Investigators who conduct research in homes face complex ethical concerns. Making observations in a home can intrude on personal privacy, including the privacy of all persons in the household. Researchers need to consider, first, whether their research is likely to present inconvenience or harm to potential third parties who are not otherwise involved with the research. Moreover, researchers cannot avoid making observations that are unrelated to the research question at hand but may be relevant to the health and well-being of the children who are the subjects of the research or other household members (Gordis, 1991). For example, researchers described to the committee sometimes observing housing code violations, illegal drug use, firearms, and even suspected child abuse or neglect. Researchers face the ethical dilemma of what if any action to take in the face of conflicting ethical and legal obligations to protect children from harm and to respect privacy and confidentiality.
All investigators carrying out research with human participants have legal and ethical responsibilities under the Common Rule (see Chapter 3) and the standards of institutional review boards (IRBs), such as obtaining informed consent, ensuring the risks of research are proportional to the expected benefits, and minimizing risks. Specific features of housing health hazard research place several additional ethical responsibilities on researchers. Previous chapters recommended involving community representatives and ensuring fully informed and voluntary parental permission for children to enroll in research. In this chapter we discuss other ethical issues and obligations: innovative research designs; reporting test results that have not been validated; potential third parties who may be affected but are not
research participants; and researchers’ role-specific obligations to develop plans for responding to risks that are incidentally observed when they enter children’s homes. On these issues we point out how the perspectives of community representatives may differ from those of researchers. Hence, the process of understanding and responding to the views of community representatives as described in Chapter 5 is also important for helping researchers clarify their ethical responsibilities.
A variety of study designs are used in housing health hazards research to obtain information on hazards and to test methods to reduce health risks. The study design chosen by researchers depends on the aims of the study, existing knowledge on the topic, and the perceived magnitude of the hazard. Regardless of the specific approach, all study designs should be scientifically and ethically sound. Well-designed and well-executed research is necessary to improve the health of children living in poor-quality housing. For some large-scale intervention studies, the approach might include conducting a small pilot study to minimize the potential for unanticipated negative consequences prior to implementation of the full research project.
At the most basic level, there are observational studies that describe or enumerate specific hazards associated with housing, such as case reports, surveillance by public health agencies, or descriptive studies by researchers. Some observational studies conduct home walk-throughs, others use existing records, reports, and data or conduct telephone or in-person interviews. Some studies include measurements of physical (e.g., radon), chemical (e.g., lead), biological (e.g., cockroach antigens), or psychosocial (e.g., accidents, violence) hazards in homes or neighborhoods. Although causality can generally not be inferred from observational studies, in some instances in which the hazard and remedy are clear, observational studies may be sufficient for devising preventive strategies, as in the case of window guards to prevent falls from windows (see Chapter 2). For many housing-related problems, however, the causal chain between a specific hazard or set of housing conditions and a health outcome is not clear. Observational studies may still be useful in determining if persons with certain baseline characteristics are more likely to experience the outcome of interest.
Even if the causal links between a housing condition and an adverse health outcome are established, however, it may be uncertain how to prevent or remedy the condition. If there are multiple contributing factors, it may not be clear which ones should be targeted for interventions. Studies on asthma in children, for example, indicate that it is associated with multiple indoor pollutants, including settled allergens (cockroach, dust mite, cat and dog), environmental tobacco smoke, and mold or fungi (Institute of
Medicine, 2000). It may not be known whether plausible interventions are in fact effective, or whether the benefits of preventing or ameliorating the hazard outweigh the adverse effects of the intervention. Moreover, with housing hazards, some effective interventions—such as moving a child to housing without the hazard—may not be feasible because of resource constraints.
To evaluate interventions, researchers conduct experimental studies that compare outcomes in a group that receives an intervention with outcomes in a control group. In order to conclude that outcome differences between the groups are the result of the intervention rather than some other factor, the only difference between the control and intervention groups should be the intervention. That is, the intervention and control groups need to be similar at the baseline and with respect to other changes over the course of the study. Quasi-experimental designs, such as those using comparison or control groups from a previous period (“historical controls”), are less able ensure that the intervention and control groups are comparable. The “gold standard” for evaluating interventions is randomized controlled trials with strict study entry criteria, detailed protocols, and well-defined treatment and control groups to which participants are assigned randomly. Randomization is important because it helps to ensure that the intervention and control groups are similar at the baseline in all pertinent characteristics. Rigorous clinical trials on housing health hazards can provide the evidence base for promulgating intervention “best practices,” regulatory standards, and health-based benchmarks for interpreting environmental or human biomonitoring measurements.
In a clinical trial, an important question is whether it is ethically acceptable to randomly assign participants to intervention and control groups. The term “research equipoise” has been coined to describe the genuine uncertainty that should exist about whether the intervention or control arm is better (Freedman, 1987, as cited in National Bioethics Advisory Committee, 2001). As described by the National Bioethics Advisory Committee (2001, p. 78), “research equipoise does not require numeric equality of intervention risks or potential benefits. Rather, research equipoise requires approximate equality in the relation between the risks and potential benefits of the study and control interventions.” The committee believed that in clinical trials expert practitioners should make judgments about research equipoise. In housing health hazards research, however, the assessments of risks and benefits by researchers may differ from those of community representatives or the parents of potential child subjects, as discussed in Chapter 5. Thus, for research on housing health hazards in children, the assessment of research equipoise should balance scientific expertise and the views of community representatives.
