Petition for Writ of Certiorari — W. Scott Harkonen, Petitioner v. United States
Supreme Court briefOct 1, 2018
Ask Donna
What actually matters in this document.
Text
No. 18IN THE
Supreme Court of the United States
W. SCOTT HARKONEN, M.D.,
Petitioner,
v.
UNITED STATES OF AMERICA,
Respondent.
ON PETITION FOR A WRIT OF CERTIORARI TO THE
UNITED STATES COURT OF APPEALS
FOR THE NINTH CIRCUIT
PETITION FOR A WRIT OF CERTIORARI
JOSHUA M. KOPPEL
ALAN E. SCHOENFELD
Counsel of Record
WILMER CUTLER PICKERING
PETER G. NEIMAN
HALE AND DORR LLP
1875 Pennsylvania Ave., NW WILMER CUTLER PICKERING
HALE AND DORR LLP
Washington, DC 20006
7 World Trade Center
250 Greenwich Street
New York, NY 10007
(212) 937-7294
alan.schoenfeld@wilmerhale.com
QUESTION PRESENTED
Whether a writ of error coram nobis should issue
for a petitioner who presents “compelling” new evidence that establishes his actual innocence of the crime
of conviction.
(i)
TABLE OF CONTENTS
Page
QUESTION PRESENTED ............................................... i
TABLE OF AUTHORITIES ........................................... v
OPINIONS BELOW .......................................................... 1
JURISDICTION ................................................................. 2
INTRODUCTION .............................................................. 2
STATEMENT ..................................................................... 6
A. Background ............................................................ 6
B. Dr. Harkonen’s Trial And Appeal .................... 11
C. Petition For A Writ Of Error Coram
Nobis ..................................................................... 15
REASONS
FOR
GRANTING
THE
PETITION ......................................................................... 20
I.
THE COURT SHOULD GRANT REVIEW TO
CONSIDER WHETHER A CLAIM OF NEW
EVIDENCE
DEMONSTRATING
ACTUAL
INNOCENCE CAN EVER JUSTIFY CORAM
NOBIS RELIEF ............................................................. 20
A. The Circuits Are Split On Whether A
Pure Actual Innocence Claim Based On
New Evidence Meets The Standard
For Coram Nobis ................................................ 22
B. Dr. Harkonen’s Compelling New
Evidence Of Actual Innocence Would
Have Warranted Relief In The Sixth,
Eighth, And Tenth Circuits ............................... 25
(iii)
iv
TABLE OF CONTENTS—Continued
Page
C. The Ninth Circuit Erred By Denying
Coram Nobis Relief Despite Finding
Dr.
Harkonen’s
New
Evidence
“Compelling” ........................................................ 29
II. THIS CASE PRESENTS A GOOD VEHICLE TO
ADDRESS THE STANDARD FOR CORAM
NOBIS RELIEF ............................................................. 33
CONCLUSION ................................................................. 34
APPENDIX A: Opinion of the United States
Court of Appeals for the Ninth Circuit,
dated December 4, 2017............................................. 1a
APPENDIX B: Opinion of the United States
District Court for the Northern District of
California, dated August 21, 2015 ............................ 5a
APPENDIX C: Order of the United States
Court of Appeals for the Ninth Circuit
denying petition for rehearing and petition
for rehearing en banc, dated June 1, 2018 ............ 29a
v
TABLE OF AUTHORITIES
CASES
Page(s)
Azzone v. United States, 341 F.2d 417
(8th Cir. 1965).............................................................. 24
Barrow v. United States, 455 F. App’x 631
(6th Cir. 2012).............................................................. 25
Byrnes v. United States, 408 F.2d 599
(9th Cir. 1969).............................................................. 23
Chaidez v. United States, 568 U.S. 342 (2013) ........ 15, 21
Fiswick v. United States, 329 U.S. 211 (1946) ............... 31
Foont v. United States, 93 F.3d 76
(2d Cir. 1996) ......................................................... 22, 24
Gavitt v. Born, 835 F.3d 623 (6th Cir. 2016) .................... 5
Harkonen v. United States, 571 U.S. 1110
(2013) (mem.) ............................................................... 15
Herrera v. Collins, 506 U.S. 390 (1993) .............. 28, 30, 31
Jimenez v. Trominski, 91 F.3d 767
(5th Cir. 1996).............................................................. 32
Kandiel v. United States, 964 F.2d 794
(8th Cir. 1992).............................................................. 24
Kuhlmann v. Wilson, 477 U.S. 436 (1986) ............... 29, 30
Maryland v. Kulbicki, 136 S. Ct. 2 (2015) ....................... 5
Matrixx Initiatives, Inc. v. Siracusano,
563 U.S. 27 (2011) ..................................... 13, 14, 26, 27
Moody v. United States, 874 F.2d 1575
(11th Cir. 1989)...................................................... 22, 23
vi
TABLE OF AUTHORITIES—Continued
Page(s)
Murray v. United States, 704 F.3d 23
(1st Cir. 2013) .............................................................. 21
Schlup v. Delo, 513 U.S. 298 (1995)........................... 30, 32
Telink, Inc. v. United States, 24 F.3d 42
(9th Cir. 1994).............................................................. 23
United States v. Bustillos, 31 F.3d 931
(10th Cir. 1994)...................................................... 25, 32
United States v. Denedo, 556 U.S. 904 (2009) .......21, 29, 31
United States v. Harkonen, 2010 WL 2985257
(N.D. Cal. July 27, 2010) .............................................. 6
United States v. Harkonen, 510 F. App’x 633
(9th Cir. 2013).............................................................. 15
United States v. Hedman, 655 F.2d 813
(7th Cir. 1981).............................................................. 24
United States v. Johnson, 237 F.3d 751
(6th Cir. 2001).............................................................. 25
United States v. Mayer, 235 U.S. 55 (1914) ................... 21
United States v. Miles, 553 F. App’x 846
(10th Cir. 2014)............................................................ 25
United States v. Mills, 221 F.3d 1201
(11th Cir. 2000)............................................................ 23
United States v. Morgan, 346 U.S. 502 (1954) ......21, 29, 31
United States v. Ridings, 569 F. App’x 73
(3d Cir. 2014) ............................................................... 32
United States v. Scherer, 673 F.2d 176
(7th Cir. 1982).............................................................. 24
vii
TABLE OF AUTHORITIES—Continued
Page(s)
United States v. Stoneman, 870 F.2d 102
(3d Cir. 1989) ............................................................... 24
DOCKETED CASES
Harkonen v. United States, No. 13-180 (U.S.) .............. 15
United States v. Harkonen, Nos. 11-10209 &
11-10242 (9th Cir.) ........................................................ 9
STATUTES, RULES, AND REGULATIONS
2 U.S.C.
§ 331 .............................................................................. 12
§ 333 .............................................................................. 12
§ 352 .............................................................................. 12
18 U.S.C. § 1343 ................................................................. 12
28 U.S.C.
§ 1254 .............................................................................. 2
§ 2241 ...................................................................... 15, 21
§ 2255 ...................................................................... 15, 21
17 C.F.R. § 243.100 ............................................................ 10
Fed. R. Crim. P. 33 .............................................................. 5
OTHER AUTHORITIES
Bajwa, Ednan K., et al., Interferon-γ1b
Therapy
in
Idiopathic
Pulmonary
Fibrosis, 128 Chest 203 (2005) .................................... 8
Briggs, William M., The Substitute for pValues, 112 J. Am. Stat. Ass’n 897 (2017) ............... 26
viii
TABLE OF AUTHORITIES—Continued
Page(s)
FDA, Guidance for Industry: Providing
Clinical Evidence for Human Drug and
Biological Products (1998), at https://
tinyurl.com/ya2aentj .................................................. 26
Friendly, Is Innocence Irrelevant? Collateral
Attack on Criminal Judgments, 38 U. Chi.
