|Home | About | Journals | Submit | Contact Us | Français|
Research in novel therapies for Alzheimer disease (AD) relies on persons with AD as research subjects. Because AD impairs decisional capacity, informed consent often must come from surrogates, usually close family members. But policies for surrogate consent for research remain unsettled after decades of debate.
We designed a survey module for a random subsample (n = 1,515) of the 2006 wave of the Health and Retirement Study, a biennial survey of a nationally representative sample of Americans aged 51 and older. The participants answered questions regarding one of four randomly assigned surrogate-based research (SBR) scenarios: lumbar puncture study, drug randomized control study, vaccine study, and gene transfer study. Each participant answered three questions: whether our society should allow family surrogate consent, whether one would want to participate in the research, and whether one would allow one’s surrogate some or complete leeway to override stated personal preferences.
Most respondents stated that our society should allow family surrogate consent for SBR (67.5% to 82.5%, depending on the scenario) and would themselves want to participate in SBR (57.4% to 79.7%). Most would also grant some or complete leeway to their surrogates (54.8% to 66.8%), but this was true mainly of those willing to participate. There was a trend toward lower willingness to participate in SBR among those from ethnic or racial minority groups.
Family surrogate consent–based dementia research is broadly supported by older Americans. Willingness to allow leeway to future surrogates needs to be studied further for its ethical significance for surrogate-based research policy.
Alzheimer disease (AD) is common, incurable, and devastating. In 2000, there were 4.5 million Americans with AD and by 2050, this number is projected to be 12.5 million if no effective interventions are found to alter the trend.1 Current treatments are of only modest benefit. Much needed innovative and promising research on AD can involve invasive procedures with unpredictable risks.2,3 Although some patients with mild AD are capable of providing their own research consent, the disease usually leads to early decisional incapacity4,5 and surrogate consent is necessary for research.6,7
Federal regulations allow surrogate consent by legally authorized representatives (LAR) (45CFR46.116), but policy uncertainties have continued for three decades, for at least two reasons. First, the regulations defer to states for defining the LAR but few states have done so.8 Thus, in most jurisdictions, the common practice of obtaining family consent with subject assent7 occurs under a cloud of policy uncertainty. This uncertainty has led to some institutions not allowing any surrogate consent for adults in research, the Federal government temporarily shutting down such research in one institution, a self-imposed temporary moratorium at another institution, and lawsuits.6 Second, there is no consensus on how much special protection is needed when subjects are enrolled based on surrogates’ consent.9-11 Perhaps the most critical issue is how to balance the risks and the potential benefits of such research.6 Three recent state laws have diverged on this issue; internationally, there are discrepancies as well.12,13 For example, the California law14 does not set a limit on the level of allowable risk for nontherapeutic research, whereas both New Jersey15 and Virginia16 laws limit such risk to minor increase over minimal.
Because policy discussions regarding surrogate- based research (SBR) have continued for three decades without a clear resolution, it may be especially important to assess the public’s views about SBR in working toward a solution. However, there have been very few attempts to understand the attitudes of the lay public or of stakeholder groups regarding SBR.17-19 We therefore designed a survey to assess the views of a nationally representative, policy-relevant sample of the general public (namely, older Americans) regarding surrogate consent for four research scenarios of varying degrees of risk and potential benefit. Because racial and ethnic minorities are underrepresented in AD research,20 we also examined their attitudes toward SBR. Finally, we examined how much latitude or leeway21,22 people are willing to confer on their surrogates. Many studies examining clinical treatment decisions have found that what people say they would want and what their surrogates think those people would want do not always match.23 These studies did not ask about leeway. But the significance of the discrepancy may depend in part on how strictly people want their preferences followed.
The HRS is a nationally representative study of persons 51 and older, designed to investigate the health, social, and economic implications of the aging of the American population with a sample of more than 30,000 individuals.24 Interviews are conducted with all HRS respondents either by telephone or in person, with the latter used preferentially for those aged 80 or older. The baseline participation rate was 80%, and the reinterview participation rates for subsequent waves have been 92 to 95%. Additional HRS information is available at http://hrsonline.isr.umich.edu.
Informed consent is obtained orally from all respondents prior to each HRS interview. The HRS was approved by the Institutional Review Board of the University of Michigan.