Study interventions may attempt to affect a clinical or behavioral outcome in child subjects, a change in their housing environment, or some
combination of these factors. Because public health practice is not strictly codified, there is uncertainty because of the ambiguity of what constitutes “accepted practice.” This uncertainty makes it more difficult to determine if “approximate equality” exists between risks and potential benefits of interventions in the experimental and control groups. Often the control group receives the best current practice; however, ethical dilemmas arise in housing health hazard research if the best current practice is considered impracticable. The situation is similar in some clinical trials, in which a shorter, less toxic, or less expensive treatment is compared with standard therapy. Other dilemmas arise when an experimental intervention is widely believed to be effective, despite the lack of rigorous empirical evidence. The ethical concern is whether the study is harming members of the control group by withholding or delaying an intervention believed to offer them the prospect of direct benefit.
Determining whether a particular clinical trial is ethically acceptable requires value judgments about the weight of the evidence, the level of uncertainty, and the potential effects of the research study on changing housing standards or policies. A weight-of-evidence approach is often used by expert panels or regulatory bodies to analyze data from experimental or quasi-experimental studies to assess the strength of association between a specific exposure and health outcome or to test the efficacy of an intervention designed to reduce mortality, morbidity, or exposure. Community residents may have different judgments than researchers or IRBs.
Community representatives often ask that research provide some direct benefit to those in both the intervention and control groups, as well as to the community. They may expect that direct benefits beyond the best available practice be provided. In the Seattle project, Partners for Healthy Communities, the community board challenged a design in which the control group would not receive a benefit. The study design was altered to address these concerns:
The group receiving the full intervention was compared with a group receiving a lower intensive intervention, rather than a control group that received only usual care. Although the researchers were concerned that this would diminish the ability of the study to demonstrate the impact of the full intervention, community members felt that providing benefit to all participants was more important (Krieger et al., 2002a, p. 366).
However, attitudes towards control group interventions changed over time, as the trade-offs between providing immediate benefit to children in the study and gaining data that would provide compelling evidence that an intervention should be widely adopted became apparent. Ultimately, the Seattle partnership reached consensus that the evaluation of the intervention was “less convincing because it did not have a usual-care control
The Children’s Lead Exposure Reduction Study (CLEARS)
CLEARS was a randomized clinical trial carried out between 1992 and 1995. It studied whether regular household cleaning to reduce house dust lead levels would reduce lead exposure in moderately poisoned young children (blood lead levels between 10 and 25 mg/dl). The households in the cleaning intervention group received education about lead and biweekly assistance in household cleaning. The control group received education about lead and accident prevention. Thus, the control group received some direct benefits that were not available to the intervention group.
The Health Outcomes and Measures of the Environment Study (HOME Study)
The HOME Study is an ongoing randomized study that tests interventions for two different hazards, lead poisoning and injuries. In one group, research staff will make repairs to reduce residential lead hazards (e.g., window replacement, paint repair, dust control, and water filters). In the other group, research staff will repair residential injury hazards (e.g., smoke alarms, window guards, stair gates, safety latches). The experimental group for one intervention serves as the control group for the other intervention. This is the first randomized trial to test the efficacy and safety of multifactorial passive environmental alterations to prevent childhood lead poisoning and residential injuries.
group. The consequence was less benefit to the community because it made it harder to find funding to sustain project activities” (Krieger et al., 2002b, p. 367). In the renewal grant for the project, the study had a usual-care control group that will receive the intervention after a delay of one year.
Innovative study designs can offer a benefit to both the intervention and control groups and thereby resolve some of the ethical concerns about housing health hazards research. First, as done in the Seattle project, the control group might receive the study intervention in a delayed manner, although there still may be objections to delaying an intervention that is believed to be effective (Krieger et al., 2002b). Second, child subjects in the control group might receive an intervention on a different housing health hazard. For example, in a clinical trial of the effect of periodic house cleaning on children’s blood lead levels, the control group received household injury reduction education visits (Rhoads et al., 1999). A variation on this is a 2×2 intervention trial that tests interventions for two different conditions, with the control group for one condition receiving the active intervention for the other condition (Bruce Lanphear, personal communication); see Box 7-1. Third, crossover designs provide one or more interven-
tions to all child subjects on a staggered basis, offering the prospect of direct benefit to all child subjects in a study (Israel et al., 1998). These various designs that provide the prospect of direct benefit to all child subjects can eliminate or reduce ethical concerns about placing vulnerable children at inappropriate risk.
