patterns of performance degradation and restoration during sleep restriction and subsequent recovery: A sleep dose-response study,” Journal of Sleep Research, 12, 1-12, (2003).

In contrast to the position of BLET/UTU, NIOSH says it may be premature to say that an employee working even Type 1 schedules will get sufficient rest, noting that if an employee has only the required minimum 8 hours off duty between duty tours, this will not allow the employee to get 8 hours of rest. Likewise, AASM suggests that the required minimum off-duty period under the regulation should be sufficient to allow for an 8-hour sleep period. FRA is comfortable with the limitations included in the rule, because of the nature of the operations in question, and the fact that the diary study of passenger train employees indicated that these employees are usually getting appropriate amounts of sleep, and most are not subjected to fatigue that would violate a threshold established in this regulation. However, FRA believes that the support of the scientific community for even more stringent limitations indicates that the limitations included in this regulation are quite reasonable.

Many comments asked for further clarification and examples to aid in the discussion of the limitation on Type 1 assignments, and these clarifications have been made throughout the final rule in the many references to this provision, and rule text has been added to clarify the application of these limitations. See § 228.405(a)(3) and the discussion of the provision in Section V, Section-by-Section Analysis.

For example, in the NPRM, FRA stated that if an employee worked only Type 1 assignments for a period of more than 6 consecutive calendar days but less than 14 consecutive calendar days, and then initiated an on-duty period involving a Type 2 assignment, the employee would be required to have the Type 2 assignment's rest period of 24 consecutive hours at the employee's home terminal, and then start the count over with regard to consecutive days or total days worked in a 14-day period. In response, MTA asks in its comment what would happen if an employee worked Type 1 assignments on 13 consecutive days, and then a Type 2 assignment on day 14. If the assignment on the 14th consecutive day had been a Type 1 assignment, the employee would have to have two consecutive calendar days off. It does not make sense to require only 24 consecutive hours off after a more fatiguing Type 2 assignment at that point. FRA has revised the rule text in § 228.405(a)(3) to clarify this issue, and other questions related to the application of these provisions.

Comments on Definitions of “Type 1 Assignment” and “Type 2 Assignment” (§ 228.5)

SEPTA, AAR and APTA each argue that the definition of “Type 2 assignment” should be modified to cover any assignment with time between midnight and 3 a.m., rather than 4 a.m., and that Type 1 assignments should be allowed to begin at 3 a.m. They point to a citation in the NPRM to the FAST validation study, which indicated a 20-percent increase in the risk of a human factors accident by working between the hours of midnight and 3 a.m. This causes AAR to conclude that 4 a.m. is an arbitrary threshold. However, 3 a.m. is actually the absolute low point for circadian rhythm, so it is actually the worst possible time to begin a shift, especially since to do so would require being awake in the period before that, in order to report for duty at 3 a.m. Indeed, NIOSH points out that even the 4 a.m. start time can have the same effect as an overnight shift because the employee must wake up earlier to report for duty at 4 a.m. Therefore, FRA has not modified the definitions as requested.

SEPTA and MTA suggest that Type 1 assignments that are delayed such that they extend past the Type 1 hours, or Type 2 assignments that model as Type 1 and are delayed, should still be treated as Type 1 assignments. This seems reasonable to FRA, as it does not seem appropriate for a schedule to have to be modeled every day if it runs a few minutes late. However, if the delay results in the employee's working in the midnight-to-4-am time period that is always to be considered a Type 2 assignment, the assignment must be considered Type 2 for that day, and the employee who worked it will have worked a Type 2 assignment for the purposes of the consecutive-days limitation. FRA has added rule text to clarify this issue. See § 228.5.

Comments About Nap Policies and Sleep Facilities (§ 228.409)

MTA suggests reducing the minimum nap period to be eligible for fatigue mitigation to 60 minutes instead of 90 minutes. The FRA-proposed 90-minute minimum nap period was the subject of significant Working Group discussion, and FRA does not see a significant reason to change it at this time. FRA notes that the

Commercial Transportation Operator Fatigue Management Reference

indicates that naps should not exceed 45 minutes and that 15-30 minutes should be allowed to fully wake up. If 15 minutes are added to allow time to fall asleep, the total is 75 minutes to 90 minutes.

MTA also suggests allowing railroads to decide on nap policies and sleep facilities unilaterally. FRA believes that the collaboration of labor and management on fatigue mitigation efforts is important to ensure successful fatigue mitigation, and FRA therefore declines to modify these provisions.