L. Rev. 142 (1970) ....................................................... 30
Wasserstein, Ronald L. & Nicole A. Lazar, The
ASA’s Statement on p-Values: Context,
Process, and Purpose, 70 Am. Statistician
129 (2016), at https://amstat.tandfonline.com/
doi/pdf/10.1080/00031305.2016.1154108?need
Access=true............................... 4, 16, 17, 18, 26, 27, 28
IN THE
Supreme Court of the United States
No. 18W. SCOTT HARKONEN, M.D.,
Petitioner,
v.
UNITED STATES OF AMERICA,
Respondent.
ON PETITION FOR A WRIT OF CERTIORARI TO THE
UNITED STATES COURT OF APPEALS
FOR THE NINTH CIRCUIT
PETITION FOR A WRIT OF CERTIORARI
Dr. Scott Harkonen respectfully petitions for a writ
of certiorari to review the judgment of the United
States Court of Appeals for the Ninth Circuit in this
case.
OPINIONS BELOW
The opinion of the Ninth Circuit affirming the denial of Dr. Harkonen’s petition for a writ of error coram
nobis (App. 1a-4a) is unpublished but is available at 705
F. App’x 606. The Ninth Circuit’s order denying the
petition for rehearing (App. 29a-30a) is unreported.
The opinion of the district court denying Dr. Harkonen’s coram nobis petition (App. 5a-27a) is unpublished
but is available at 2015 WL 4999698.
2
JURISDICTION
The Ninth Circuit entered judgment on December
4, 2017, and denied a timely rehearing petition on June
1, 2018. App. 1a, 29a-30a. On July 30, 2018, Justice
Kennedy extended the time to file a petition for a writ
of certiorari to October 1, 2018. No. 18A107. This
Court has jurisdiction pursuant to 28 U.S.C. § 1254(1).
INTRODUCTION
W. Scott Harkonen, M.D. is an experienced medical
researcher. In 2009, he was convicted of wire fraud,
based on statistical evidence the government claimed
showed a press release he issued was false. In 2016, the
American Statistical Association formally disavowed
the statistical theory at the heart of the prosecution.
The court below found that the evidence that Dr.
Harkonen is innocent was “compelling.” App. 2a. But
the court declined to vacate Dr. Harkonen’s conviction,
holding that more than actual innocence was required
before a writ of coram nobis vacating the conviction
could be issued. This holding conflicts with the holdings of three other circuits (the Sixth, Eighth, and
Tenth) that actual innocence itself is a basis for coram
nobis relief. The Court should grant review to address
the circuit split on this important question, and should
hold that at least where advances in science establish
innocence, coram nobis relief is appropriate.
The press release at issue reported the results of a
clinical trial. The trial was designed to evaluate the
safety and efficacy of a biologic drug, Actimmune, to
treat a rare, progressive invariably fatal lung disease
known as idiopathic pulmonary fibrosis (IPF).
The trial showed something compelling: There was
a 40% reduction in mortality in the patients taking Ac-
3
timmune compared to those taking placebo. And
among patients treated at relatively early stages of the
disease, there was a 70% reduction in mortality for
those taking Actimmune compared to those taking placebo. As required by the securities laws, the company
issued a press release reporting these results. The
press release described the study as “demonstrat[ing] a
… survival benefit.” C.A.E.R. 1047.
The government conceded that the study was well
designed and had been properly conducted, and that
every piece of data in the press release was entirely accurate. But the government argued that it was a crime
to describe these results as demonstrating anything at
all. The government’s theory was that there was a
hard and fast law of science: Conclusions about causation could only be drawn from data that demonstrated
statistical significance as measured by what is known
as the “p-value test.” In the government’s world, statistical significance, measured by a p-value less than or
equal to .05, was the gateway to drawing scientific conclusions. See C.A.E.R. 594-597. The government elicited testimony supporting this theory from a statistician.
Summing up this theory in closing argument, the prosecution told the jury that because the study failed to
meet the p-value test, “you can’t draw any conclusions
from this trial.” Id. at 595.
More than six years after the jury convicted Dr.
Harkonen, the American Statistical Association (ASA)
issued a statement of fundamental principles declaring
the government’s statistical theory wrong. In particular, the ASA explained that the concept of statistical
significance had been “misused and misinterpreted,”
that reliance on p-values “leads to considerable distortion of the scientific process,” and that a “p-value, or
statistical significance, does not measure the size of an
4
effect or importance of a result,” and is “not equivalent
to scientific, human, or economic significance.” Wasserstein & Lazar, The ASA’s Statement on p-Values:
Context, Process, and Purpose, 70 Am. Statistician 129,
131-132 (2016) (ASA Principles)). In fact, p-values “do[]
not provide a good measure of evidence regarding a
model or hypothesis,” and “[s]cientific conclusions …
should not be based only on whether a p-value passes a
specific threshold.” Id.1
Because Dr. Harkonen was no longer in custody, he
petitioned for a writ of error coram nobis to avoid the
lingering adverse consequences of his conviction. The
Ninth Circuit found Dr. Harkonen’s evidence of actual
innocence “compelling” but, because it found Dr.
Harkonen had not shown an intervening change in law
or an independent constitutional error in the conduct of
the trial, it held that he had not “establish[ed] that his
trial resulted in a manifest injustice warranting issuance of the writ.” App. 2a. In doing so, the Ninth Circuit deepened a circuit split over whether a petitioner
is entitled to coram nobis relief upon a showing of new
evidence demonstrating his actual innocence, or whether a petitioner must also allege a constitutional or jurisdictional error in the proceedings leading to his conviction. This Court should grant certiorari to resolve
that important question.
Although the question whether new evidence of actual innocence justifies coram nobis relief can arise in a
variety of contexts, it is particularly important where,
as here, the new evidence consists of scientific developments that discredit key evidence or theories used at
1
The ASA Principles are available at https://amstat.tandfon
line.com/doi/pdf/10.1080/00031305.2016.1154108?needAccess=true.
5
trial. The history of criminal law is replete with examples of convictions founded on bad science, whether it
be Comparative Bullet Lead Analysis, see Maryland v.
Kulbicki, 136 S. Ct. 2, 3-4 (2015) (per curiam), outdated
understandings of burn patterns in arson science, see
Gavitt v. Born, 835 F.3d 623, 630-636 (6th Cir. 2016), or,
as here, misplaced reliance on the concept of statistical
significance. Fundamental fairness requires that there
be a mechanism to challenge such convictions, but a defendant cannot bring a motion for a new trial when, as
is often the case, the scientific developments come to
light more than three years after the defendant is convicted, see Fed. R. Crim. P. 33(b)(1), and constitutional
challenges to these convictions are often unavailing, see
Kulbicki, 136 S. Ct. at 4-5 (rejecting an ineffectiveassistance-of-counsel claim because counsel could not
have been expected to predict that Comparative Bullet
Lead Analysis would be discredited).