REPORTING OF TEST RESULTS
Researchers commonly carry out experimental tests on samples collected from child subjects or their environments. Unlike tests used in routine clinical practice, the significance of experimental test results may be unknown. Indeed, a goal of the research might be to establish the validity of such tests or to determine the health implications of the results. Ethical dilemmas may arise if parents of child subjects or community representatives believe individual parents should have the results of those tests. On one hand, information whose significance is unknown or uncertain may not provide objective benefit to parents; on the other hand, respect for persons may imply that parents should be the ones to judge whether they want information that researchers have obtained about themselves or their homes. The particular salience of the issue of reporting test results in housing health hazards research was highlighted by the Grimes case (see Chapter 3), in which plaintiffs charged that the researchers failed to provide them information about the level of lead in dust in the homes in a timely manner; the significance of the measurements was not known.
Studies of housing health hazards often collect human tissue samples, such as blood or urine, to establish the child subject’s exposure, health effects, or susceptibility to a specific health outcome. These tissue samples are used to measure a variety of parameters, such as immune response to specific allergens or levels of specific compounds in body fluids, such as blood lead levels (Centers for Disease Control and Prevention, 2003; Needham and Sexton, 2000), or levels of urinary metabolites of tobacco smoke, pesticides, or persistent compounds (Quandt et al., 2004; Adgate et al., 2001; Hecht et al., 2001). In addition, such studies may collect environmental samples to establish the magnitude of exposure to a variety of potential hazards, such as chemical or biological agents. Levels can be measured from samples collected in or on household carpets, furniture, or other surfaces (Lanphear et al., 1996; Platts-Mills et al., 2000; Rich et al., 2002), as well as in air (Institute of Medicine, 2000), water and house dust (Colt et al., 2004; Rudel et al., 2003), or soil (Aschengrau et al., 1994; Mielke and Reagan, 1998). Several issues arise with regard to the reporting of test results to parents: the validity of the tests, the time frame for reporting results, and the parents’ expectations. Similar issues of reporting test
results occur in biomedical research, when the validity and clinical significance of certain tests may not be known.
Validity of Tests
In housing health hazards research, both routine clinical tests and experimental tests may be done, depending on the study design and the problem under investigation. For clinical tests, there are standardized testing methods, age-specific normal ranges, predictive value for various diagnoses or conditions, and levels at which clinical interventions should be instituted. Blood lead measurements are an example of a clinical test often used in housing health hazards research: test results can be obtained from certified laboratories within days or weeks of sample collection and have established health benchmarks (Centers for Disease Control and Prevention, 1991). If researchers carry out tests that are also used in clinical practice, they typically provide patients or their physicians with the results, a description of the normal range of values, and the implications of results outside the normal range. Providing results of clinical tests in the range of concern to parents of child subjects in a timely manner is ethically required because it allows appropriate medical follow-up to be obtained.
In contrast, experimental tests may have uncertain validity. Indeed, one goal of the research may be to determine the validity of a new method of measuring a variable or the strength of an association between a new measurement and a clinically meaningful outcome. Two types of validity are pertinent (Secretary’s Advisory Committee on Genetic Testing, 2000): analytic validity and clinical validity. Analytical validity indicates how well the test measures the property or characteristic it was intended to measure: it encompasses reliability—the probability of obtaining the same result with repeated measurements. Clinical validity refers to the probability that a test result correctly diagnoses a condition or predicts a disease or clinical condition.
In research on housing health hazards, some experimental tests may be carried out to help characterize the extent of potential exposure. For most such results, a description of the normal range of values and an assessment of the implications of the results is uncertain or unknown. Many types of tissue and environmental sample measurements do not have established analytical protocols, well-established laboratory quality control and assurance processes, or normative reference ranges or health benchmarks that permit ready interpretation of the test results (Centers for Disease Control and Prevention, 1991, 2003). The significance of results from such experimental tests for an individual subject may be unknown or uncertain until long after the samples have been collected, often not until all study data have been analyzed, and sometimes not even then.
Ethical Issues in Reporting
The threshold criteria in determining whether to share test results are typically whether the results are valid in the two senses discussed above. There is no clinical benefit to reporting the results to individual parents if they cannot be meaningfully interpreted. In biomedical research, when the validity of experimental tests on biological specimens is not established, individual results generally are not reported to participants. For example, a model consent form for genetics studies states that “the study is not meant to test your personal medical status” and that participants will not receive the results of research on their sample (Beskow et al., 2001). In some cases, some validity of the test can be established at the completion of the study; if so, the researchers may agree to then offer the tests results.
In some studies, community representatives or parents of prospective child subjects may believe that parents have a “right to know” information that researchers have obtained about their children regardless of whether the validity of the tests have been established. They may not trust or may disagree with researchers’ judgments that the significance of the results is not known. Or they may simply want to have information about themselves even if there are no actions they could take that are known to reduce their risk of health hazards. That is, they may value the information about themselves for its own sake, even though its significance is unclear. Researchers report that in some cases community groups would like the results of experimental tests (such as urinary pesticide metabolite levels) without clear clinical implications to be nonetheless made available to the tested individuals if requested (Eskenazi et al., 2005).
