science-of-solitary.html.pm

#lang pollen

◊define-meta[page-title]{The science of solitary confinement}
◊define-meta[original-date]{2019-02-09}
◊define-meta[edited-date]{2019-06-25}

◊declare-work[#:type "magazine/news" #:author "Treacy Ziegler" #:title
"Memory of Space" #:publication "The Prison Arts Coalition" #:date "15
October 2014" #:url
"https://theprisonartscoalition.com/2014/10/15/memory-of-space/" #:id
"Ziegler1" #:short-form "Ziegler, \"Memory of Space\""]

◊declare-work[#:type "magazine/news" #:author "Treacy Ziegler" #:title
"The art of absolute loneliness" #:publication "Broad Street Review"
#:date "3 June 2014" #:url
"https://www.broadstreetreview.com/art/the-art-of-absolute-lonelines"
#:id "Ziegler2" #:short-form "Ziegler, \"The Art of Absolute
Lonliness\""]

◊declare-work[#:type "magazine/news" #:publication "Quirks and Quarks,
CBC" #:date "8 February 2019" #:title "Months locked in a tiny box —
how solitary confinement can erode mental health" #:url
"https://www.cbc.ca/radio/quirks/feb-9-2019-psychology-of-solitary-confinement-mind-over-genes-genocide-and-climate-change-and-more-1.5008739/months-locked-in-a-tiny-box-how-solitary-confinement-can-erode-mental-health-1.5008744"
#:id "Quirks"]

◊declare-work[#:type "legal-case" #:title "British Columbia Civil
Liberties Association v Canada (Attorney General)" #:citation "2018
BCSC 62" #:url
"https://www.canlii.org/en/bc/bcsc/doc/2018/2018bcsc62/2018bcsc62.html"
#:id "BCCLA" #:short-form "*BCCLA v Canada*"]

◊declare-work[#:type "legal-case" #:title "British Columbia Civil
Liberties Association v Canada (Attorney General)" #:citation "2019
BCCA 228" #:url
"https://www.canlii.org/en/bc/bcca/doc/2019/2019bcca228/2019bcca228.html"
#:id "BCCLA-CA" #:short-form "*BCCLA v Canada* (CA)"]

◊declare-work[#:type "legal-case" #:title "Canada (Attorney General) v
Bedford" #:citation "2013 SCC 72" #:url
"https://scc-csc.lexum.com/scc-csc/scc-csc/en/item/13389/index.do"
#:id "Bedford"]

◊fig[#:src "assets/raymond-palmore-cell.jpg"]{Used with permission
from Raymond Palmore. Drawn during his time at Corcoran State
Prison.◊note{Read how art can help people in solitary confinement
express their experiences. ◊see["Ziegler1"] ◊see-also["Ziegler2"]}}

CBC's ◊em{Quirks and Quarks} had a segment this week about the
psychology of solitary confinement.◊note-cite["Quirks"] Bob McDonald
interviewed Dr. Craig Haney, one of the expert witnesses in the BC
Supreme Court case that resulted in portions of Canada's
solitary-confinement regime being struck down as
unconstitutional.◊note-cite["BCCLA"]◊note{Canada appealed that
result. The appeal was heard on November 13, 2018. The decision had
not been announced at the time of this blog post's original
publication, but on June 24, 2019, the Court of Appeal for British
Columbia held that "a legislative provision that authorizes the
prolonged and indefinite use of administrative segregation in
circumstances that constitute [solitary confinement]" are an
unjustified violation of s 7 of the ◊em{Charter}. The Court of Appeal
accepted all of the trial judge's factual findings, which are the
subject of this blog post. ◊cite["BCCLA-CA"]}

This is a short comment on that case, ◊em{British Columbia Civil
Liberties Association v Canada (Attorney General)} (◊em{BCCLA v
Canada}), focusing on the judge's handling of the scientific evidence.

◊heading{Context}

This case turned on the judge's understanding of nuanced scientific
arguments between experts with competing opinions. The scientific
evidence was essential to the judge's finding that the way that we
currently use solitary confinement "places all Canadian federal
inmates subject to it at significant risk of serious psychological
harm, including mental pain and suffering, and increased incidence of
self-harm and suicide."◊note-cite["BCCLA" #:pinpoint "para 247"]

I did not listen to or read transcripts of the testimony of the
scientific experts in this case. This comment is based only on the
presentation of the evidence by Justice Leask in his written
decision. In the following, I assume that he has given a fair
presentation of that evidence. I am not a psychologist, but I am a
scientist. I understand evidence, experiments, statistical tests and
power, causation, and the strengths and weaknesses of meta-analyses
and systematic reviews.