In these cases, the only potential remedy is a freestanding claim of actual innocence based on new evidence. And where the science has taken years to develop, the defendant will frequently be out of custody
before the evidence proving his innocence is discovered, and the only mechanism for seeking relief is a coram nobis petition. Under the Ninth Circuit’s rule—
and that of four other circuits—innocent defendants
whose convictions are premised on discredited science
have no plausible avenue for relief. This Court should
grant certiorari to resolve the circuit split and affirm
that—at least where new scientific evidence irrefutably
establishes actual innocence—defendants may obtain
relief through a petition for a writ of error coram nobis.
6
STATEMENT
A. Background
Dr. Harkonen is an accomplished medical doctor,
researcher, and biotechnology executive. His work has
led to the approval of FDA labeling for five new disease
indications, including three for rare, orphan diseases.
In 1998, Dr. Harkonen founded InterMune, a biotechnology company focused on developing new treatments
for unmet medical needs. InterMune began trading as
a public company in 2000, and Dr. Harkonen served as
InterMune’s CEO until 2003. See United States v.
Harkonen, 2010 WL 2985257, at *1 (N.D. Cal. July 27,
2010).
One of the therapies under development at InterMune was Actimmune. Harkonen, 2010 WL 2985257,
at *1. Actimmune is a bio-engineered form of interferon gamma that occurs naturally in the human body and
that inhibits fibrosis that scars lung tissue. Actimmune
had been approved for two other orphan diseases. Researchers found that in IPF, patients have abnormally
low levels of interferon gamma in their bodies. This
observation, along with Actimmune’s ability to inhibit
fibrosis, suggested it might be useful in treating IPF,
which involves fibrotic scarring of lung tissue. Id.
There were no approved treatments for IPF.
In 1999, the New England Journal of Medicine published the results of a randomized, controlled trial involving 18 patients with IPF (the “Phase II trial”). InterMune had no role in the trial. After twelve months,
the patients receiving interferon gamma had meaningful improvement in lung function tests. See C.A.E.R.
529. The long-term survival status of the Phase II
study patients was assessed over five years; 78% of the
patients who received interferon gamma remained
7
alive, while only 16% in the control group survived.
The p-value for this result was .009. Id. at 282, 1048.
Because Actimmune was already approved by the
FDA for other uses, InterMune had the option of conducting no further research, and letting doctors choose
to prescribe Actimmune based on the Phase II trial.
Instead, after consultation with the FDA, InterMune
decided to fund a gold-standard, double-blind, multicenter clinical study (the GIPF-001 study) to develop
further information on the safety and efficacy of Actimmune as a treatment for IPF. See App. 7a. A Steering Committee of leading experts in pulmonary medicine and IPF supervised the trial, which enrolled 330
patients at 58 different study locations. C.A.S.E.R. 626,
652.
The results of the GIPF-001 study were reported
to InterMune in August 2002. On safety, the drug was
very well tolerated, confirming that physicians could
prescribe Actimmune for IPF patients without harming them. C.A.E.R. 1048. On efficacy, the results were
mixed. The primary efficacy endpoint—progressionfree survival—showed approximately a 10% reduction
in death or disease progression, but the effect was too
small to be clinically meaningful to patients with IPF.
Id. The result also did not achieve statistical significance on the “p-value” test. Id. at 1047.
While the data did not meet the study’s primary
endpoint, the results on the mortality endpoints were
compelling. Across the entire study population, the
mortality rate among those who took Actimmune was
40% lower than in the group taking the placebo,
C.A.E.R. 1048, a result consistent with the overall mortality benefit observed in the long-term follow-up to the
Phase II trial. The p-value for this result was .084. Id.
8
The mortality results were even more pronounced
among patients with mild-to-moderate IPF (254 out of
the 330 patients in the study, id.). In that group, the
survival rate was 70% higher for patients taking Actimmune than for those taking placebo. Id. Only 6 of
the 126 patients on Actimmune died during the study.
Id. More than three times as many—21 of the 128 patients on the placebo—died. Id. The p-value for this
result was .004. Id.
The size and consistency of the mortality results
across two independent studies, and the persistence of
the results across the entire population in the Phase III
trial, and across various large subgroups in that trial,
made a powerful case that there was a real mortality
benefit.2
The graphs below depict what is known as a
Kaplan-Meier analysis, and illustrate the dramatic survival advantage for patients taking Actimmune as compared to placebo in the Phase III trial.
2
Indeed, a later, independent meta-analysis of the data from
the Phase II and Phase III trials, among other trials, concluded
that interferon gamma “treatment is associated with decreased
mortality.” Bajwa et al., Interferon-γ1b Therapy in Idiopathic
Pulmonary Fibrosis, 128 Chest 203, 206 (2005).
9
12/4/2009 Topel Decl. Ex. S at 6 (D. Ct. Dkt. 249-6);
C.A.E.R. 2366, United States v. Harkonen, Nos. 1110209 & 11-10242 (9th Cir. Oct. 31, 2011). The analyses
show the large separation between the survival rates
for the drug and placebo group, and that the separation
increased over time.
10
Other data points in the study also supported the
inference that Actimmune reduced mortality for IPF
patients. There was a trend in favor of Actimmune in
terms of both improved breathing and reduced need for
supplemental oxygen. C.A.E.R. 1048. The results
were reviewed at an off-site meeting on August 27 with
the company’s employees and with the study’s Steering
Committee and principal investigators.
Once the GIPF-001 study’s results were disclosed
to the corporate outsiders for their review and analysis,
InterMune was obligated by federal securities law prohibiting selective disclosure promptly to disclose those
results to investors and the public. C.A.E.R. 753, 808809, 811-815; see 17 C.F.R. § 243.100. Accordingly, InterMune issued a press release, held a conference call,
and posted the top line results of the GIPF-001 study
on its website on August 28, 2002. C.A.E.R. 1047-1050;
12/4/2009 Topel Decl. Ex. SS (D. Ct. Dkt. 249-13).3
The press release’s headline stated: “InterMune
Announces Phase III Data Demonstrating Survival
Benefit of Actimmune in IPF.” C.A.E.R. 1047 (capitalization altered). The subtitle read: “Reduces Mortality
by 70% in Patients with Mild to Moderate Disease.” Id.
The release further stated that the “preliminary”
data “demonstrate a significant survival benefit in pa3
At trial, the government focused heavily on the mechanics
of the preparation of the press release, urging the jury to find a
culpable mental state from Dr. Harkonen’s efforts to limit the
number of people involved in drafting the release. That argument
was misdirected, because limiting the number of people involved
in coordinating the release of material, non-public information was
routine and appropriate, particularly where, as here, the release
included revisions to its future earnings projections.
11
tients with mild to moderate disease randomly assigned
to Actimmune versus control treatment (p=0.004).”
C.A.E.R. 1047. It stated in the first paragraph that
there was a 10% reduction in progression-free survival
and that this result was not statistically significant. Id.
at 1047, 1048. It explained that the preliminary results
were part of a continuing study to track “longer term
outcomes.” Id. at 1047-1050. And the press release
provided the underlying data on the survival benefit,
including the number of patients who lived and died
and the associated p-values for the study group as a
whole and for those patients who comprised the mildto-moderate-IPF subgroup. Id. at 1048; see also id. at
723:20-21.
