The Ethics of Withdrawal from Study Participation

Abstract

While considerable attention has been directed toward the informed consent process and ethical recruitment of human subjects into research protocols, relatively less attention has focused on possible ethical concerns pertaining to the withdrawal process and dis-enrollment of participants once recruited. While there are eight main and six additional Federal requirements for informed consent to enter subjects/participants into research studies, two of them pertain to leaving a study. The U.S. Code of Federal Regulations (CFR) regarding voluntary withdrawal from research clearly requires communicating terms of withdrawal and that requests for withdrawal will be granted without repercussions to the participant. The Federal Common Rule requires informed consent documents to include information on withdrawal from research participation: “…the subject may discontinue participation at any time without penalty or loss of benefits to which the subject is otherwise entitled” [45 CFR §46.116(a)(8)]. Additionally, the CFR stipulates that the consent document must address, “when appropriate”: “The consequences of a subject’s decision to withdraw from the research and procedures for orderly termination of participation by the subject” [45 CFR §46.116(b)(4)]. The U.S. Food and Drug Administration (FDA) also maintains identical respective regulations [21 CFR §50.25(a)(8)] and [21 CFR §50.25(b)(4)], although most research conducted under FDA regulations is supported by the pharmaceutical industry. Such regulations make clear that subjects/participants are free to withdraw from study participation at any time. However, the point at which withdrawal procedures become too cumbersome to subjects and are perceived as undue influences to remain in the study remains to be determined.

Little is known about whether consent forms actually adhere to these two Common Rule regulations. Additionally, the kinds of procedures requested and/or required of subjects to withdraw from research have not been examined. The withdrawal section on informed consent forms can include additional stipulations necessary for withdrawal suggesting that subjects cannot simply leave a study at any time. Ascertaining whether consent forms include provisions regarding withdrawal and/or whether such provisions include additional stipulations necessary for withdrawal is important because subjects/participants may perceive them as simply persuasive tactics or as undue influences, and possibly even coercion, to continue to participate in a research study, undergo burdens, and be exposed to unwanted health risks. But the distinction between how subjects perceive and/or experience these concepts is unclear and should be clarified.

Persuasion occurs when “a person must be convinced to believe in something through the merit of reasons advanced by another person” (Beauchamp and Childress 1986:164). Persuasion is an ethically acceptable and common practice among health care professionals in the context of treatment in an effort to ensure that patients undertake the most healthful course of action (e.g., Scales and Miller, 2003). However, there is controversy over the role of persuasion to recruit or retain subjects within clinical research. It is unethical to undermine voluntary consent, but when research is therapeutic and no other options are available off-study, the use of persuasion may be justified (Nelson and Merz, 2002).

Undue influence “occurs through an offer of an excessive, unwanted, inappropriate or improper reward or other overture in order to obtain compliance” (The National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research, 1979). Inducements may become undue influences if the subject is especially vulnerable. (The National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research, 1979). The Belmont Report makes clear that:

“Unjustifiable pressures usually occur when persons in positions of authority or commanding influence – especially where possible sanctions are involved – urge a course of action for a subject. A continuum of such influencing factors exists, however, and it is impossible to state precisely where justifiable persuasion ends and undue influence begins. But undue influence would include actions such as manipulating a person’s choice through the controlling influence of a close relative and threatening to withdraw health services to which an individual would otherwise be entitled” (The National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research, 1979).

In contrast, coercion occurs when “the choice a research participant makes is constrained by another person, such that the research participant has no alternative but to do what the other person wants him or her to do” (Agrawal, 2003:S27). In the context of withdrawing from clinical research, when investigators request or require that subjects undergo additional actions before withdrawing from study participation, subjects may perceive coercion because they may prefer to leave a study but feel that they have no choice but to remain, or the incentive to stay is too high, or the effort to leave is too high.

Certainly, there is a subjective component to assessments of undue influence, and as the Belmont Report states, there is no clear distinction between persuasion and undue influence. How subjects perceive withdrawal stipulations likely depends on various strategies employed by clinical investigators that could potentially make subjects reconsider or feel uneasy about leaving. These strategies may be induced in terms of financial factors or time and effort, and less obvious psychological strategies pertaining to subjects’ sense of obligations to clinical investigators/physicians. The uncertainty over how to interpret withdrawal stipulations in light of concerns over persuasion, undue influence, and coercion warrants further ethical attention in order to make further distinctions and protect human research subjects.

