To summarize our conclusions, we suggest that the following four criteria may be used to define research subjects:
1. An individual who is directly intervened upon by an investigator for research purposes;
2. An individual who is deliberately intervened upon via manipulation of his/her environment by an investigator for research purposes;
3. An individual with whom an investigator interacts for the purpose of collecting data;
4. An individual about whom an investigator obtains identifiable private information for the purpose of collecting data.
A detailed discussion of these criteria and their implications for CRTs follows.
A research subject is an individual whose interests may be compromised as a result of interventions in a research study including:
1. An individual who is directly intervened upon by an investigator for research purposes
Individuals are research subjects if, in the context of a research study, they are the recipients of an experimental intervention (active or control) or if they undergo an intervention to collect data, such as an additional examination.
If an intervention is targeted at individual cluster members but random assignment is done at the cluster level (typically to avoid treatment contamination or for logistical reasons) then the individuals receiving the intervention should be considered research subjects. In healthcare, this would include CRTs evaluating therapeutic or health promotion modalities aimed at individual patients, as well as CRTs evaluating new modes of health service delivery. An example of the former includes a CRT evaluating the effect of individualized exercise prescriptions for patients, randomized by physician practice40. An example of the latter includes a CRT evaluating the effectiveness of asthma management using specialist nurses34. In these studies, the individuals themselves are being
manipulated, and should therefore be considered research subjects.
In healthcare CRTs, the intervention under study is often not administered to patients, but rather to healthcare professionals, and the outcomes are evaluated using patient data. It may be reasonably asked whether the health professionals who receive an educational intervention in a CRT are research subjects or collaborators. Collaborators are individuals who contribute to the design of, or participate in the conduct of, a research study.
Collaborators are not recipients of experimental interventions. In healthcare CRTs, the health professionals are receiving an experimental educational or quality improvement intervention. When they are directly intervened upon in this way, health professionals participating in a CRT meet the definition of a research subject5,6,41.
Some healthcare CRTs evaluate complex interventions that may include
combinations of health professional education, novel modes of health service delivery and patient-level interventions. An example is a CRT evaluating a primary care program aimed
at reducing obesity42. In this trial, primary care providers in the intervention arm received an educational intervention on motivational techniques and physical activity for obese patients. Patients in the intervention arm were screened for obesity, and obese patients were counseled by the physician and referred to local sports foundations to receive
individualized exercise counseling. In determining whether or not patients need be considered research subjects for these CRTs, an ethics committee needs to examine each intervention in such a CRT to determine whether a particular intervention is directed at patients, or whether data collection includes interaction between researchers and patients or the use of identifiable private information. Interventions on health professionals mean that the health professionals will be research subjects. Patients will not necessarily be research subjects because of interventions on health professionals. But, patients will be research subjects if there are patient-level interventions (either therapeutic interventions or direct interventions to collect data) or novel modes of health service delivery are used, if researchers interact with patients, or if the study uses patients’ identifiable private information to evaluate outcomes.
2. An individual who is deliberately intervened upon via manipulation of his/her environment by an investigator for research purposes
Individuals who are intervened upon via manipulation of their environment are research subjects. This includes individuals who will be affected by CRTs of public health interventions, whether the unit of randomization is a municipality, a neighborhood, a family, or some other group whose environment may be manipulated. Because these
individuals are research subjects, they are entitled to regulatory protections, including the determination by a research ethics committee that the risks to their interests do not outweigh the potential benefits offered by the CRT. Many of these studies would meet regulatory criteria for a waiver of informed consent because the interventions in these studies pose only minimal risk and would likely be unfeasible without a waiver of consent.
We concluded above that the indirect effects that a CRT may have on an individual are not sufficient to warrant considering that individual to be a research subject. In
healthcare CRTs, patients may be indirectly affected by educational or quality
improvement interventions that are directed at healthcare professionals or institutions. The physicians under study continue to have an obligation to act in patients’ best interests, and have no competing obligations to the study itself. The physician-patient relationship is preserved. If there are no patient-level interventions, if the researcher has no interaction with individual patients, and there is no use of identifiable private information for research purposes, patients are not research subjects.
3. An individual with whom an investigator interacts for the purpose of collecting data Any individual from whom an investigator, in the context of a research study, obtains data through any kind of interaction should be considered a research subject. Interaction includes any kind of communication or interpersonal contact between investigator and subject, for example interviews, focus groups, or questionnaires. Such modes of interaction may be employed in CRTs when collecting data reported by individual cluster members.
Any data collection through interaction means that the respondents are entitled to protections as research subjects8,11.
4. An individual from whom an investigator obtains identifiable private information for the purpose of collecting data.
Obtaining identifiable private information about individuals within a cluster will make these individuals research subjects, and therefore make them entitled to protections. Conversely, there is no risk to an individual’s privacy if the researchers are only collecting anonymized or aggregate group-level information43. Individuals whose data have been anonymized before transfer to the investigators, or whose administrative or health-related information is used to generate aggregate measures for a cluster are not research subjects unless they are manipulated in some other way.