The press release was only InterMune’s initial disclosure related to the study. As promised in the release, that same day, InterMune conducted a conference call with investors at which additional details of
the study results were discussed. See 12/4/2009 Topel
Decl. Ex. SS (D. Ct. Dkt. 249-13). And over the next
several weeks, leading physicians involved in the study
provided comprehensive analyses of the study results,
including the same survival analyses presented in the
press release, at various medical conferences. See
12/4/2009 Topel Decl. Exs. C, Z (D. Ct. Dkts. 249-1, 2496). Thereafter, InterMune met with the FDA to review
the study results and discuss requirements for including an IPF indication in Actimmune’s labeling.
B. Dr. Harkonen’s Trial And Appeal
The government launched an investigation of InterMune largely focused on allegations of “misbranding,” that is, that InterMune encouraged doctors to
prescribe Actimmune for IPF without obtaining FDA
approval to market it for that use. In 2006, InterMune
12
entered into a deferred prosecution agreement and paid
a settlement. C.A.S.E.R. 4031.
In 2008, Dr. Harkonen was indicted, both for misbranding, in violation of 2 U.S.C. §§ 331(k), 333(a)(2),
and 352(a), and for wire fraud, in violation of 18 U.S.C.
§ 1343. C.A.E.R. 1036-1045. The premise of the wire
fraud charge was that the press release “falsely portrayed the results of the GIPF-001 Phase III trial as
establishing that Actimmune helped IPF patients live
longer.” Id. at 1041. The only statements specifically
alleged to be false were the press release’s headline and
subheading. Id. at 1041-1042. These statements were
allegedly made “to induce doctors to prescribe, and patients to take, Actimmune to treat IPF.” Id. at 1041.
At trial, the government acknowledged that the
press release accurately reported the data from the
GIPF-001 study. But the government took issue with
the conclusions about those data as stated in the headline and subheading of the release. The government’s
theory was that “no reasonable scientist could have, in
good faith, reported the trial to be anything but an abject failure because its results did not meet certain statistical principles [the government] argued were immutable.” App. 6a-7a.
Specifically, the government argued that because
the study failed to achieve statistical significance, no
conclusions could be drawn from the trial. C.A.E.R.
212-214; see also id. at 3, 594-596. As the prosecution
argued in summation: “The only meaningful p-value for
this trial is the p-value for the primary endpoint. And
that p-value was 0.5; nowhere near statistical significance. What that p-value means … [is] that you can’t
draw any conclusions from this trial.” Id. at 595. Thus,
the government’s case expressly relied on a particular
13
(and subsequently-discredited) view of statistical significance: that because the study results did not meet
the p-value test for statistical significance, the data was
unreliable and no inferences could be drawn from the
study.4
On September 29, 2009, the jury found Dr. Harkonen guilty of wire fraud and not guilty of misbranding.
C.A.E.R. 550. Dr. Harkonen moved for a judgment of
acquittal for insufficient evidence and on First
Amendment grounds and for a new trial. Id. The district court denied the motion on July 27, 2010. Id. at
550-582. In doing so, the court relied on the testimony
of the government’s witnesses that “a p-value of 0.05 is
somewhat of a magic number,” id. at 557, and that if the
primary endpoint fails to achieve a p-value of less than
.05, no conclusions can be drawn from the trial, id. at
558-559.
On February 14, 2011, prior to sentencing, Dr.
Harkonen filed a second new trial motion. Dr. Harkonen argued that an amicus brief filed by the United
States in Matrixx Initiatives, Inc. v. Siracusano, 563
U.S. 27 (2011), constituted newly-discovered exculpatory evidence because the government took a position in
that brief that “directly conflict[ed] with the testimony
it elicited at trial and argued to the jury.” 2/14/2011
Mot. for New Trial 12 (D. Ct. Dkt. 322). Specifically,
the United States took the position in Matrixx that “a
determination that certain data are not statistically
4
Just as the prosecution was taking this position in Dr.
Harkonen’s trial, the NIH was issuing a press release touting the
success of an HIV vaccine trial despite argument that “per protocol” results were not statistically significant. See 12/4/2009 Motion
to Dismiss 19 (D. Ct. Dkt. 247).
14
significant … does not refute an inference of causation”
and that “causation can appropriately be inferred
through consideration of multiple factors independent
of statistical significance.” Id. at 1 (internal quotation
marks omitted; alteration in the original).
In opposing the new trial motion, the government
sought to recharacterize its brief in Matrixx. The Matrixx brief, according to the government, was about
whether investors might care about safety data drawn
from a statistically insignificant number of adverse
events. But the Matrixx brief, the government argued,
did not resolve the role of statistical significance in
drawing conclusions about efficacy from study data. In
the government’s view, it remained improper to “assert
that statistically insignificant data proves a drug
works.” 3/7/2011 Opp. to Mot. for New Trial 7 (D. Ct.
Dkt. 343).
In the midst of the briefing on the new trial motion,
this Court issued its decision in Matrixx. Relying in
large part on the government’s amicus brief, this Court
rejected the argument that “statistical significance is
the only reliable indication of causation,” and stated
that “[a] lack of statistically significant data does not
mean that medical experts have no reliable basis for
inferring a causal link between a drug” and observed
clinical outcomes. Matrixx, 563 U.S. at 40. Moreover,
the Court acknowledged that “medical experts rely on
other evidence to establish an inference of causation.”
Id. For example, researchers and clinicians consider
the “strength of the association” and “temporal relationship” between a drug and the observed outcome;
“consistency across studies; biological plausibility; consideration of alternative explanations; specificity …;
the dose-response relationship; and the clinical and
pathological characteristics of the event.” Id. at 40 n.7.
15
Notwithstanding Matrixx, the district court denied
Harkonen’s motion for a new trial. C.A.E.R. 228. On
April 13, 2011, the court imposed a sentence of three
years’ probation and six months’ home detention. Id. at
232-237.
The Ninth Circuit affirmed Dr. Harkonen’s conviction and sentence. United States v. Harkonen, 510 F.
App’x 633, 635 (9th Cir. 2013). Dr. Harkonen petitioned
for certiorari, arguing that a conclusion about the
meaning of scientific data cannot be a “false or fraudulent” statement within the meaning of the wire fraud
statute and that applying that statute to scientific conclusions would violate the First Amendment’s proscription against viewpoint discrimination. Pet., Harkonen
v. United States, No. 13-180 (U.S. Aug. 5, 2013). This
Court denied certiorari on December 16, 2013. 571 U.S.
1110 (2013).
C. Petition For A Writ Of Error Coram Nobis
On July 30, 2014, Dr. Harkonen filed a petition for a
writ of error coram nobis in the district court challenging the statistical basis for the government’s theory at
trial. 7/30/2014 Pet. for Writ of Error Coram Nobis (D.
Ct. Dkt. 399).5 Dr. Harkonen explained that the government statistical witness’s testimony regarding the
5
The writ of error coram nobis provides a way to collaterally
attack a conviction for a person, like Dr. Harkonen, who is no
longer in custody and therefore cannot seek habeas relief or file a
§ 2255 motion. Chaidez v. United States, 568 U.S. 342, 345 n.1
(2013); see 28 U.S.C. § 2241(c) (“The writ of habeas corpus shall not
extend to a prisoner unless … [h]e is in custody ….”); id. § 2255(a)
(“A prisoner in custody under sentence of a court … may move
the court which imposed the sentence to vacate, set aside or correct the sentence.” (emphasis added)).