When such disagreements arise, researchers have several ethical obligations that are not spelled out in the federal regulations. As a first step, they should discuss with parents of potential child subjects and community representatives what tests they will be conducting, explaining the limitations of the experimental tests and the potential misinterpretation of results. They also need to discuss whether test results should be made available. In some cases, researchers may persuade community representatives that there is little benefit and much risk to making results of unvalidated experimental tests available. In other cases, the community may persuade researchers that the results of individual tests should be made available to all the parents whose children are in the study.
If researchers decide to make results of experimental tests available, they need to consider how to do so in ways that minimize the harms and maximize the benefits of providing results. First, the researchers should offer parents a choice of whether or not to receive results of experimental tests. Some parents will want to know such information, while others will not. Respect for persons requires that individual parents be given a choice. Second, parents need to understand the potential significance and limita-
tions of the results of experimental tests (Needham and Sexton, 2000). Parents also need to know where, if anywhere, they might go for help (e.g., personal physician, community clinic, or public health office). Third, the researchers need to make clear during the informed consent process whether, when, and how the results of experimental tests will be offered to parents. Some tests are run shortly after samples are taken, while others may be run in batches, sometimes at the end of the study after all samples have been obtained. In some cases, the health significance of particular test results will become known over time, for example, as research on a particular biomarker advances. Parents should be told in the informed consent process what the researcher will do in such situations. Similarly, some test results (e.g., measurement of a particular hazard in the home) may suggest a hazard to other residents in the household; researchers should consider how to provide them information about the findings, taking into account the wishes of the subjects’ parents.
To address these ethical dilemmas, researchers should discuss experimental tests that are part of the research protocol with the community (see Chapter 5) and should ensure that the informed consent process includes thorough disclosure of whether, when, and how the results of such tests will be shared with parents (see Chapter 6).
ETHICAL OBLIGATIONS TO THIRD PARTIES
All research can have unintended consequences. Research on housing health hazards in homes can have unintended consequences for individuals who are not study participants (“subjects,” as defined in the federal regulations)—individuals with whom researchers interact or about whom they collect identifiable information. If third parties meet either of these criteria, researchers must obtain their informed consent (see Chapter 6). However, even if the federal regulations do not consider them research participants, people living in the same household, the same multiunit dwelling, or the same neighborhood as a study subject may be affected by research. Researchers may have ethical obligations to such third parties.
Other household residents may experience psychosocial harms, such as embarrassment or shame, if they or other residents are observed to be living in substandard housing or engaging in certain behaviors, such as alcohol abuse. Residents may also face legal liability if they are identified as carrying out illegal activities in their homes. In addition, residents may encounter physical risks, such as exposure to noise or dust, resulting from procedures carried out as part of the study: if these other residents are not notified, they cannot make plans to be out of the home.
Researchers need to anticipate and make plans for the effect of their research on other household residents. Often the researchers can simply ask
a parent of the child subject to notify other residents about the study and the interventions. Such notification gives other residents an opportunity to be absent from the home when the research interventions are carried out so that they are not inconvenienced by interviews, inspections, or repairs and so that their privacy is not compromised.
Research carried out in rental properties can have consequences for landlords. Researchers need to examine the specific terms of a lease for any restraints on the normal right of the occupant to invite any law-abiding person into the dwelling and to make minor improvements, such as installing battery-powered smoke alarms. If researchers propose to make significant structural changes to the home, such as installing new windows, the permission of the owner needs to be obtained. However, minor modifications that a tenant would have authority to make do not ordinarily require additional permission. Researchers should also take reasonable steps to provide information to landlords about possible public resources for helping to correct housing hazards and code violations, particularly if such hazards might be reported to authorities.
Neighbors may also experience adverse consequences of research. For example, pest management carried out in one unit of an apartment building may cause pests to flee to other units. Researchers need to anticipate such unintended results and take steps to minimize them. In the case of pest control, for example, researchers might reframe the study intervention to carry out pest management throughout a building rather than in a single unit to avoid causing harms to residents of other units. In other research studies, such as when repairs are made to a single unit, researchers should provide neighbors whatever notice would be expected if the landlord or tenant were carrying out similar activities outside the research context. If potentially disruptive activities are planned, informing neighbors of the plans gives them the opportunity to act as they wish in response to the activities (such as by leaving their units).
It is important to note that third parties may benefit from research as well as suffer inconvenience or risks. Educational activities may benefit neighbors as well as the family participating in the study. For example, in a study of a pesticide intervention involving rural farm workers, some participating parents brought friends or family to meetings to discuss strategies for reducing pesticide exposure (Salvatore et al., 2004). The diffusion of information from families participating in a research study to neighbors may result in the neighbors’ altering their health behaviors and reducing their exposure to hazards. Landlords benefit if the research involves such improvements as pest control, the installation of smoke detectors, or other interventions that increase the quality and value of their properties.