What remains to be determined is the point at which withdrawal stipulations bear undue influence or become potentially coercive to the subject/participant to stay in the study. While this information can only be obtained through interviews with research subjects (which is beyond the scope of this paper), it is worth pointing out the continuum by which withdrawal stipulations may generate concerns about persuasion, undue influences, or coercion based on an increased level of effort expected of subjects for withdrawal. Certainly, the most expedient option is to require (or request) subjects/participants to do nothing other than state their intention in order to withdraw from a study. While this option confers the greatest freedom to subjects/participants, it is associated with other problems, which will be examined below. A slightly more cumbersome withdrawal procedure would be to ask patients their reasons for deciding to withdraw from research. An even more demanding withdrawal procedure would require participants to engage in further behavior and extensive withdrawal procedures, such as additional tests and visits. These options span a continuum along which Principal Investigators (PIs) and their Clinical Investigator (CI) collaborators may request or require additional behaviors of study subjects/participants for which their ethical soundness must be examined.

The occurrence of additional withdrawal stipulations in consent forms may very well be related to different study characteristics. In particular, four study characteristics -- phase of clinical trial, whether the research was therapeutic, study sponsorship, and IRB review level -- should be considered for the following reasons. First, phase of clinical trial is relevant because it may very well be that the risk-benefit ratio differs by trial phase thus requiring different levels of attention to patient monitoring. Second, the therapeutic status is important to consider because studies pose different levels of risk depending on their therapeutic status, which accordingly may require different levels of patient oversight. For example, therapeutic studies generally present higher risks to subjects than nontherapeutic studies. Third, study sponsorship is relevant to how consent forms are written because sponsors vary in the level of oversight of study participants and in the review of consent forms. Sponsors with more rigorous data collection requirements will likely have greater subject oversight mechanisms. For example, the NIH requires considerable oversight of study participants. Conversely, it may be that other funding sources require less attention to addressing issues of withdrawal and subject oversight. Lastly, the level of IRB review is important because it reflects differences in the level of study oversight. For example, studies requiring full IRB review require greater clinical and ethical oversight given the greater potential for harm to subjects. PIs of protocols presenting greater risks to subjects may desire greater monitoring of such patients by including additional withdrawal stipulations that request subjects to notify them or undergo further tests.

Ensuring that subjects/participants can voluntarily withdraw from study participation without consequences is a basic tenet of patient self-determination. Reviewing consent forms for withdrawal provisions is a good way to a) raise awareness of both the variation in withdrawal procedures on consent forms, and b) begin an assessment of the ethical parameters of withdrawal. This study set out to determine whether barriers to withdrawing from study participation exist on consent forms. The study aims were designed to:

1. Measure the frequency in which consent forms adhered to the Common Rule regulations;
2. Assess whether certain study characteristics (phase of clinical trial, whether the study proposed to do therapeutic research, study sponsorship, and type of IRB review) are related to the presence of additional withdrawal stipulations;
3. Measure the frequency in which consent forms explain why withdrawal procedures are necessary for the subject’s welfare; and
4. Examine the language by which withdrawal procedures are framed to explore whether messages of persuasion or undue influence might be perceived.

Our hypotheses pertain to study aim #2. It is hypothesized that phase I trials would have greater details pertaining to withdrawal stipulations than phase II or phase III trials. It is hypothesized that therapeutic studies would have greater specificity in detail regarding withdrawal stipulations than nontherapeutic studies. It is also hypothesized that NIH-sponsored studies would have the greatest level of details regarding withdrawal stipulations than protocols with other funding sources, and that studies requiring full IRB review would include more additional withdrawal stipulations than studies requiring expedited review.

This paper presents findings on withdrawal provisions as noted on consent forms and discusses the ethical implications of additional stipulations to withdrawal from study participation beyond CFR guidelines. Discussion focuses on the potential ethical reasons for requiring subjects to remain in the study beyond their preference.