◊heading{My take on the evidence as presented}
◊sub-heading{Plaintiff's experts}

The evidence supports a claim that solitary confinement causes various
harmful effects.◊note-cite["BCCLA" #:pinpoint "paras 163--194"] There
is a large body of clinical and experimental research establishing
that when you restrict a person's environmental and social
stimulation, they suffer. Dr. Grassian has observed, in prisoners in
solitary confinement, a rare confluence of psychiatric symptoms which
we already know can be caused by restricting environmental and social
stimulation. There is a correlation with the onset of the
intervention: introduction of solitary confinement predictably shifts
EEG patterns towards that characteristic of stupor and
delirium. Psychiatric disturbances spring up anew when prisoners are
segregated. When the segregated confinement is prolonged, there is a
higher incidence of psychosis.

Further (and this relates to how we should view the government's
evidence), the nature of solitary confinement and the psychiatric
symptoms make self-reporting a poor measure of effects. Prisoners will
avoid acknowledging psychological harm out of spite against
authorities whom they see as trying to "break them down". Psychotic
disturbances often have a dissociative character: individuals often do
not recall events which occurred during confusional psychosis.

Dr. Grassian's evidence must be seen in the context of the ◊a[#:href
"https://en.wikipedia.org/wiki/Bradford_Hill_criteria"]{Bradford Hill
criteria} (useful for assessing causation from epidemiological
evidence). In the case of solitary confinement, it would not be
ethical to perform a randomized controlled trial. The government's
experts, Dr. Mills and Dr. Gendreau, point to the inability to deduce
causation from this research. While this may be true, due to the lack
of direct experimental evidence, the Bradford Hill criteria give
guidance for making (abductive) inferences in spite of this.

The harmful effects are consistent across populations, the effects
happen after the purported cause, there is a biological gradient
(longer exposure is associated with higher incidence), there is a
plausible mechanism for the causation, and the hypothesis is
consistent with laboratory experiments on deprivation of environmental
and social stimulus. This confluence of evidence supports an abductive
inference that solitary confinement causes the harmful effects.

◊sub-heading{Government's experts}

The conclusions drawn by the government experts are not consistent
with the evidence they claim to base it on. They look to low-power
observational studies that used subjective self-reporting as a
measure. These studies have a poor ability to detect an effect even if
an effect were to be present. Because these studies failed to find an
effect, the government experts concluded that there was no effect
("segregated inmates without mental illness do not experience
debilitating psychological or psychiatric symptoms due to their
placement in segregation"), or an effect of lower magnitude than
argued by the plaintiffs ("that it is much milder than that predicted
by the plaintiffs' experts"), but these conclusions cannot be drawn
from a negative results in low-powered studies.

Suedfeld ◊em{et al} (1982), a study relied upon by the government,
gives contradictory and weak evidence.◊note-cite["BCCLA" #:pinpoint "para
200"] "A relationship was found between the length of stay and
measures of depression and hostility" (a dose-response relationship;
one of the Bradford Hill criteria that points in favour of finding an
effect). But because of the small magnitude of the observed effect,
the authors of that study gave a measured conclusion: it "did not
support the claim that solitary confinement ◊elide was overwhelmingly
aversive, stressful, or damaging to inmates." Again, this is just
failure to find evidence for the effect, not positive evidence against
the effect.

The government experts pushed the court to rely on longitudinal
studies, which are more like an experiment in that they observe what
happens before and after the onset of an intervention. They believe
the "Zinger study" to be worth significant weight.◊note-cite["BCCLA"
#:pinpoint "paras 203--204"] The population for this study was 83
segregated plus 53 general-population prisoners. The segregated
population initially reported more depressive symptoms, but not on a
measure of hopelessness. Both segregated and non-segregated prisoners
improved over the 60 days of the study on measures of depression,
anxiety, hopelessness, and psychological adjustment. The fact that
both populations improved over the 60-day study is interesting
scientifically. It puts into question whether the fact of observation
◊u{on its own} had an effect on the population or on their
reporting. Perhaps when prisoners receive increased attention from
others regarding their well-being and are asked to perform meaningful
introspection and tell someone about, they feel better. Or maybe the
participants felt a subconscious pressure to report
improvement. Without blinding, this kind of observer effect is common.

Of course, a longitudinal study that tracks the response to an
intervention over time has the potential to be more useful
scientifically than a set of clinical observations as presented by
Dr. Grassian, but the Zinger study has methodological weaknesses that
make its negative finding unsurprising and definitely not positive
evidence for the claim that there is no effect.

Another study that the government experts emphasized was a pair of
meta-analyses by Morgan and Gendreau. Meta-analyses (and systematic
reviews especially) can be a great tool for evaluating an entire body
of evidence generated by many separate research teams. By pooling
together participants, meta-analyses can find effects when individual
studies were too small to find one. They can tease out patterns of a
decreasing effect size with increased rigor (a marker of there being
only a small or no effect). They can detect publication bias (where
research teams across an entire field withhold studies that don't show
statistically significant effect).