16
proper interpretation of data and the import of data
showing statistically-insignificant results was inconsistent with generally-accepted principles of biostatistics. 7/30/2014 Memo ISO Pet. for Writ of Error Coram
Nobis 72-74 (D. Ct. Dkt. 399-1). The district court denied the petition as untimely on August 21, 2015. App.
5a, 17a-22a.
Dr. Harkonen appealed the district court’s denial of
his petition. On appeal, he argued that a recent scientific development had disproven the theory upon which
he was convicted. Dr. Harkonen explained that the
American Statistical Association had, in 2016—more
than six years after the jury found him guilty—
promulgated a statement of six fundamental principles
that govern the field of statistics. Pet. C.A. Br. 43-44
(citing the ASA Principles). Dr. Harkonen argued that
the ASA Principles “declare[] to be false” “precisely
the statistical propositions” on which the “government
convicted Dr. Harkonen.” Id. at 44. Accordingly, “his
conviction must now be declared void.” Id.
The ASA issued these principles because the concept of statistical significance and the use of p-values
had been “commonly misused and misinterpreted,”
which had led to “considerable distortion of the scientific process.” ASA Principles 131. Through the Principles, the ASA sought to “improve the conduct or interpretation of quantitative science” by “clarifying …
principles underlying the proper use and interpretation
of the p-value” based on “widespread consensus in the
statistical community.” Id.
The conclusions from the ASA Principles are the
opposite of the government’s theory of guilt. The government presented statistical significance as the gateway through which clinical trial data must pass, with p-
17
values as “magic number[s]” that determine the reliability of study results. C.A.E.R. 557. The government
flatly told the jury that the p-value on the primary
endpoint, by itself, showed the study had “failed” and
precluded drawing any conclusions from it. Id. at 594595.
The ASA Principles establish that the scientific
consensus is now squarely to the contrary. The ASA
Principles explain that the consensus of statisticians is
that statistical significance is not the gateway to sound
inferences from study data, and p-values are not magic
at all. Indeed, “[p]-values do not measure the probability that the studied hypothesis is true, or the probability that the data were produced by random chance
alone.” ASA Principles 131.6 Thus, “[a] conclusion does
not immediately become ‘true’” if data generates a particular p-value (such as less than .05) or (more importantly here) “‘false’” if it does not. Id.7 Nor does a
p-value “measure the size of an effect or the importance
of a result.” Id. at 132.
All that is consistent with this Court’s statement in
Matrixx that a lack of statistical significance is not a
bar to causation, and that scientists appropriately consider other factors. But the ASA Principles go much
further. They explain that p-values are not an im6
The district court, in upholding the verdict, had said just the
opposite: “The lower a p-value is, the greater probability that the
result perceived in the data is not due to chance.” C.A.E.R. 557.
7
Again, the district court in upholding the verdict had said
precisely the opposite: Simply because the “data for the secondary endpoint of survival time yielded a p-value [higher than .05] …
the jury could have concluded beyond a reasonable doubt that [the
press release] … was false.” C.A.E.R. 563.
18
portant tool in assessing causation. To the contrary,
“[b]y itself, a p-value does not provide a good measure
of evidence regarding a model or hypothesis.” ASA
Principles 132. A low p-value “offers only weak evidence against the null hypothesis.” Id. And a “relatively large p-value does not imply evidence in favor of
the null hypothesis.” Id.8 Accordingly, as the ASA
Principles state, “[s]cientific conclusions and business
or policy decisions should not be based only on whether
a p-value passes a specific threshold.” Id. at 131. Rather, researchers “should bring many contextual factors into play to derive scientific inferences, including
the design of a study, the quality of the measurements,
the external evidence for the phenomenon under study,
and the validity of assumptions that underlie the data
analysis.” Id.
Dr. Harkonen argued (Pet. C.A. Br. 43-48) that the
current understanding of statistical significance and pvalues—as reflected in the ASA Principles—directly
undermines the government’s theory at trial that no
reasonable scientist could “conclude that Actimmune
has a survival benefit” simply because the GIPF-001
trial did not yield statistically significant results.
C.A.E.R. 595. With p-values in their proper place—not
as magic numbers, or a gateway that must be crossed
before drawing conclusions, but rather as evidence too
“weak” to form the basis for scientific conclusion—the
“logical conclusion from [the survival] data was that it
showed or ‘demonstrated’ a mortality benefit,” as stated in the press release. Pet. C.A. Br. 44. Accordingly,
8
The “null hypothesis” in the context of a drug trial is the
hypothesis that there is no evidence of an association between the
drug and the effect.
19
the conclusions stated in the press release could not be
false or fraudulent.
In response, the government acknowledged that
“experts can use statistically insignificant data coupled
with other information to infer a link between a drug
and a particular effect in patients.” Resp. C.A. Br. 66.
But it maintained that this pivotal concession “has
nothing to do with Harkonen’s conviction,” which it instead attributed to the supposed fact that Dr. Harkonen “did not rely on other evidence in asserting that
the clinical trial results supported the inference that
Actimmune was responsible for a survival benefit.” Id.
That argument, of course, was never made to the jury,
which was instead told that the statistical results, by
themselves, precluded drawing conclusions from the
study. It is also contrary to the plain language of the
press release, which refers expressly to other evidence,
such as the Phase II study results, and the supportive
trends on other endpoints. And it does not reckon with
the ASA Principles, which make clear something the
jury was never told: The p-value results do not count
as important evidence against the efficacy conclusion.
The Ninth Circuit rejected the government’s effort
to salvage the factual basis for the conviction. Instead,
the court of appeals found Dr. Harkonen’s “proffered
evidence” of innocence “compelling, especially in light
of Matrixx.” App. 2a. Nevertheless, it affirmed the
judgment below. For the Ninth Circuit, innocence was
not enough. To be entitled to relief, Dr. Harkonen
needed to point to a “change in controlling law.” Id.
Because he had not done so, the court held that he had
failed to “establish that his trial resulted in a manifest
20
injustice warranting issuance of [a] writ” of error coram nobis. Id.9
Dr. Harkonen filed a petition for rehearing and rehearing en banc arguing that the Ninth Circuit had
erred in denying his coram nobis petition despite finding his claim of actual innocence “compelling.” Pet.
C.A. Reh’g Pet. 10-16. Dr. Steven Goodman, Associate
Dean of Clinical and Translational Research and Professor of Medicine and of Health Research and Policy at
Stanford University, filed an amicus brief in support of
Dr. Harkonen. Dr. Goodman confirmed that the ASA’s
statement of principles “utterly discredited” the “government’s ‘immutable’ theory of statistics upon which it
obtained Dr. Harkonen’s conviction” and demonstrated
Dr. Harkonen’s innocence. Goodman C.A. Br. 9. Dr.
Goodman added that “leaving this conviction in place
would represent an implicit endorsement of scientific
falsehoods,” which “does not serve the courts, science
or society well.” Id. at 7. The Ninth Circuit called for a
response, but denied Dr. Harkonen’s petition for rehearing on June 1, 2018. App. 29a-30a.