Community groups may also be third parties in many housing-related research studies. Community representatives sometimes ask for benefits to
the community beyond whatever benefits individual subjects might receive, as well as the general benefit of increased knowledge (see Chapter 5). Often, community groups seek to have local residents hired as research staff and receive training that will enhance their employability in the future. Furthermore, the community may want researchers to specify how they will present the project’s findings when it is completed so that the community can use the new knowledge, for example, to advocate for funding of effective interventions. Researchers should also present relevant findings—either on their own or in conjunction with community representatives—to local, state, or federal officials and testify at pubic hearings to support evidence-based public policies that would ameliorate housing health hazards. Researchers cannot be expected to ensure that research findings are fully implemented, but these steps can help the community benefit from the findings. Researchers should develop a plan to disseminate results to the families participating in the study, as well as the affected community.
The appropriate actions regarding third parties will vary according to the particular study and need to be determined on a case-by-case basis, following the general ethical guidelines of respect for persons, beneficence, and justice. Consulting with community representatives and parents of potential child subjects will help researchers understand the full range of a project’s potential effects, as well as what those people believe researchers need to do to fulfill their ethical obligations (see Chapter 5).
In considering risks to third parties, researchers need to focus on risks that are foreseeable and significant rather than those that are conceivable but extremely unlikely or of minor importance. It is important not to place requirements on researchers that are overly broad, vague, or open-ended, lest they deter important, soundly designed research that is intended to alleviate housing health hazards that are disproportionately severe in vulnerable populations.
Although research staff are not technically “third parties,” the committee notes the responsibilities of a project’s leaders regarding the physical safety of staff working in the communities and entering child subjects’ homes. These risks to staff may be greater than when research is carried out in a medical institution. As with all research, housing health hazards researchers have an obligation to consider the safety of their staff and to develop plans appropriate to their particular research project: such plans might include having staff conduct home visits in pairs or only during daylight hours or providing staff with cell phones.
RISKS OBSERVED IN HOMES
When researchers enter homes for their work, they cannot avoid making observations beyond the information collected as part of the research.
They may observe risks that are not related to the topic of the study. If a researcher learns about environmental hazards or behaviors by others that place a child in imminent risk of serious harm, there may be a legal requirement to report such information to specific authorities. However, interventions to reduce risks could violate confidentiality and could be ineffective or even counterproductive. In addition, the child at risk or the person(s) engaging in behavior that puts a child at risk may not be a participant in the research study (as defined in the federal regulations). The researcher may have no prior relationship with those being observed and may be viewed as invading their privacy.
Confidentiality must be distinguished from the related concept of privacy. A loss of privacy occurs if others intervene in “zones of secrecy, anonymity, seclusion, or solitude” or in intimate relationships, causing a person “to be observed, touched, or intruded upon” against his or her wishes (Beauchamp and Childress, 2001, pp. 295-296). Privacy is also violated if others obtain information about a person that he or she wants to keep inaccessible. Confidentiality refers to limits on the dissemination of information disclosed by a person within a special professional relationship, such as the doctor-patient relationship or participant-researcher relationship (Beauchamp and Childress, 2001). Within these special relationships, the disclosed information is protected against disclosure to third parties by professional codes of conduct and by law. Furthermore, researchers often promise confidentiality of research data, with certain limitations, during the informed consent process. Thus, for example, when physicians have permission to gather medical information about a child patient, they may learn that the child is at risk for child abuse or domestic violence or places others at risk because of a contagious disease. To take steps to protect the patient or third parties, the physician would have to breach confidentiality; the ethical issue is whether it is appropriate to do so. In housing health hazards research, a researcher who has permission to enter a home to collect research data might incidentally observe evidence of child abuse or domestic violence, even though these are not the topic of the research. Here the ethical issue is whether it is appropriate to use information obtained under permission to collect research data for purposes that go beyond the scope of this permission.
Overriding confidentiality in such situations poses dilemmas for researchers because several strong ethical guidelines may be in conflict. First, researchers have an ethical or professional obligation to try to prevent harm to children who cannot protect themselves. In some situations, they may also have legal responsibilities through statutory reporting duties, which
depend on who is making the observation and in what state the research is being conducted. Second, well-intended actions may be ineffective or counterproductive and actually cause greater harm. Actions intended to alleviate risks may have unintended adverse effects (see the discussion below), such as the attempted eviction of the household after a unit has been reported for housing code violations. Third, researchers have an ethical obligation to respect the privacy and confidentiality of the residents of the homes in which their research is being carried out. Privacy and confidentiality show respect for persons affected by the research. Far-ranging interventions by researchers, even if intended to benefit residents of the household, may be considered meddlesome intrusions by them. In addition, privacy and confidentiality have instrumental research value by making it more likely that people will agree to participate in research.
In clinical medicine and public health, confidentiality may be overridden in certain situations to protect a person or third party from harm without legal repercussions; in some situations confidentiality must or may be overridden. For instance, confidentiality must be overridden in some circumstances to protect someone from child abuse, domestic violence, or elder abuse. In addition, confidentiality must be overridden to protect third parties, as when specified infectious diseases are required to be reported to public health officials. Widely accepted ethical guidelines (Beauchamp and Childress, 2001; Gostin, 2000; Lo, 2000) identify such situations in which confidentiality may or must be overridden to protect a person or third party from harm:
The potential harm to identifiable persons is of serious magnitude and high likelihood.
