Medical Students' Attitudes toward Torture, Revisited.

This paper reports the findings of a survey of medical students' attitudes toward torture and discusses variables that may correlate with those attitudes. In late 2016, 483 enrolled medical and MD-PhD students at the Weill Cornell Medical College received an anonymous, institutional review board-approved survey that included questions about torture and its effectiveness, demographic questions, inquiries about personal experiences of harassment or discrimination, and questions regarding engagement in human rights activities. Some questions were drawn from a 2008 University of Illinois survey of medical students' attitudes toward torture, the only prior such survey at a US medical university. Of the 483 students who were contacted, 121 (25%) returned completed questionnaires, with responses indicating strong opposition to torture and skepticism about its usefulness. Respondents expressed greater opposition to torture in this survey than those who participated in the 2008 survey. Respondents' involvement in Weill Cornell's human rights program was associated with significantly stronger opposition to torture, while personal experiences of harassment were associated with a trend toward weaker opposition to torture. Respondents' answers closely approximate the clearly stated ethics of the profession, suggesting that human rights education during medical school may contribute to the development of proper values in young physicians even before they proceed into practice.

Introduction

Physicians are subject to broadly accepted standards governing ethical and professional conduct. The condemnation of physicians’ direct or indirect participation in torture is one example. This is asserted by the World Medical Association, American Medical Association, American Psychological Association, World Psychiatric Association, American College of Physicians, American Psychiatric Association, and innumerable countries’ medical societies. However, as Steven Miles reminds us, many medical personnel continue to participate in and condone torture: “Many torture survivors report that clinicians monitored their mistreatment. The presence of a physician during torture compounds the victim’s suffering by emphasizing that even the humanity of medicine is turned against the prisoner.” This paper explores the knowledge and attitudes of medical and MD–PhD students at one institution on matters relevant to the practice of torture and physician complicity with torture.

Many physicians and medical personnel who participate in torture do so as they succumb to conflicts between professional ethics and the demands of their work, superiors, and peers. The authorization of torture during the Bush administration by government authorities (Departments of State, Defense, and Justice) enabled its legitimization through the military chain of command. In military settings, many may fear retaliation or disciplinary consequences should they fail to follow orders. From a psychosocial perspective, Myles Balfe identifies factors that may contribute to a physician’s participation in torture, such as the passionate assumption of the need to defend the United States from grave danger. Balfe further notes that the capacity for rationalization (such as euphemistically referring to torture as “enhanced interrogation procedures”) and cognitive distortions (such as the belief that without medical supervision, greater harm might ensue) are clear factors. Additionally, the splitting of roles such that responsibility can be diffused among many participants, each believing that their individual contribution to torture was minor or insignificant, creates an environment that enables medical personnel’s participation in torture.

The present study explores future US physicians’ attitudes toward the permissibility and utility of torture, as well as their beliefs about physicians’ participation in torture. A previous survey of medical students’ attitudes toward torture, conducted in 2008 at the University of Illinois College of Medicine-Chicago (UIC), provided a precedent for our project. That six-question survey of 336 medical students across the four years of study revealed a level of support for torture that the authors reported as distressing, given medical associations’ widespread condemnation of physicians’ participation in torture. Specifically, the authors found that 35% of their sample would condone torture under certain circumstances; 24% agreed with the use of torture if a chance to elicit life-saving information existed; and 22% agreed that it was permissible for physicians to treat individuals so that torture could be initiated or continued. These and other findings led the authors to recommend the implementation of medical school curricular assessments to address ignorance or attitudes among students that are at odds with the universally and clearly stated ethics of the profession.

International surveys of medical students’ attitudes toward torture or the mistreatment of prisoners have generally revealed a somewhat greater tolerance for such practices among medical students than in the UIC survey. For example, in a study conducted in Mauritius, 37.4% of surveyed medical students were in favor of beating individuals in police custody to obtain information, and in a study of New Delhi medical students, nearly 30% of respondents indicated approval of this practice. Taken together, these studies demonstrate the importance of assessing medical students’ attitudes toward such a major human rights issue.

The UIC survey has not, to our knowledge, been repeated at any other medical university in the eight years since it was administered. In addition to replicating the prior study, our project attempts to delineate personal and demographic factors that may be associated with attitudes toward torture. Clearly a question of great political, ethical, and medical sensitivity, this topic invites medical training institutions to examine curricula and human rights educational efforts. It is important that medical students have solid grounding on matters such as torture and the obligations of the profession before they begin to practice independently in society.