REASONS FOR GRANTING THE PETITION
I.
THE COURT SHOULD GRANT REVIEW TO CONSIDER
WHETHER A CLAIM OF NEW EVIDENCE DEMONSTRATING ACTUAL INNOCENCE CAN EVER JUSTIFY CORAM
NOBIS RELIEF
A petition for a writ of error coram nobis provides
a way to collaterally attack a criminal conviction when
the petitioner is no longer “in custody,” and therefore
9
The court did not reach the question whether Dr. Harkonen’s coram nobis petition was timely. App. 4a n.1.
21
cannot seek relief under 28 U.S.C. §§ 2241 or 2255.
Chaidez v. United States, 568 U.S. 342, 345 n.1 (2013)
see also United States v. Morgan, 346 U.S. 502, 506-511
(1954). Like habeas corpus, the writ of error coram
nobis was originally confined to cases where the tribunal lacked jurisdiction or where other errors rendered
the proceeding invalid. At common law, coram nobis
was available to correct “errors in matters of fact which
… were material to the validity and regularity of the
legal proceeding itself; as where the defendant, being
under age, appeared by attorney, or the plaintiff or defendant was a married woman at the time of commencing the suit, or died before verdict or interlocutory
judgment.” United States v. Mayer, 235 U.S. 55, 68
(1914). But the scope of coram nobis—like that of habeas corpus—has been expanded to provide a remedy
for a variety of constitutional errors or otherwiseunjust verdicts. As the Court said in Morgan, although
the writ of error coram nobis is generally described by
reference to a narrow set of instances warranting relief, see, e.g., Mayer, 235 U.S. at 68, “its use has been by
no means so limited,” Morgan, 346 U.S. at 507-508; see
also United States v. Denedo, 556 U.S. 904, 910-911
(2009).
Although the Court has spoken of coram nobis in
general terms, it has not “addressed the precise standards that lower courts should use in deciding whether
to issue” the writ. Murray v. United States, 704 F.3d
23, 29 (1st Cir. 2013). One question, in particular, that
arises on coram nobis review is whether a court may
grant the writ based on new evidence of actual innocence. It often takes years for new evidence to materialize, by which point defendants have frequently completed any term of imprisonment and coram nobis is the
only avenue for seeking relief. The courts of appeals
22
have taken differing positions on whether an actualinnocence claim warrants coram nobis relief. The Court
should grant certiorari to resolve that split and confirm
that where a petitioner presents compelling evidence of
actual innocence, coram nobis relief is available.
A. The Circuits Are Split On Whether A Pure
Actual Innocence Claim Based On New Evidence Meets The Standard For Coram Nobis
The Second, Third, Seventh, Ninth, and Eleventh
Circuits have held that a freestanding claim of actual
innocence does not warrant coram nobis relief. In those
circuits, a petitioner must allege a constitutional or jurisdictional error. In contrast, courts in the Sixth,
Eighth, and Tenth Circuits will grant coram nobis relief
if the petitioner brings newly-discovered evidence
demonstrating his actual innocence.
1. Five circuits, including the Ninth Circuit in this
case, have held that a “claim of newly discovered evidence relevant … to the guilt or innocence of the petitioner is not cognizable in a coram nobis proceeding.”
Moody v. United States, 874 F.2d 1575, 1577 (11th Cir.
1989). In those circuits, even a petitioner who presents
new evidence conclusively demonstrating his innocence
cannot obtain relief unless that evidence reveals an independent “constitutional or jurisdictional error.”
Foont v. United States, 93 F.3d 76, 80 (2d Cir. 1996).
In Moody, for example, Walter Moody had been
convicted of possessing an unregistered destructive
device. 874 F.2d at 1576. Moody petitioned for coram
nobis relief on the ground that new evidence proved
that he did not commit the crime. Id. The Eleventh
Circuit rejected the claim without considering its merits, stating that “newly discovered evidence affords no
23
entrée to the writ.” Id. at 1577 (internal quotation
marks and brackets omitted); see also United States v.
Mills, 221 F.3d 1201, 1205 (11th Cir. 2000) (“allegations
of newly discovered evidence are not cognizable in a
petition for coram nobis”).
The Ninth Circuit has taken a particularly narrow
view of coram nobis, stating that, to warrant relief, a
coram nobis petition must present a “constitutional
controversy.” Byrnes v. United States, 408 F.2d 599,
602 (9th Cir. 1969); see also Telink, Inc. v. United
States, 24 F.3d 42, 45 (9th Cir. 1994) (“The writ of error
coram nobis affords a remedy to attack an unconstitutional or unlawful conviction in cases when the petitioner already has fully served a sentence.”). A freestanding claim of actual innocence—one in which the
petitioner has discovered new evidence that proves
that he did not commit the crime of conviction, but
where he cannot establish that an independent constitutional error infected the proceedings that led to conviction—does not warrant coram nobis relief.10
The Second, Third, and Seventh Circuits have similarly held that a petitioner cannot obtain coram nobis
relief by showing only that he is innocent of the crime
of conviction; the petitioner must also demonstrate an
independent constitutional or jurisdictional error. See
10
The Ninth Circuit applied this heightened standard here.
The court recognized that Dr. Harkonen’s “proffered evidence”—
which demonstrated that the conclusions stated in the press release were well within the realm of reasonable scientific debate—
were “compelling.” App. 2a. The court of appeals denied the writ
nonetheless, finding that this demonstration of actual innocence
did “not establish that [Dr. Harkonen’s] trial resulted in a manifest
injustice warranting issuance of the writ,” because he had not
“point[ed] to [any] change in controlling law.” Id.
24
Foont, 93 F.3d at 80 (“Claims of new evidence, … without constitutional or jurisdictional error in the underlying proceeding, cannot support a coram nobis claim.”);
United States v. Stoneman, 870 F.2d 102, 106 (3d Cir.
1989) (“The error must go to the jurisdiction of the trial
court, thus rendering the trial itself invalid.”); United
States v. Scherer, 673 F.2d 176, 178 (7th Cir. 1982)
(“[T]he burden is on the petitioner to demonstrate that
‘the asserted error is jurisdictional or constitutional,
and involves an error of law that results in a complete
miscarriage of justice.’” (internal brackets omitted));
United States v. Hedman, 655 F.2d 813, 815 (7th Cir.
1981) (coram nobis “[r]elief … is available only if the
asserted error is jurisdictional or constitutional”).
2. Three circuits, in contrast, will recognize a
claim of new evidence demonstrating actual innocence
in a coram nobis petition. The Eighth Circuit, for example, has rejected the argument “that coram nobis
relief is not available for claims based on newly discovered evidence.” Kandiel v. United States, 964 F.2d 794,
797 n.1 (8th Cir. 1992) (per curiam). The Kandiel court
ultimately did not grant the writ, finding that the new
evidence did not “constitute[] positive proof” of the petitioner’s innocence. Id. at 797. But the Eighth Circuit
confirmed that coram nobis relief is available for compelling new evidence that undermines the factual basis
for a conviction. Id. at 797 n.1; see also Azzone v. United States, 341 F.2d 417, 419 (8th Cir. 1965) (per curiam)
(considering petitioner’s claim “that he has newly discovered evidence which proves that he did not violate
18 U.S.C.A. § 1073”).
Similarly, the Tenth Circuit has not limited coram
nobis relief to cases in which the petitioner can demonstrate a constitutional or jurisdictional error, stating
more generally that “a writ of error coram nobis is
25
available … to correct errors resulting in a complete
miscarriage of justice, or under circumstances compelling such action to achieve justice.” United States v.