Our module survey was a modified version of a paper-based survey that we had used in a previous study,17 adapted for the brief interview format of the HRS modules (appendix e-1 on the Neurology® Web site at www.neurology.org). After a brief introductory background on AD and the rationale for the survey, the subjects were given one of four SBR scenarios: a lumbar puncture (LP) study, a randomized controlled trial (RCT) of a new drug, a vaccine study, and a first-in-human gene transfer neurosurgical study. These scenarios approximated real studies in AD.2,3 Then the subjects were asked three questions (see table 2, column 1).
The module was administered to a random subsample of the 2006 HRS participants. Of the 1,517 subjects approached for the module, two declined at the outset. The remaining 1,515 subjects were randomized to one of the four SBR scenarios: 374 to the LP scenario, 398 to the new drug RCT scenario, 375 to the vaccine scenario, and 368 to the gene transfer scenario. Depending on the survey question, 52 to 71 persons (out of 1,515) answered “don’t know” or refused to answer; they were excluded from the analyses below. The randomization was successful in that the groups were not significantly different on age, gender, race/ethnicity, education, net worth, marital status, or religion.
In analyzing our module survey responses, we included additional variables from the HRS parent survey listed in table e-1. We also examined four additional HRS parent survey questions for the leeway question analysis: 1) “How much do they [spouse or partner] really understand the way you feel about things?” 2) “How much can you rely on them if you have a serious problem?” The same questions were also asked regarding their children. The response choices were a lot, some, a little, or not at all. These questions were chosen because surrogates of patients with AD tend to be spouses and children. We wished to examine if the nature of those relationships predict the respondent’s attitudes regarding leeway, since identifying strong predictors of leeway could have policy implications. Although responses to these extra questions are available only in subsets of participants who have spouses or children (ranging from n = 494 to n = 633 of our module sample of 1,515, depending on the question), neither the availability nor distribution of these variables was associated with the four scenarios.
For each SBR scenario, the proportions of respondents who find the SBR research acceptable from a societal perspective and from a self perspective were calculated. The proportions of respondents willing to allow no, some, or complete leeway were calculated for each scenario.
The effect of the scenario on willingness to allow family consent for SBR, on willingness to participate in SBR, and on willingness to allow leeway were each assessed using χ2 tests. We assessed the effect of perspective for each scenario, using the McNemar test.
Race/ethnicity categories were defined as non-Hispanic white, non-Hispanic black, Hispanic, or other. A logistic regression model was fit for each scenario using responses to each of the three main survey questions as the dependent variable. Since our primary interest was in assessing how the responses to the three main survey questions vary by race/ethnicity, adjusting for potential confounders and other variables potentially related to attitudes toward SBR, the logistic regression models were adjusted for all of the variables in table 1 and table e-1.
We examined two types of models. First, we evaluated the predictors of willingness to allow leeway (dichotomized as some/complete leeway vs not) using logistic regression models, where the potential predictors examined were the HRS variables in table 1 and table e-1, the scenario dummy variables, as well as a variable for willingness to participate (because it turned out to be a strong predictor in bivariate analysis). The covariates were modeled carefully to assess their appropriate functional relationships to the response variable by using polynomials (e.g., for variables such as age) and categorical or dichotomized dummy variables. Second, similar but separate logistic regression models were used to test if allowing leeway was associated with relying on spouse or children and feeling understood by spouse or children.
Statistical analyses were performed using Stata 9.2 (Stata Corp, College Station, TX). All analyses except for McNemar tests were adjusted for the complex sampling design of the HRS survey using the HRS respondent level sampling weights (http://hrsonline.isr.umich.edu/meta/tracker/desc/wghtdoc.pdf). Dr. H.M. Kim conducted the statistical analysis.
Table 2 summarizes the responses for the three main survey questions.
Most of the respondents are supportive of allowing families to make surrogate consent decisions for dementia research (67.5% to 82.5%). Even for a first-in-human gene transfer scenario, nearly 68% state that our society should allow families to make such surrogate decisions. The response patterns were significantly different across the four scenarios for the societal question and for the self perspective question, and the pattern trended toward a difference for the leeway question.