Subjects and methods

The Institutional Review Board of the Weill Cornell Medical College in New York City approved this study. All enrolled four-year medical and MD–PhD students (483 total) received an email in November 2016, shortly after the conclusion of the national elections, containing links to a 28-item questionnaire (see Appendix). Their participation in the survey was anonymous and completely voluntary. Appropriate encryption procedures were employed to ensure that the identification of participant/ non-participant status was impossible. Institutional review board-approved consent was obtained from all participants through encrypted procedures, and participants could access the survey only after first providing their consent.

The survey instrument contained 10 items addressing specific torture activities, justifications for torture, and ways that physicians might participate in torture. Five of these items matched questions from the 2008 UIC study. To identify factors that might influence students’ attitudes, the survey also included demographic inquiries related to age, gender, ethnicity, religious affiliation, sexual orientation, and stage of medical school training; questions examining individual and family histories of exposure to trauma, harassment, or discrimination; and items surveying respondents’ participation in student human rights activities at Weill Cornell, their familiarity with certain human rights statistics, and their opinions about human rights curricula in medical school. Participation in the survey was initiated by 146 students, but 19 were excluded because they did not complete any sections. Responses from six more participants were discarded, either because those respondents failed to complete the 10 torture-specific items or because they neglected to indicate their gender or age. The final sample size was thus 121 respondents.

We employed two methods to identify associations between participants’ attitudes toward torture and their responses to the non-torture questions. In the first strategy, we used an aggregate metric, termed the “attitude toward torture scale” (ATS), to compare pools of participants grouped by their responses to individual non-torture questions. To calculate the ATS, we created a standardized scale of 0–4 for each of the 10 torture-specific items. Higher scores on this scale correspond to greater support for torture; “strongly agree,” for example, was coded as a 4 if this response indicated the strongest support for torture (questions 1.1a-c, 1.2-1.5), whereas “strongly agree” was coded as a 0 if it instead corresponded to the greatest opposition to torture (questions 1.6-1.8). An individual’s ATS was then calculated by summing that person’s scores on these 10 questions. The ATS therefore ranged from 0 to 40, with a neutral position represented by 20. We performed Mann-Whitney U tests to compute p-values for the differences found between mean ATS values of paired subgroups. We employed the Bonferroni method to account for multiple hypothesis testing; differences in mean ATS values were considered statistically significant if their associated p-values fell below 0.05/N, in which N is the number of hypotheses tested. The ATS metric was developed solely for this study and has not been validated elsewhere.

Considering the 10 torture-specific items and ATS as dependent variables, the second method entailed creating a statistical model for each dependent variable as follows. We performed ordinal regressions between the dependent variable and each independent variable using the polr function of the MASS package in R. The false discovery rate was controlled at a level of 0.1 using the Benjamini-Hochberg method to account for testing multiple hypotheses. Those independent variables that did not survive the multiple hypothesis correction were then excluded. A final ordinal regression was performed on the dependent variable and all the remaining significant independent variables to arrive at a multivariate model for the dependent variable. Multivariate models were not computed for torture questions 1.6 and 1.7, as these questions were not found to have significant dependences on any of the independent variables. Lastly, to investigate associations between individuals’ attitudes toward various facets of torture, we calculated the Pearson correlation coefficient (r) for each pair of torture-specific questions.

Results

Of the 483 students who received the survey, 25.1% responded (Table 1). The lowest response rates were seen from the third-year medical school class and from MD–PhD students engaged in the PhD phase of their training. More females than males (75:46) completed the survey (the entire student body annually approximates a 50:50 male:female ratio).

Table 1.

Respondent demographics

Year[†]	Number of respondents	Response rate
1st	35	34.7%
2nd	37	40.2%
3rd	14	12.8%
4th	24	22.4%
MD–PhD	11	14.9%
Total	121	25.1%

The questionnaire began with ten items surveying respondents’ attitudes toward torture, five of which were based on questions from the 2008 UIC study. Using a five-point Likert scale, participants were asked to indicate their agreement or disagreement with each statement. For ease of comparison between the studies at UIC and the Weill Cornell Medical College (WCM), the “strongly disagree” and “disagree” responses were pooled together, as were the “strongly agree” and “agree” responses.