Breaching confidentiality will allow steps to be taken to prevent harm.
There is no less invasive alternative to overriding confidentiality for warning or protecting those at risk.
Harms resulting from the breach of confidentiality are minimized and acceptable. Disclosure should be limited to information essential for the intended purpose, and only those persons with a need to know should receive the information.
The more of these criteria that apply, the stronger are the ethical reasons for an obligation or a license to disclose confidential information. These guidelines form the ethical underpinnings of laws and regulations regarding the confidentiality of personal health information. This approach of balancing the likelihood of preventing serious harm against the harms caused by a breach of confidentiality can help researchers respond to risks they observe. The informed consent process (see Chapter 6) should clearly specify the situations in which confidentiality will be overridden.
Role-Specific Researcher Obligations
When physicians provide medical services to patients, they have a clear ethical duty to place the interests of the patient above their own self-interest or the interests of third parties, such as insurers. The doctor-patient relationship, based in trust, obligates physicians to behave in a competent and professional manner and assures patients that only interventions thought to be in their best interests will be recommended. As discussed in Chapter 4, research participants often conflate research and clinical care (“therapeutic misconception”). This is particularly understandable when the researchers are physicians or other health care professionals. In addition, research participants are familiar with relating to health professionals in the clinical setting and expect such professionals have their interests at heart.
Unlike clinical care, the primary objective of research is not to benefit participants directly, but rather to produce generalizable knowledge. While investigators are not obligated to act solely in the interests of individual participants, they have legal duties to protect them by obtaining informed consent, ensuring that the risks of research are proportional to the expected benefits, and minimizing risks. In particular, investigators need to inform participants about the interventions that will be carried out during the research and their risks and benefits. In addition, researchers must guard against the possibility of conflating research and clinical goals.
As discussed above and in previous chapters, two role-specific obligations are (1) to provide information beyond that required in the federal regulations, if appropriate, for parents to decide whether to enroll their children in the research project (see Chapter 6) and (2) to have a plan to deal with risks observed in homes that are unrelated to the research.
In addition, researchers may be in a special position to prevent harms from housing health hazards because they have expertise about housing hazards and a unique opportunity to intervene because they are in subjects’ homes. If researchers do not act, an opportunity to prevent serious harm may be lost because no one else with similar expertise observes the conditions. Moreover, parents of research subjects are likely to expect that investigators will inform them if environmental hazards detrimental to their health are discovered in the course of the research, even if those hazards are not the focus of study. These characteristics—their expertise, the unique opportunities to prevent harm, and the reliance and expectations that parents place on them—distinguish researchers from a plumber or electrician who enters someone’s home and is expected to carry out only a specific task.
Although some research critics argue that researchers who observe children in poor housing environments should “rescue” these children from harm and provider better living conditions (see Sherav, 2003), this seems an unreasonable expectation in most situations. Nor do researchers have an obligation to eliminate all housing health hazards in the households they
study: indeed, this task is beyond their power and resources. However, they do have an obligation to do what is reasonable under the circumstances. Above all, they should not make the situation worse for those already at risk. The appropriate course of action will depend on the specific situation: the type of risk identified, who experiences the risk, the nature of the research project, and the availability of community resources.
Researchers have an ethical obligation to anticipate what kinds of risk they are likely to observe, to develop reasonable plans to address anticipated risks which include procedures for how staff will be expected to respond to specific risks, and to train staff on these plans. Researchers also have an obligation to discuss how they will respond to such risks as part of the informed consent process. The first step for researchers is to consult with community representatives to identify likely risks and community resources.
Community representatives can help researchers identify and understand risks that they are likely to observe. They can also help researchers understand what responses to these risks might be considered acceptable and effective by families likely to participate in the research. They can inform researchers about community service organizations, and perhaps particular contacts within these organizations, that can provide information, support, and advocacy services to child subjects and their families. For example, researchers may want to be aware of the alternative housing resources in the community. The next step is to develop a plan based on the level of risk.
Risks That Are Imminent and Serious
In cases of suspected child abuse and neglect, where there is evidence of “imminent risk of serious harm,” such as clearly excessive force in disciplining children, the situation should be reported to child protective services for appropriate action. Each state has a child abuse and neglect reporting law and an agency designated to receive such reports; see Box 7-2. If a research staff member belongs to a profession that makes her or him a mandated reporter under the state’s law, the staff person is legally required to report such cases of neglect or abuse.1
There is considerable variation in each state’s child welfare system and
The National Clearinghouse on Child Abuse and Neglect Information reported that as of June 2003, 18 states (Delaware, Florida, Idaho, Indiana, Kentucky, Maryland, Mississippi, Nebraska, New Hampshire, New Jersey, New Mexico, North Carolina, Oklahoma, Puerto Rico, Tennessee, Texas, Utah, and Wyoming) have universal reporting requirements; see http://nccanch.acf.hhs.gov/general/legal/statutes/manda.pdt [May 2005].