The proportion of those willing to participate in the various SBR scenarios is generally similar to the level of support for family surrogate consent, but a significantly greater proportion would allow family consent in the vaccine study than would themselves want to participate (odds ratio [OR] = 3.09; 95% confidence interval [CI] = 1.89–5.25; p < 0.001). For other scenarios, the societal to self perspective comparisons showed an OR of 1.24 (95% CI = 0.77–2.00; p = 0.42) for the lumbar puncture study, 1.59 (0.96–2.68; p = 0.07) for the drug RCT scenario, and 1.29 (0.81–2.06; p = 0.31) for the gene transfer scenario.
The relationship between race/ethnicity and attitudes toward the various SBR scenarios are summarized in table 3.
The direction of the race/ethnicity effect for the societal perspective question is not consistent across scenarios and none of the effects are significant. In absolute terms, the majority in all race/ethnicity groups responded that society should allow family consent for all scenarios.
For the self perspective question, the overall pattern trends toward black and Hispanic respondents being less willing to participate in SBR, although the only significant results are that black respondents are less willing than white respondents to participate in the drug RCT and Hispanic respondents are less willing to participate in the gene transfer study. For the leeway issue (not shown in the table), the only significant difference is that Hispanic respondents are less likely than white respondents to allow some/complete leeway for the gene transfer study (adjusted OR = 0.26; 95% CI = 0.09–0.79; p = 0.02).
Depending on the scenario, between 55 and 67% would allow their surrogates some or complete leeway to decide against their stated wishes regarding SBR participation. Among those who do not desire to participate in SBR, 26% would allow some or complete leeway; among those who desire to participate, 76% would allow some or complete leeway.
Table 4 summarizes the adjusted ORs of the variables retained in the final logistic regression model to predict leeway question responses. The strongest predictor of whether one is willing to allow at least some leeway remains whether one is willing to participate in research, with an adjusted OR = 8.31. There were other significant, although weaker, predictors of willingness to grant leeway. Married people, women, those in excellent health, and respondents to the gene transfer scenario (although not among those who considered religion very important) were more likely to give at least some leeway than not.
In separate models, the more the respondent felt understood by a child, the more likely he or she would allow leeway (adjusted OR = 1.37, 95% CI = 1.07–1.75; p = 0.01), corresponding to a 37% increase in the likelihood of willingness to allow leeway with every one level increase in feeling understood (out of four levels, ranging from “not at all” to “a lot”). Being understood by one’s spouse or partner had a similar effect (adjusted OR = 1.44, 95% CI = 1.07–1.94; p = 0.02). The degree of reliance on one’s spouse or child, however, was not related to one’s willingness to allow leeway.
We found broad majority support—ranging from 68% to 83%, depending on the scenario—for a societal policy of family surrogate consent for AD research. A previous survey of adults at increased risk for AD also showed majority support for family consent for the same four SBR scenarios.17 Another previous study of research participants at risk for AD19 found very high willingness to participate in SBR of relatively low risk. A survey in Quebec18 showed similar results but its emphasis on examining a variety of surrogate types (legal guardian, family) makes it difficult to compare with our results.
The responses from the societal and self perspectives are closely related, but they should not be treated as equivalent. Personal willingness to participate is not a proxy for views about the ethical appropriateness of SBR because many people may choose not to participate for a variety of reasons. If the point is to characterize public opinion on the ethical permissibility of SBR, then respondents’ views from a societal perspective should be primary. Of course, it may not be ideal if many more people are willing to allow SBR as a societal practice than those who are willing to participate. However, even for the vaccine study scenario, in which the gap is greatest, the majority (57%) of respondents still desire to participate.
The significant underrepresentation of racial and ethnic minorities in AD research is a perennial concern.20 Recent studies have shown that racial and ethnic minorities, despite the legacy of abuses such as the Tuskegee study, are similar to whites in their attitudes toward and participation in research.25,26 We found that regarding the social policy of family surrogate consent for SBR, the differences among racial and ethnic groups are small and the overall level of support is high. In terms of a desire to participate in SBR, the overall pattern among minority respondents is a somewhat lower willingness to participate. However, responses vary a great deal by scenario. For instance, although black Americans are significantly less likely to volunteer for a drug RCT, they appear more likely than whites to endorse the gene transfer neurosurgical study for AD with family consent (and are no less willing to participate).