A larger proportion of those surveyed in the present study expressed opposition to torture than in the 2008 UIC study (Figure 1). The most striking difference was seen in response to the statement “It is permissible for interrogators to use psychological intimidation (e.g. mock executions, sexual humiliation, religious humiliation, threatening loved ones)”; here, 93% of WCM respondents disagreed, compared to 30% of those surveyed at UIC (Figure 1A). In addition, 63% of WCM respondents disagreed that “the use of torture can be justified if the information obtained will save lives,” compared to 51% of UIC respondents (Figure 1B). The item “It is permissible for physicians to treat individuals to verify their health so torture could begin or continue” garnered a disagreement rate of 83% at WCM and 58% at UIC (Figure 1C). Furthermore, 83% of WCM respondents agreed that “the use of torture should be prohibited as a matter of state policy,” compared to 64% of UIC respondents (Figure 1D). Finally, for the statement “The use of torture to elicit information is immoral and intrinsically wrong,” the 86% agreement rate among WCM respondents exceeded the rate of 63% reported for UIC respondents (Figure 1E). WCM respondents’ answers to these two final questions were strongly correlated (r = 0.84). Interestingly, the rate of neutral responses was lower among WCM respondents than among UIC students in all but one item: “The use of torture can be justified if the information obtained will save lives.”

Figure 1.

Comparison between survey results from the 2008 study at the University of Illinois College of Medicine– Chicago (UIC) and from the present study at the Weill Cornell Medical College (WCM)

Our study sought to expand upon the 2008 UIC survey in several ways. First, because torture comes in different forms, we asked participants whether it is permissible for interrogators to employ “psychological intimidation (e.g. mock executions, sexual humiliation, religious humiliation, threatening loved ones),” “deprivation or exposure to environmental extremes (e.g. exposure to extreme heat/cold/noise, sensory deprivation, sleep deprivation, starvation, or forced feedings),” or “physical distress or injury (e.g. forced positions, asphyxiation, beating, electrocution).” More than 90% of WCM respondents disagreed that any of these three interrogation classes are permissible (Figure 2A). Approximately 8% of WCM students felt neutral or agreed that psychological intimidation and deprivation or exposure to environmental extremes are permissible interrogation strategies. Interestingly, WCM students were most opposed (97.5%) to interrogators employing tactics that would cause physical distress or injury. Respondents’ answers to these three questions were strongly correlated (r > 0.7).

Figure 2.

Medical students’ attitudes toward torture

Believing that information elicited through torture is reliable will influence individuals’ views on the justification of torture. Therefore, our survey also asked participants whether they agreed that “torture is an effective means of obtaining information”; more than 80% of students disagreed with this statement (Figure 2B). A similar percentage of respondents also disagreed with the premise that “torture can be justified to gain important information” (Figure 2C). Only 60%, however, disagreed with the statement that “torture can be justified if the information obtained will save lives” (Figure 1F), with 27.6% responding with a neutral answer. Participants who disagreed that torture is justified if it yields important information were also likely to disagree that life-saving information justifies torture (r = 0.75). Their responses to these two items, however, correlated less well with their belief in the effectiveness of torture as a means of obtaining information (r = 0.57 and 0.49, respectively).

Third, the 2008 questionnaire explored whether those surveyed believed it was acceptable for physicians to participate in torture. Taking this one step further, our study looked at whether respondents felt that physicians’ participation in torture should warrant punishment. In this regard, 75% of WCM participants agreed that “health professionals who are found to have designed, committed, or otherwise facilitated torture should face disciplinary or legal action” (Figure 2D).

Finally, in an effort to identify factors that may inform individuals’ beliefs about torture, our survey contained a number of additional questions (Appendix). Several items specifically addressed human rights issues, including whether respondents had participated in the existing programs at the medical college. Students involved with the Weill Cornell human rights program (item 5.3 of the survey) had significantly (p < 10-4) lower mean ATS values (see “Subjects and methods” section), indicating greater opposition to torture than those who had not participated (Figure 3 and Table 2). Participants were also asked about individual or familial experiences of harassment or discrimination (item 4.3). A trend toward stronger pro-torture attitudes, as measured by mean ATS values, was seen among individuals who had experienced discrimination or harassment (p = 0.0046), but this trend did not reach statistical significance after applying a Bonferroni correction for multiple hypothesis testing (Figure 3 and Table 2). Strong correlations were seen between ATS values and responses to all of the torture-specific questions, with the strongest being for the item “The use of torture can be justified in order to gain important information” (r = 0.85).