Bustillos, 31 F.3d 931, 934 (10th Cir. 1994). And the
Tenth Circuit has held that “[a] colorable showing of
factual innocence demonstrates a fundamental miscarriage of justice.” United States v. Miles, 553 F. App’x
846, 848 (10th Cir. 2014) (internal quotation marks
omitted). Indeed, in Bustillos, the court rejected the
petitioner’s claim that an insufficient factual basis was
presented at the plea hearing to support his guilty plea
because, the court stated, he had “not assert[ed] his innocence of the charge to which he pleaded guilty.” 31
F.3d at 934.
The Sixth Circuit’s standard for coram nobis relief
also permits the writ to issue upon a showing of compelling new evidence. The court has stated that to obtain relief, a coram nobis petitioner must show “(1) an
error of fact; (2) unknown at the time of trial; (3) of a
fundamentally unjust character which probably would
have altered the outcome of the challenged proceeding if
it had been known.” United States v. Johnson, 237
F.3d 751, 755 (6th Cir. 2001) (emphasis added); see also
Barrow v. United States, 455 F. App’x 631, 634-637 (6th
Cir. 2012) (affirming denial of writ because petitioner
did not present new evidence that would have “altered
the outcome of the entire proceeding”).
B. Dr. Harkonen’s Compelling New Evidence Of
Actual Innocence Would Have Warranted Relief In The Sixth, Eighth, And Tenth Circuits
The Ninth Circuit found Dr. Harkonen’s evidence
of actual innocence “compelling.” App. 2a. That finding
was plainly correct. The government told the jury that
the press release was false because the trial results
26
were not statistically significant. The ASA Principles
are squarely to the contrary: P-values are not a measure of a proposition’s truth or falsity. A p-value is
“weak evidence” insufficient to determine the clinical
meaning of study data. ASA Principles 132. Indeed, no
“[s]cientific conclusions” at all—much less a mandate to
ignore all evidence from a study—should be “based only on whether a p-value passes a specific threshold.”
Id. at 131.11
Absent the government’s statistical theory that the
p-value precluded drawing any conclusions from the
study data, the statements in the press release were
uncontroversial. In two consecutive studies—the longterm follow-up to the Phase II trial, and the Phase III
trial InterMune conducted—patients taking Actimmune were much more likely to survive than those taking placebo. The trend was large, and continued for
years in both trials.
Such “consistency across more than one study” is
(unlike p-values), highly relevant in assessing causation, since it “greatly reduc[es] the possibility that a
biased, chance, site-specific, or fraudulent result will
lead to an erroneous conclusion that a drug is effective.” FDA, Guidance for Industry: Providing Clinical
Evidence for Human Drug and Biological Products 5
(1998), at https://tinyurl.com/ya2aentj; see also Matrixx, 563 U.S. at 41 (“‘consistency of findings across
available data sources’” is relevant in assessing causation). Moreover, the study was well designed, there
11
Many statisticians now conclude that “[t]here are no good
reasons nor good ways to use p-values,” and that reporting them
in scholarly journals should be proscribed. Briggs, The Substitute
for p-Values, 112 J. Am. Stat. Ass’n 897, 897 (2017).
27
appeared to be a dose-response relationship,12 the survival benefit appeared to increase over time, and there
was a plausible biologic basis to believe Actimmune—
which had known anti-fibrotic properties—would work
for IPF patients. See Matrixx, 563 U.S. at 41 (discussing role of “‘dose-response,’” “‘temporal relationship,’”
and “‘biologic plausibility’” in assessing causation);
ASA Principles 132 (discussing role of “good study design” in assessing causation). In short, the inferences
stated in the press release are amply supported by the
kinds of evidence that scientists should rely on. The
government told the jury something very different. It
insisted that lack of statistical significance, measured
by p-values, was enough to make the results unreliable
and the press release fraudulent. But the ASA Principles now establish p-values have no magic power to
outweigh compelling evidence of causation; they are
instead evidence too “weak” to form the basis for “scientific conclusions.” ASA Principles 131-132.
Dr. Harkonen’s claim of new evidence is thus quite
unlike, and much stronger than, the usual case in which
new evidence casts doubt on the facts proven at trial.
The new evidence in this case demonstrates that the
prosecution’s own version of the facts does not establish the crime alleged. This is not a case where Dr.
Harkonen is relying on after-the-fact affidavits, or witness recantations, or belatedly challenging who said
what to whom. Rather, the new evidence establishes
that the metric by which the government asked the jury to judge the truth or falsity of Dr. Harkonen’s undisputed statements in the press release was itself flatly wrong. The jury was told that because the study’s p12
See 3/23/2009 Topel Decl. Ex. 1 at 1 (D. Ct. Dkt. 89-1).
28
values failed to hit a “magic” number, no conclusions
could be drawn from the study; on that premise, the jury had little choice but to find the press release inaccurate. Had the jury instead been told that p-values “do[]
not provide a good measure of evidence regarding a
model or hypothesis,” ASA Principles 132, and that scientists instead focus on issues such as consistency of
results across trials, the jury would have had no option
other than to acquit.
Because the new evidence in this case undermines
the government’s entire theory of guilt, and not just
one piece of evidence, Dr. Harkonen has established his
actual innocence under any possible standard. Cf. Herrera v. Collins, 506 U.S. 390, 417 (1993) (assuming that
“a truly persuasive demonstration of ‘actual innocence’
made after trial would render the execution of a defendant unconstitutional”); id. at 429 (White, J., concurring) (“To be entitled to relief, … petitioner would at
the very least be required to show that based on proffered newly discovered evidence and the entire record
before the jury that convicted him, ‘no rational trier of
fact could find proof of guilt beyond a reasonable
doubt.’” (alterations omitted)); id. at 444 (Blackmun, J.,
dissenting) (a petitioner would be entitled to habeas
relief if he could “show that he is probably actually innocent, in light of all the evidence”). The new evidence
conclusively establishes that the government was relying on faulty scientific theories in arguing that rules of
statistical significance made it impossible to draw any
conclusions from the GIPF-001 study. No rational jury
could convict Dr. Harkonen in light of the ASA’s rejection of the government’s false statistical theory.
As the Ninth Circuit recognized, there is compelling evidence of innocence. In the Sixth, Eighth, and
Tenth Circuits, that would be enough to warrant coram
29
nobis relief. By the fortuity of having been tried in the
Ninth Circuit, however, Dr. Harkonen remains convicted of a crime that the evidence now indicates he did not
commit. That ought to be an intolerable result.
C. The Ninth Circuit Erred By Denying Coram
Nobis Relief Despite Finding Dr. Harkonen’s
New Evidence “Compelling”
As explained above (at 21), coram nobis has been
used to remedy a wide range of errors in criminal proceedings and generally “under circumstances compelling such action to achieve justice.” Morgan, 346 U.S.
at 511. Thus, the Court has “found that a writ of coram
nobis can issue to redress a fundamental error” as well
as “mere technical errors.” Denedo, 556 U.S. at 911. In
Morgan, for example, the Court held that coram nobis
was available to review a claim of deprivation of counsel in violation of the Sixth Amendment. 346 U.S. at
512-513. Coram nobis has also been used to examine
claims of insanity, id. at 510, that the government coerced witnesses to commit perjury, id., and ineffective
assistance of counsel, Denedo, 556 U.S. at 907.