The 1974 Child Abuse Prevention and Treatment Act (CAPTA) (42 USC 5101, et seq.) as amended and reauthorized by the Keeping Children and Families Safe Act of 2003 (P.L. 108-36), imposes certain requirements on states to be eligible for federal funds, including a child protective service and a mandatory child abuse and neglect reporting law. Although each state has its own definition of child abuse, federal law (42 USC 5106g(2)) sets the minimum definition for child abuse or neglect as “any recent act or failure to act on the part of a parent or caretaker, which results in death, serious physical or emotional harm, sexual abuse or exploitation, or an act or failure to act which presents an imminent risk of serious harm.”
All 50 states have a mandatory reporting statute, but there is considerable variation in the laws, most notably in the types of situations that must be reported and the kinds of persons who must report. In some states only helping professionals, such as health care workers, school personnel, child care providers, law enforcement workers, and mental health professionals, must report; in other states, any person who suspects abuse or neglect has a mandatory duty to report (Sieber, 2004). Almost every state imposes penalties, in the form of a fine or imprisonment, on those who “knowingly” or “willfully” fail to report.
CAPTA also requires states to provide legal immunity for good faith reporting of suspected abuse or neglect (42 USC 5106a(b)(2)(A)(vii)). The combination of potential penalties for failing to make a report and the immunity protection creates legal incentives to report: as long as there is a good faith reason to suspect abuse or neglect, it is in a potential reporter’s best legal interest to make the report rather than remaining silent.
governing status, and specific state requirements may change frequently. State laws differ, for example, in their definitions of mandatory reporters, definitions of abuse and neglect, the obligations and protections for discretionary reporters, and the time frames for submitting a report. In developing their anticipatory plans, researchers should consult with the local child protective services office to ensure that they understand applicable laws. The state or local protective services office may also be a resource in developing plans for responding to particular situations that are likely to arise in a research project. However, even when legal obligations don’t exist, there is an ethical obligation to act in cases when the risk is imminent and serious. Researchers who conduct housing health hazards research will likely also encounter situations when they need to determine on an individual basis if the presence of physical hazards in the home present an imminent risk of serious harm and therefore warrant reporting the situation to child protective services.
For example, if a researcher enters a home to conduct a vigorous clean-up of household allergens and discovers a large hole in the middle of the floor with a drop-off to the floor below, does this obvious safety risk to any child in the home fall within the definition of child abuse or neglect? In developing plans for how to respond to this physical hazard, investigators should first consider the relevance of the state child abuse and neglect law. If this situation falls within the scope of state law and the researcher is a mandated reporter, it must be reported to child protective services. In addition, all researchers should consider whether the parents can take steps to eliminate the hazard. For example, the parents might be able to report the situation to the local housing authority so that they can initiate repairs, get the landlord to make repairs, or obtain help from community organizations to set up barriers around the hazard. In contrast, if the child has already fallen into the hole and suffered injuries and the parents refuse to acknowledge the presence of risk, the investigators may be obligated to report the cause of the child’s injuries to child protective services.
Researchers have countervailing concerns when faced with evidence indicating potential child abuse or neglect. There are strong reasons to report, such as the desire to help and protect a child or the threat of civil and criminal liability. Yet, many professionals may believe that reporting would have a negative effect on the trust and the alliance they have built with the research participants. In addition, some researchers believe that child protective service agencies lack the necessary resources and discretion to effectively handle a reported case (Goldstein, 1999, Ch. 9). Physicians have given several reasons for not reporting suspected child abuse and neglect: lack of sufficient evidence that abuse had occurred; belief that the abuse was not serious enough; belief that the case had already been reported; belief that the report would disrupt current treatment; belief that the situation had already resolved itself; belief that they could help the child better themselves; and belief that child protective services agencies were of poor quality or that the agencies overreacts to reports (Kerns et al., 1994). Researchers may be concerned that reporting suspicions to child protective services agencies will undermine parents’ willingness to continue participation in the study. Although the committee acknowledges these concerns, the ethical and in some cases legal obligation to act in cases of imminent and serious risk is binding.
Risks That Are not Imminent and Serious
Researchers need to consider options for preventing harms that are not imminent and serious, and for each option, to analyze its effectiveness,
adverse consequences, steps to mitigate those adverse consequences, and acceptability to the parents. Researchers need to balance the harm that is likely to be averted against the harms resulting from breaches of confidentiality. Community organizations may have useful programs or services to help families address such risks, and they may be able to reduce the potential adverse consequences of responding to harms. For instance, legal services organizations may help families enforce their rights as tenants or protect them from attempts at retaliatory evictions, which are illegal.
Responses to these risks need not be as urgent as in the case of suspected abuse and neglect, and in some situations no response may be necessary. Discussing the observed risk with parents and providing information about community resources may be sufficient, particularly in observational studies. In some cases, a parent may not be aware that a particular situation presents a risk (e.g., improper application of pesticides, cleaning products in easy reach of children), and a sensitive, diplomatic discussion with the parent could resolve the situation. It is clearly not the researchers’ responsibility to eliminate all risks in homes.