Figure 3.

A comparison of attitude-toward-torture scale values (error bars indicate standard error of the mean, and the asterisk denotes p <10-4)

Table 2.

Mean attitude-toward-torture scale values

Condition	Number of respondents	Mean ATS ± SEM†
All	121	6.43 ± 0.61
WCCHR event[a]	51	3.71 ± 0.62
No WCCHR event[b]	70	8.41 ± 0.88
Harassed[c]	69	7.80 ± 0.91
Not harassed[d]	52	4.62 ± 0.68

SEM = standard error of the mean.

Attended a human rights event hosted by the Weill Cornell Center for Human Rights (WCCHR).

Did not attend or left this question blank.

Respondent or family member experienced harassment (bullying/cyberbullying, stalking, intimidation, threats, etc.).

Was not harassed, was unsure, or left this question blank.

Our multivariate regression analyses identified several additional factors that strongly predicted respondents’ attitudes toward certain torture-specific items. Responses to the statement “Medical schools’ curricula should include mandatory Human Rights coursework” (item 5.1) provided the strongest predictor for when individuals would express weaker opposition to torture. Those who disagreed with this statement were many times more likely to support the use of torture to obtain important information (odds ratio (OR) 42.14, 95% confidence interval (CI) {3.8,1059}) or life-saving information (OR 58.63, 95% CI {3.89,1752}), as well as the use of interrogation techniques that result in physical distress or injury (OR 18.22, 95% CI {2.22,183.7}). Conversely, those who agreed with mandatory human rights coursework were less likely to support the use of psychological intimidation (OR 0.27, 95% {0.09,0.79}), and those who strongly agreed were less likely to support the use of deprivation or exposure to environmental extremes compared to those who disagreed, felt neutral, or did not answer the question (OR 0.17, 95% CI {0.05,0.5}). The conclusions we can draw are limited, however, as only four individuals disagreed that medical school curricula should include mandatory human rights coursework.

Interestingly, students who were trained by the Weill Cornell Center for Human Rights to perform forensic evaluations of asylum seekers (item 5.3) were less likely to agree that obtaining life-saving information justifies the use of torture (OR 0.4, 95% CI {0.20,0.78}) or that it is permissible for physicians to treat individuals so that torture can begin or continue (OR 0.28, 95% CI {0.13,0.61}). We also found that the importance of an individual’s belief system and experiences of harassment or gender-based discrimination yielded predictors of that person’s attitude toward certain torture-specific items (data available upon request). Other factors, including ethnicity and age, were not found to be significant predictors for responses to individual torture questions (additional information available upon request).

Discussion

As part of their education, medical students would benefit from understanding that participation in torture may take active and passive forms. Helping design torture programs that leave no physical evidence, such as those designed by James Mitchell and John Jessen for the US Department of Defense, would be considered active participation. Passive physician participation can include ignoring torture when it occurs, deliberately failing to diagnose injuries caused by torture, and covering up occurrences through non-documentation or the alteration of medical records. When doctors condone and participate in torture, a demoralizing impact is felt both by victims and by staff—and given the respected authority of physicians, a strong message of support for torture may be inferred. The same can be said of physicians’ participation in certain activities in correctional institutions, where human rights abuses often occur and where a doctor’s behavior and attitude may convey tolerance for insensitive, inhumane conduct. Given the parallels between torture and correctional medicine abuses, both topics should be central to any medical school’s human rights curriculum.

Our results support the implementation of human rights educational programs in medical school curricula. However, beyond the one item in our survey that asked about support for disciplinary or legal action for physicians who have participated in torture (Figure 2D), we have not addressed an important aspect of the issue: professional accountability. The settlement, in August 2017, of a lawsuit brought by the American Civil Liberties Union (on behalf of three men who were kidnapped by the CIA in 2002 and tortured according to a protocol designed by psychologists Mitchell and Jessen) sends a clear message: health care personnel who violate professional codes of conduct may indeed be held accountable for their actions. The two psychologists had been paid millions by the CIA to design and implement an “enhanced interrogation program” to deal with post-9/11 terrorism suspects. Whether a physician contributes actively to the conduct of torture, in the manner of Mitchell and Jessen, or passively, as might a prison MD who fails to contravene a course of solitary confinement, awareness that their actions or inaction will be scrutinized may help combat misbehavior.