Comments About Training (§ 228.411)

Comments about training were centered on the timing of both initial training of existing employees subject to the subpart and immediate supervisors of those employees, and initial training of new employees. The NPRM proposed initial training of such existing employees and supervisors “as soon as practicable.” This description of the deadline was deemed too uncertain. NIOSH suggested initial training should be provided to existing employees and supervisors within 90 days of the effective date of the final rule, while SEPTA recommended delaying the deadline for compliance with the initial training requirement for existing employees and supervisors until December 2012, so that it could be aligned with other railroad training

schedules. FRA believes that SEPTA's proposal is reasonable, has the benefit of certainty, and is consistent with the period for providing training in certain other FRA rules. Consequently, FRA has amended the training provision to require initial training of existing employees and supervisors no later than December 31, 2012.

With regard to initial training of new employees, which FRA proposed to require within 90 days of an employee's working an assignment that would be subject to this rule, AAR commented that this time frame will not allow employees to be trained within the railroads' normal training schedules. FRA has revised the time period in which new employees must be trained to be consistent with the latest version of FRA's forthcoming training standards, which was discussed in the Working Group as a standard with which it was agreed that the training provision in this regulation should be consistent. Therefore, new employees will have to be trained prior to December 31, 2012 or before they begin work, whichever is later.

Other Comments

BLET/UTU and TTD request that FRA require a “10-hour call” prior to an assignment (

i.e.,

notification of the time to report 10 hours in advance of the time at which the employee is requested to report for duty). While FRA agrees that such a requirement would provide predictability as to when an employee will be called to work, adopting a 10-hour call requirement is not possible at this time, as it was not a part of the proposed rule. FRA notes, however, that a 10-hour call is one of the fatigue mitigation tools that was discussed. The regulation requires labor involvement in the determination of fatigue mitigation tools to be applied, so there may be opportunities to voluntarily make use of this scheduling practice.

SEPTA suggested that the rule should place responsibility on the employee not to violate the regulation. FRA agrees that in some circumstances the employee may bear some responsibility, but the railroad bears responsibility for scheduling, so it will also bear some responsibility for scheduling an employee for an assignment that would violate the regulation. The applicable civil penalty provision (49 CFR 228.21) includes a reference to the liability of individuals for civil penalties for violating a requirement or causing the violation of any requirement of part 228, and the penalty schedule for part 228 includes a footnote, common to the penalty schedules of many FRA regulations, providing for the possibility of individual liability for a civil penalty for a willful violation.

Finally, NIOSH says this regulation should be part of a comprehensive fatigue management plan. FRA agrees, and notes that the fatigue mitigation plans applied to particular schedules found to violate the fatigue threshold will be part of overall fatigue management. Appendix D to this rule provides guidance on fatigue management plans. Additional requirements will likely result from other ongoing FRA rulemaking projects.

VI. Section-by-Section Analysis

Subpart A—General

Section 228.1 Scope

FRA is revising this section by adding paragraph (c), which indicates that the regulation prescribes substantive hours of service requirements for train employees engaged in commuter or intercity rail passenger transportation.

Section 228.3 Application

Existing paragraph (a) of this section states that part 228 applies to any railroad or contractor or subcontractor to a railroad except as provided in paragraph (b) of the section.

Paragraph (b) of this section excludes from the scope of this part railroads or a contractor or subcontractor of a railroad that operates only on track inside an installation which is not part of the general railroad system of transportation. This provision would exclude from the coverage of subpart F some tourist, scenic, excursion or historic railroads because they operate off the general system. FRA has otherwise specifically included these operations within the coverage of this regulation, as provided by § 228.401, because if they are not covered by this regulation, their train employees would be subject to the statutory freight hours of service requirements of 49 U.S.C. 21103. As is explained in more detail in the discussion below of § 228.401, FRA believes that Congress intended these operations to be subject to this regulation, rather than the statutory requirements, and FRA does not believe the statutory requirements are appropriate for these operations. Accordingly, FRA is revising paragraph (b) of this section to refer to § 228.401, which is the specific applicability provision for new subpart F.

Paragraph (b) of § 228.3 also excludes from the application of part 228 rapid transit operations in an urban area that are not connected with the general railroad system of transportation. Section 228.401 contains an exclusion for these operations.

Section 228.5 Definitions

FRA is amending this section to add definitions of “Associate Administrator” and “FRA” as used in this part. Section 101 of the RSIA refers to FRA's “Associate Administrator for Railroad Safety” and emphasizes that the Associate Administrator is the Chief Safety Officer. Thus, in this final rule “Associate Administrator” means FRA's Associate Administrator for Railroad Safety/Chief Safety Officer.