METHODS

To investigate adherence to Federal Regulations pertaining to withdrawal, and procedures for withdrawal from study participation, we conducted a content analysis of withdrawal parameters as stated in a sample of consent forms from all IRB-approved active studies approved by one Midwestern, academic hospital’s Institutional Review Board (IRB) between 1992 and November, 2002. To be clear, withdrawal parameters stated on the consent form boiler plate were not included in this analysis so as to highlight the practices of individual investigators rather than one institution’s policy. A total of 477 protocols involving human subjects research were active and eligible for inclusion. Retrospective chart review studies were excluded from analysis because there is a waiver of informed consent, and thus no document to review. This resulted in a total of 342 protocols eligible for analysis. A random sample of 33% resulted in a sample of 114 consent forms included for analysis. The withdrawal section of consent forms was assessed for the existence of statements of omission and commission pertaining to specific circumstances under which withdrawal occurs. Most, but not all, consent forms reviewed were for clinical studies, with physicians serving as the principal investigator. While most protocols included adult participants, some protocols included children or both adults and children. Insufficient information about subject eligibility on some consent forms prevented us from definitively ascertaining subject age status and undertaking an analysis of this variable. IRB approval for this research was granted by Loyola University Medical Center.

Independent variables pertaining to study characteristics included: phase of clinical trial, whether the study proposed to do therapeutic research (e.g., offering direct benefit to the subject/patient), study sponsorship (e.g., no funding, pharmaceutical company funded, or federally funded e.g., by a National Cancer Institute cancer clinical trials cooperative group), and type of IRB review (e.g., expedited review, full board review). We defined therapeutic research as studies in which participants may be expected to receive direct benefit from their involvement in the study. An example of therapeutic research is a phase III clinical trial designed to determine the effectiveness of an investigational drug compared to an FDA-approved drug for metastatic non-small cell lung cancer. Conversely, we defined nontherapeutic research as a study that is designed to obtain further knowledge but is not expected to be of direct benefit to the participant(s). An example of a nontherapeutic study is a study designed to determine the toxicity levels of four different doses of an investigational drug for possible use to decrease blood clots in bypass grafts. We used the stated study benefits section within consent forms at face value regarding the protocol’s therapeutic status.

All study characteristics and statements conveying additional stipulations to withdrawal beyond that required by the Common Rule regulations were entered verbatim into the statistical database, SPSS 12.0 (SPSS, Inc., Chicago, IL). Analysis entailed descriptive statistics, and univariate analysis involving testing differences in proportions using Pearson’s Chi-square test or Fisher’s exact test if expected frequencies were less than five. Logistic regression (backwards) was used to examine multivariate relationships between study characteristics and the dependent variables of sufficient sample size in the distribution of scores. These dependent variables included notify the PI, and a combination of three variables: notify the PI, return for another visit, and return for further testing, called “All Stipulations Combined.” All tests were two-tailed and p < 0.05 was considered statistically significant.

A content analysis was conducted for each consent form by systematically searching for themes and repetitions pertaining to additional steps to withdraw from study participation beyond those required by the Common Rule that emerged from the data (Luborsky, 1994; Huberman and Miles, 1994). Themes were developed by grouping coded segments into larger domains, reviewing the categorization schema for appropriate thematic fit, and then adjusting and reviewing the schema again until categorization and interpretation were sound. Two research assistants and one investigator (EJG) reviewed consent forms, and the investigators examined consistencies in the results.

RESULTS

Of the 114 consent forms examined, exactly half were for studies conducting therapeutic research and had no external sponsorship, and over three-quarters were phase I clinical trials and were fully board reviewed (see Table 1). The studies varied in clinical conditions researched: the most common conditions were cancer, cardiovascular, and infectious diseases; the least commonly studied conditions included dermatology, orthopedics, and renal.

DISCUSSION

To our knowledge, this is the first study to empirically examine whether consent forms from IRB-approved studies adhere to human subjects protections regarding withdrawal from study participation, and examine additional steps to withdraw from study participation beyond those required by the Federal Regulations.

Acknowledgements

REFERENCES