A recent global survey by the International Committee of the Red Cross reported that 46% of Americans approved of torture to obtain information from enemy combatants, with only slightly more than half indicating that torture was “wrong.” Only Israelis, Palestinians, and Nigerians matched Americans in their endorsement of torture. By comparison, over 80% of Afghans and Colombians surveyed disapproved of torture. The results of our survey of medical students at one American university reflect strongly divergent attitudes toward torture compared to this global survey.

Conclusive findings from our survey are limited by several factors. Our study, like the 2008 survey, was conducted at a single institution. The number of respondents (121) is relatively small, although our response rate of 25.1% is comparable to other survey response rates of physicians and medical students: 24% of medical students and 18.9% of medical residents completed the 2012 Canadian National Physician Survey, and 27% of medical students responded to Australia’s 2013 National Mental Health Survey of Doctors and Medical Students. The more robust response rates of first- and second-year students (34.7% and 40.2%, respectively) suggest that the pre-clinical years, before students disperse to disparate locations with different schedules, might be a more optimal time for surveying students.

The response rate of our study may have been affected by the very subject matter, torture being an uncomfortable topic for many. Questions about personal or familial experiences of discrimination, harassment, being a victim of a crime, or experiencing sexual or physical abuse may similarly have been too off-putting for some students.

Responses to some of our survey items may have been influenced by our detailing specific torture methods that were not made explicit in the UIC questionnaire (Figure 1A). Similarly, we attempted to ascertain whether respondents believed that torture is an effective way to obtain information, a question that was not included in the UIC study.

Although temporal, geographic, and compositional differences limit comparisons of responses to similar and identical questions between the 2008 UIC survey and our own, some limited observations seem merited. A trend toward stronger anti-torture attitudes was seen among students in 2016 (Figure 1). Why might such a trend be taking place? General contributing factors might include formal and informal human rights educational initiatives that have arisen in the intervening years; continued writing about Abu Ghraib and the role of medical personnel; the ongoing horrific human rights crises in Syria and in other countries; increased awareness of all human rights abuses, including torture; and the agitating polemics of the most recent presidential campaign, in which torture specifically and an atmosphere of hostility toward immigrants fleeing oppression generally may have seeped into the consciousness of medical students.

Using the ATS metric, we examined whether students’ participation in the school’s student-run asylum clinic, the Weill Cornell Center for Human Rights, affected their views on torture. Founded in 2010, this voluntary program attracts students interested in human rights activism and trains them in providing pro bono medical, mental health, and gynecologic evaluations to individuals seeking asylum in the United States. Between one-quarter and one-third of Weill Cornell students will have participated in this program by graduation. In the course of these evaluations, students have helped examine more than 300 asylum seekers, the majority of whom are survivors of torture, and have absorbed their histories. In our study, students who had participated in the human rights program had significantly lower (more opposed to torture) mean ATS values than students who had not been involved (Figure 3 and Table 2).

The medical education experience can be isolating for many students. Human rights education, formalized or not, can ameliorate such isolation and can enhance a student’s preparedness for his or her life as a physician. A 2010 study documented deficits in medical students’ knowledge of torture that could be remediated through a structured curriculum. The experiences of Croatian medical students during the Croatian War of Independence (1991–1995) bears on the value of human rights experiences in medical school. As has been described:

Our experience during the five years of the wars… showed that the best option for the medical students was to continue their studies and engage very actively in a number of activities where their education and medical experience were important …Psychologically, intensive engagement in extracurricular activities related to their profession was of enormous benefit to the students … They became a symbol of safety, consolation, and help in many complicated situations.

Of course, torture is only one of many human rights issues that physicians will encounter in their careers. However, learning about torture equips clinicians to attend to victims of other abuses, including victims of human trafficking; domestic abuse; child abuse and child labor; bullying; religious, political, and ethnic persecution; LGBT discrimination; and, as previously mentioned, human rights abuses in prisons. Specifically, being able to elicit painful and difficult histories, developing empathy, and recognizing human resilience are skills that will serve all physicians in all settings.

Gender
Female	75
Male	46
Age (yrs)
Mean	25.4
Range	21–38

Medical school year or MD–PhD students in PhD training

Main text figure	1A	1B	1C	1D	1E	2A	2B	2C	2D
Survey, Part 1 question	1.a	4	5	6	7	1	2	3	8