Leeway is an important concept because it may signify a subject’s recognition that the future cannot be pinned down in detail and it allows subjects to convey their trust in their surrogates who will have more complete information at the time of actual decision-making. The majority of our respondents would allow some or complete leeway for all the scenarios. However, it should be noted that a significant minority (up to 45% for the gene transfer scenario) would not allow leeway.
Those who are willing to participate in research are highly likely to allow leeway. Relational factors (such as feeling understood by their potential surrogates) clearly play a role as well—and this may have important consequences on how we view current practices. Because there is ample cross-sectional evidence that even impaired dementia patients generally express “reasonable” preferences (i.e., similar to healthy controls) for treatment27 and for research participation,28 if such preferences are predictive of a desire for conferring leeway then the current practice of surrogate consent with subject assent seems well supported. However, only a longitudinal study would provide direct evidence regarding whether a potential subject’s current preference (as reflected in the subject’s assent) is reflective of a longstanding desire for granting leeway.
Even among those who state that they do not desire to participate in SBR, a significant minority (over a quarter) are willing to confer some or complete leeway on their future surrogates. If these subjects are added to those who are willing to participate in the four SBR scenarios, the proportions of potential participants would rise to approximately 80% for the LP study, 88% for the drug RCT, 79% for the vaccine study, and 74% for the gene transfer study. A recent study of elderly subjects from senior centers and primary care clinics regarding an LP SBR scenario found similar results.22 However, this point needs to be balanced by the fact that the surrogates of those who are willing to participate and are willing to grant leeway will have the latitude to override that willingness to participate; leeway does not necessarily imply an increase in enrollment.
This study’s main limitation is that the scientific and policy issues surrounding surrogate consent for research are quite complicated. This brief survey captures only a snapshot impression regarding SBR. It may be that as people learn and reflect more, their attitudes may change. We are therefore currently conducting a series of all-day deliberative democracy sessions with laypersons and caregivers regarding SBR.29 Another limitation of the study is that it focuses on dementia research and may not generalize to other areas of SBR.30,31
An important strength of this study is its large and representative sample with an excellent response rate. The Federal government is revisiting the issue of surrogate consent and recently solicited public input.32 As a nationally representative study on the topic,our study provides new data for an important policy discussion.
The primary ethical goals in overseeing SBR are to minimize the potential harms of SBR (already mandated by the current regulations) and to minimize the potential for enrolling persons whose values may not be compatible with participation in SBR. In this regard, the current evidence, including the results of our study, provides some guidance. First, it appears that even for invasive studies the prior probability of an older American’s willingness to participate in SBR is high. Second, even among those who are not willing to participate, there is a sizable minority who are willing to confer some leeway on their surrogates. More research on attitudes toward leeway may provide further guidance to policy makers and to researchers at the time subjects are approached for research. Third, even when patients suffer from dementia with impaired decisional capacity, they may still retain some ethically relevant abilities. Persons with AD express personal choices regarding research participation that are similar to those expressed by competent adults, suggesting that risk-related preferences are relatively preserved.28 Thus, the practice of requiring the assent of even incapable subjects may provide another level of assurance. Patients with AD may also have relatively preserved abilities to appoint a trusted relative to help make research decisions for them.33 Finally, it is quite likely that persons acting as family surrogates will be protective and provide a gatekeeper function.17 Future research should aim to provide further guidance on how to minimize the enrollment of those whose values are not compatible with SBR.
Address correspondence and reprint requests to Dr. Scott Kim, Bioethics Program, 300 North Ingalls St, 7C27, Ann Arbor, MI 48109 ude.hcimu@mikttocs
Supplemental data at www.neurology.org
Dr. Kim was supported by a grant from the NIA (R01 AG029550) and Greenwall Foundation Faculty Scholars in Bioethics award. Dr. Langa was supported by a grant from the NIA (R01 AG027010) and a Paul Beeson Physician Faculty Scholars in Aging Research award. Dr. Karlawish was supported by a Greenwall Foundation Faculty Scholars in Bioethics award and the Marian S. Ware Alzheimer Program. The National Institute on Aging (NIA) provided funding for the Health and Retirement Study (U01 AG09740) which is performed at the Survey Research Center, Institute for Social Research, University of Michigan.
Authors’ affiliations are listed at the end of the article.
Disclosure: The authors report no disclosures.
Received August 12, 2008. Accepted in final form October 3, 2008.