In these situations, input from parents should be part of the plan. Specific responses will have to take into account parental values because the parents will have different views about childrearing and will have to live with the consequences of researchers’ actions. Although staff may intend to reduce risks, their actions may have unintended adverse consequences, such as disruption of the household, retaliation, or identification of a family that includes illegal immigrants or is living in overcrowded conditions in violation of a lease. Staff also need to understand why parents may be reluctant to accept a recommended action for reducing risk. Parents will often have legitimate concern that the intervention will not be as effective as expected or that the risk of unintended adverse consequences is greater than the researcher appreciates. Generally, the staff and parents can agree on an approach that both reduces risk and respects the parents’ values.
For example, a researcher might observe mold in the home. Mold presents a health risk to children and is a code violation in some jurisdictions. However, if more than one family is living in the unit, the parent may not want to report (or have the researcher report) the situation to the landlord for fear of eviction. An alternative might be to share information regarding how to control growth of mold, such as using particular cleaning products or using exhaust or other fans. In another situation, the family might not be aware that the landlord is responsible for the condition. This family might be advised to follow up with the landlord and be given information about possible community resources (e.g., legal services, code enforcement) that are available.
The informed consent process should discuss reasonably foreseeable situations in which researchers may have to breach confidentiality to protect study subjects from imminent and serious harm (e.g., reporting of suspected child abuse or neglect). The discussion should be tailored to the situations that are likely to arise in the project; researchers cannot give an exhaustive list of all situations in which confidentiality might have to be overridden.
If researchers observe risks that they believe require them to override confidentiality, they need to discuss confidentiality again with the parents. The discussion at the time of enrollment in the study was hypothetical; after a serious risk has been identified, the researcher needs to discuss the specific situation that poses the risk, the various alternatives for preventing the risk, and the parents’ preferences on how to proceed. Ideally, researchers and parents can agree on a course of action that resolves the risk without requiring a breach in confidentiality. However, this may not be possible in the case of suspected child abuse and neglect, particularly if the staff person is a neighbor and does not feel responsible for correcting the situation. In some cases, it may be possible to make an anonymous report to child protective services, which would be the appropriate response to ensure the safety of both the child and the staff. If the case needs to be reported to a responsible official (e.g., if a child in the study is being abused by someone else in the household), it is ethically desirable to obtain parents’ concurrence with reporting, as well as to help the parents and child find supportive services to protect them against retaliation. Such interactions show respect for parents.
Researchers should train their research field staffs about what they might encounter and how to respond to risks they observe in homes, develop policies and procedures to address them, and provide necessary oversight and back-up. Research staff who visit homes will be the ones facing the dilemma of how to respond to observed risks. Training should include information on privacy and confidentiality, any applicable legal obligations related to suspected child abuse and neglect (below), and the appropriate procedures for reporting such cases. When unanticipated situations arise, staff should be able to contact the principal investigator or a senior investigator for guidance on how to respond.
The research team should be aware of relevant available community resources to help enable them to respond in a timely, helpful manner. They should also identify specific contact people in those organizations for urgent assistance so staff can provide such information to parents.
Senior staff should be able to help resolve situations in which the immediate, intuitive response may be misguided. For example, field staff who notice firearms in the home within access of children may consider calling the police to prevent violence. However, doing so may cause many unintended problems, including embarrassment, recriminations by other family members, reprisals from a gun owner, or attempted eviction from housing. Harm can be averted in other ways, as by having the gun locked and unloaded. Unless situations present an imminent and serious harm, staff responses need to respect parents’ preferences.
IRBs are charged with protecting participants in research; hence, they need to review and approve researchers’ plan for responding to risks they observe in the home, including any anticipated situations that warrant overriding confidentiality.
CONCLUSIONS AND RECOMMENDATIONS
The committee reached several conclusions about researchers’ responsibilities. The particular context of housing health hazards research raises ethical obligations for researchers beyond those addressed in the applicable federal regulations. Community concerns related to the provision of benefit to all children enrolled in research can be addressed though the use of innovative designs. Research conducted in the home might affect third parties who are not participating in the research. In addition, researchers may observe a range of behaviors and conditions not related to the research. Researchers have an ethical obligation to thoroughly consider these issues and to inform the parents of potential child subjects of any circumstances under which the researchers anticipate they will need to breach confidentiality.
Recommendation 7.1: Researchers designing intervention studies on housing health hazards involving children should consider using innovative designs in which all child subjects receive a prospect of direct benefit.
Recommendation 7.2: Researchers carrying out research on housing health hazards involving children should discuss in their protocols and IRB submissions their legal and ethical obligations to potential third parties affected by their research.
Recommendation 7.3: Researchers designing research on housing health hazards to children need to anticipate the risks and behaviors that may be observed in the home, including observations that are not part of the research protocol, develop anticipatory plans that specify how to assess
and respond to risks when they are identified, and educate their staffs about the plan.
Recommendation 7.4: Institutional review boards that review housing health hazards research should examine the researchers’ plans for responding to risks observed in the home and require that they be appropriate in the context of the research and the affected community.