There is no error more fundamental than the conviction of an innocent person. The Court should take
this opportunity to hold that coram nobis relief is available to a petitioner who has made a compelling showing
of actual innocence, at least where (as here) that showing is based on a new scientific consensus.
In
Kuhlmann v. Wilson, a plurality of the Court recognized that “a prisoner retains a powerful and legitimate
interest in obtaining his release from custody if he is
innocent of the charge for which he was incarcerated.”
477 U.S. 436, 452 (1986). Thus, a defendant who shows
that “a constitutional violation has probably resulted in
[his] conviction” though he is “actually innocent” can
30
show a “fundamental miscarriage of justice” that will
excuse any procedural default of a claim on habeas review. Schlup v. Delo, 513 U.S. 298, 321 (1995) (internal
quotation marks omitted). The fundamental miscarriage of justice is just as clear here, and warrants extending collateral relief to a defendant who can demonstrate his innocence without regard to whether there
was also independent constitutional error in the proceedings leading to his conviction.13
In Herrera v. Collins, the Court assumed without
deciding that, at least in a capital case, “a truly persuasive demonstration of ‘actual innocence’ made after trial would render” punishment “unconstitutional, and
warrant federal habeas relief.” 506 U.S. at 417. Although the Court only assumed this rule, a majority of
the justices agreed that a capital prisoner’s demonstration of actual innocence would be grounds for relief
from the conviction. Justice O’Connor, joined by Justice Kennedy, wrote in a concurrence that “the execution of a legally and factually innocent person would be
a constitutionally intolerable event.” Id. at 419. Justice
Blackmun, joined by Justices Stevens and Souter,
meanwhile, wrote in dissent that “if a prisoner can
13
In recognizing that a showing of actual innocence can excuse a procedural default, the Court adopted a standard proposed
by Judge Friendly. Kuhlmann, 477 U.S. at 454 (plurality opinion)
(citing Friendly, Is Innocence Irrelevant? Collateral Attack on
Criminal Judgments, 38 U. Chi. L. Rev. 142 (1970)); see also
Schlup, 513 U.S. at 328 (adopting Judge Friendly’s description of
the actual-innocence inquiry, that the habeas court must determine the petitioner’s innocence “‘in light of all the evidence’”).
Judge Friendly recognized that substantive claims of actual innocence should be considered on collateral review, even without an
independent constitutional error. 38 U. Chi. L. Rev. at 159 n.87,
160.
31
show that he is probably actually innocent, in light of all
the evidence, then he has made ‘a truly persuasive
demonstration [of innocence],’ … and his execution
would violate the Constitution.” Id. at 444.
While the consequences of error are obviously
uniquely high in the capital context, a petitioner who
has completed his term of incarceration maintains a
strong interest in avoiding the lingering consequences
of a conviction for an offense of which he is actually innocent. Subsequent convictions may carry heavier
penalties, Morgan, 346 U.S. at 512-513; a non-citizen
may face deportation, Denedo, 556 U.S. at 907-908; and
the individual may be deprived of certain civil rights,
such as the ability to serve on a jury, vote, or hold office, Fiswick v. United States, 329 U.S. 211, 222 & n.10
(1946). A criminal record may also limit or prohibit individuals from accessing employment, occupational licensing, housing, education, adoption services, and other opportunities.14 There is no legitimate reason to impose such burdens on a person who has demonstrated
factual innocence. Indeed, doing so only undermines
public confidence in the judicial system.
Where the new evidence demonstrating the defendant’s innocence consists of changed science, that
new evidence often will not be available for many years
following the verdict because it takes time for science
to change and new theories to find acceptance in the
scientific community. By that time, other remedies—
such as motions for a new trial—may no longer be
available. Yet that kind of new evidence can be among
14
Dr. Harkonen, for example, is subject to a lifetime bar from
working for any company seeking FDA approval for a product.
32
the most compelling because it demonstrates that the
evidence the jury relied upon does not prove what the
government argued or expert witnesses testified that it
proved. Moreover, where, as here, scientific developments discredit the entire premise of the government’s
case, justice requires a mechanism for the defendant to
challenge his conviction.
The Court has recognized that the conviction of an
innocent person is a “fundamental miscarriage of justice.” Schlup, 513 U.S. at 321. There was just such a
miscarriage here. The government told the jury that
any conclusion drawn from the study was false because
the study had a p-value greater than .05. But we now
know that it was the government’s statistical premise—not the press release’s conclusions—that was
false. The writ of coram nobis is an appropriate mechanism to obtain relief from such errors. See, e.g., United
States v. Ridings, 569 F. App’x 73, 75 (3d Cir. 2014)
(per curiam) (denying coram nobis because petitioner
had not “made the necessary showing of a ‘complete
miscarriage of justice’”); Jimenez v. Trominski, 91 F.3d
767, 768 (5th Cir. 1996) (coram nobis “will issue only to
correct errors resulting in a complete miscarriage of
justice”); Bustillos, 31 F.3d at 934 (“a writ of error coram nobis is available only to correct errors resulting
in a complete miscarriage of justice, or under circumstances compelling such action to achieve justice”). The
Court should take this opportunity to confirm that coram nobis can be used to grant relief from a conviction
for a defendant who is able to make a compelling showing of innocence based on new evidence.
33
II. THIS CASE PRESENTS A GOOD VEHICLE TO ADDRESS
THE STANDARD FOR CORAM NOBIS RELIEF
This case presents an ideal vehicle to address the
question presented because the Court need not resolve
any factual issues to reach that question. In particular,
the Ninth Circuit has already found Dr. Harkonen’s
new evidence of actual innocence “compelling.” App.
2a. It denied relief because Dr. Harkonen had not
pointed to any “change in controlling law.” Id. The
question whether new evidence of actual innocence justifies coram nobis relief is thus cleanly presented here,
and requires no fact finding from this Court.
Many coram nobis cases raising claims of actual innocence, by contrast, will raise subsidiary factual questions. Which pieces of evidence or what witnesses’ testimony are called into doubt by the new evidence? Was
the affected evidence harmless? Would the unaffected
evidence have been sufficient to convict? But the Court
need not address any of those questions in the present
case because the new scientific evidence undermines
the government’s entire theory of guilt. Nor need the
Court worry that it will not be able to reach the question presented because Dr. Harkonen could not demonstrate his innocence under the appropriate standard in
any event. Because the scientific consensus undermines the theory of conviction, Dr. Harkonen can
demonstrate his actual innocence, however stringent
the standard.
34
CONCLUSION
The petition for a writ of certiorari should be
granted.
Respectfully submitted.
JOSHUA M. KOPPEL
ALAN E. SCHOENFELD
Counsel of Record
WILMER CUTLER PICKERING
PETER G. NEIMAN
HALE AND DORR LLP
1875 Pennsylvania Ave., NW WILMER CUTLER PICKERING
HALE AND DORR LLP
Washington, DC 20006
7 World Trade Center
250 Greenwich Street
New York, NY 10007
(212) 937-7294
alan.schoenfeld@wilmerhale.com
OCTOBER 2018
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.