Paragraph 206 summarizes the Morgan and Gendreau studies: "the two
studies ◊elide found that the differences between segregated and
non-segregated inmates ranged from small to moderate on a wide range
of psychological indicators."◊note{Isn't this actually a finding in
favour of an effect?} But, they also found that "the size of these
differences were significantly reduced when only the studies rated as
more scientifically rigorous were considered." As I mentioned above,
this pattern in a body of research is an indicator of a small or
non-existent effect. But, using this move depends on an accurate
assessment of a study's scientific rigor, and Dr. Grassian, one of the
plaintiff's experts, criticized the selection criteria for the Morgan
study.◊note-cite["BCCLA" #:pinpoint "para 234"]

Regardless, the government expert concludes: "the findings do not
support the notion that, as a group, inmates in segregation suffer
◊elide when compared with inmates outside of segregation."

Two things are important from this conclusion. It is a conclusion "as
a group".◊note-cite["BCCLA" #:pinpoint "para 207"] And again, even
taken at face value, it is only a failure to find positive evidence
for the effect, not positive evidence that there is no effect. With an
effect that might vary within specialized subpopulations, and
subjective psychological measures involving self-reporting that the
participants may have motivations or incentives for hiding, it isn't
unexpected that a meta-analysis would find no evidence for an effect.

Last, the government relies on the "Colorado Study". But, there is one
criticism of the study that is so fatal to its relevance that I will
not even present its results. The study relied on self-reporting
scales. But, the corrections staff also kept objective data in the
form of psychiatric crisis records (for self-harm, suicide, psychotic
disturbances). These crisis records squarely contradicted the
prisoners' self-reporting. According to the self-report scales, there
was no deterioration within the segregated population over the course
of 12 months. The crisis records show the opposite. Only 12 episodes
occurred in the first six months; 25 occurred in the second six
months.◊note-cite["BCCLA" #:pinpoint "para 237"]

The methodological weaknesses in both the "Colorado study" and the
"Zinger study" also affect the relevance of the Morgan and Gendreau
meta-analysis because they were both included in the small set of
studies meeting the selection criteria◊note-cite["BCCLA" #:pinpoint
"para 234"], and the "Colorado study" was one of the studies
identified to be weighted as having a "stronger" quality
design.◊note-cite["BCCLA" #:pinpoint "para 243"]

◊heading{Discussion}

This is a difficult scientific question. The plaintiff's experts do
not point to evidence that experimentally establishes a causal link
between solitary confinement and psychological harm. They do, however,
point to a confluence of evidence that, when viewed in light of the
Bradford Hill criteria, strongly suggests an effect.

The evidence presented by the government is based on methods that have
the potential to be more revealing. However, due to methodological
weaknesses involved in the specific applications of those methods,
this evidence is actually not revealing.

I have emphasized the distinction between (1) failing to find an
effect and (2) finding positive evidence that there is no effect. Of
course, one cannot prove a negative; failure to find positive evidence
that there is no effect should not be fatal to any
counter-evidence. But, when you present a study that you intend to
negate the claim that there is an effect, that study should at least
have been set up to be able to find an effect if the effect
exists. That is, the study should be high powered. Failure to find an
effect in a high power study can be very compelling evidence.

That is a tricky distinction and I am impressed that Justice Leask saw
it. In this case, evidence that was obtained by methods that are
generally less robust was owed more weight than evidence obtained by
methods that are generally more robust.◊note-cite["BCCLA" #:pinpoint
"para 254"]

In a Section 7 challenge◊note{That is, when a plaintiff challenges a
government law or act on the basis that it infringes upon their
◊a[#:href
"https://www.justice.gc.ca/eng/csj-sjc/rfc-dlc/ccrf-ccdl/check/art7.html"]{Section
7} right to life, liberty, or security of the person.}, the plaintiffs
only need to show a "sufficient causal connection" between the
government action and the harm. They don't need to show that solitary
confinement is "the only or the dominant cause" of the harms. And,
they only need to establish "a reasonable inference, drawn on a
balance of probabilities."◊note-cite["Bedford" #:pinpoint "para 76"]
Justice Leask found that the plaintiffs met that burden.

◊heading{Related media}

◊format-work[#:type "magazine/news" #:publication "BC Civil Liberties
Association" #:title "End Solitary Confinement -- BobbyLee's Story"
#:date "26 May 2016" #:url
"https://www.youtube.com/watch?v=lfnDpXHgZio"]

◊format-work[#:type "book" #:author "Susan Haack" #:title "Defending
Science -- within Reason" #:publisher "Prometheus Books" #:year "2007"
#:publisher-location "Amherst, New York"]

◊format-work[#:type "article" #:author "Susan Haack" #:title "Of
Truth, in Science and in Law" #:journal "Brooklyn Law Review" #:volume
"73" #:issue "3" #:year "2008" #:url
"https://brooklynworks.brooklaw.edu/cgi/viewcontent.cgi?article=1315&context=blr"
#:pages "985--1008"]

◊format-work[#:type "article" #:author "Craig Haney" #:title
"Restricting the Use of Solitary Confinement" #:journal "Annual Review
of Criminology" #:volume "1" #:pages "285--310" #:url
"https://www.annualreviews.org/doi/10.1146/annurev-criminol-032317-092326"
#:year "2017"]

◊cite["Ziegler1"]

◊cite["Ziegler2"]