FRA is also adding definitions of the terms “Type 1 assignment” and “Type 2 assignment.” As was previously discussed in Section IV, above, these definitions were the subject of significant discussion in the Task Force and the Working Group, particularly because of the implications of a particular schedule's status as a Type 1 assignment or a Type 2 assignment for determining the application of the limitations on consecutive days in § 228.405 and the requirements for analysis of schedules and submission of schedules to FRA for approval in § 228.407. FRA believes that the definitions accommodate the concerns expressed in the Working Group regarding schedules outside the time parameters for a Type 1 assignment that may still present very little risk of an effectiveness score that would violate the fatigue threshold and compromise safety. At the same time, however, the definitions recognize the increased risk of fatigue associated with working late night and very early morning hours, which justifies the application of the more stringent requirements.

FRA added language to these definitions as they appeared in the NPRM to make clear that if an assignment is delayed so that the assignment that an employee actually worked includes any period of time between midnight and 4 a.m., the assignment must be treated as a Type 2 assignment for that employee for purposes of the consecutive days limitations and corresponding rest requirements in section 228.405. As was discussed in Section V, Responses to Public Comments on the NPRM, some commenters suggest that Type 1 assignments, or assignments having some time within the definition of a Type 2 assignment but that modeled acceptably to be treated as Type 1 assignments, should continue to be treated as Type 1 assignments even if delayed.

In most circumstances, this makes sense to FRA, in that railroads should not be expected to model assignments on a daily basis if they extend a few minutes past the 8 p.m. limits of a Type

1 assignment, or past the scheduled end time of a Type 2 assignment that was acceptable to be treated as Type 1. However, if the assignment as delayed includes time between midnight and 4 a.m., such an assignment is always considered Type 2, and an employee working that assignment should have Type 2 consecutive-days limitations and corresponding rest requirements.

FRA has added these terms to this general definitions section for part 228, rather than the definitions specific to subpart F, because these terms are also used in the recordkeeping provisions of subpart B, as amended by this rule.

Subpart B—Records and Reporting

Section 228.11 Hours of Duty Records

Paragraph (c) of this section indicates that paragraphs (b)(13) through (b)(16) do not apply to the records of train employees providing commuter or intercity passenger rail transportation. Paragraphs (b)(13) through (b)(16) relate to substantive provisions of the HSL for train employees, added by the RSIA. As was described above in Section II, these requirements were not extended to train employees on commuter and intercity passenger railroads. The requirements referred to in paragraphs (b)(13) through (b)(16) are not required by this rule and therefore would continue not to apply to train employees providing commuter and intercity rail passenger transportation.

Paragraph (c) of this section now also requires two additional pieces of information, relating to the provisions of § 228.405(a)(3). First, paragraph (c)(1) requires that the record must note the date that begins the series of at most 14 consecutive calendar days that includes the duty tour being recorded. Second, paragraph (c)(2) requires that the record note the date, if any, of a calendar day on which the employee did not initiate an on-duty period prior to the current duty tour in the current series of at most 14 consecutive calendar days. This information will allow the railroad and FRA to determine compliance with the limitations established by paragraph (a)(3), both with respect to calendar days on which the employee did not initiate an on-duty period and consecutive days including one or more Type 2 assignments.

FRA recognizes that most railroads and employees subject to this subpart are currently keeping their hours of service records manually, and it may be burdensome for an employee to be required to keep track of his or her series of at most 14 consecutive days and mark its starting date on the hours of service record each day, as well as indicating whether there had been a prior day off during the series. However, the railroad will have to have some way to track this information. Therefore, if a railroad wishes to keep this information centrally for all of its employees, this will be considered sufficient to satisfy the requirements that the hours of service record include the start date of the at-most 14-day series and the date, if any, that the employee did not initiate an on-duty period during the at-most 14-day series, provided this information is made available to FRA upon request.

Section 228.19 Monthly Reports of Excess Service

FRA is revising paragraph (c) of this section to require railroads to report to FRA instances of excess service related to new substantive limitations contained in § 228.405(a)(3) of this rule. That paragraph limits the number of consecutive days or total days within a series of at most 14 consecutive calendar days that train employees engaged in commuter or intercity passenger railroad transportation may initiate an on-duty period, and requires a minimum amount of time off duty or not initiating an on-duty period after an employee has reached the maximum number of consecutive or total days within the prescribed period, before the employee may return to duty, with different requirements depending on the time of day of the employee's assignments.

Excess service under § 228.405(a)(3)(ii) occurs when an employee has initiated on-duty periods on six consecutive days, including one or more Type 2 assignments, and then initiates a new on-duty period without having had the required 24 consecutive hours off at the home terminal. Paragraph (c)(5) addresses this excess service in the situation when the employee is at his or her home terminal at the end of the duty tour that triggers the rest requirement. Paragraph (c)(6) addresses this excess service, including the exception for an additional initiation of an on-duty

This text is long and has been trimmed here. Open the source document for the complete record.

This is a copy of a public record, reproduced as it was published. It is not legal advice, and it may not be the version a court would rely on. Check the official source before you cite it.