html xmlns=”http://www.w3.org/1999/xhtml” xmlns:mml=”http://www.w3.org/1998/Math/MathML” xmlns:epub=”http://www.idpf.org/2007/ops”>
A key factor in generalizability is the target population, the people who are the object of the study. In this chapter we discuss the target population versus the actual study pool of available individuals. We focus on the eligibility criteria, since this is critical to the validity (scientific integrity) of the study. The actual individuals who participate in the study, the study group, are recruited from the group of eligible individuals. The study design must specify in detail which individuals within the study pool are eligible for the study to ensure the validity of the study.
12.1 The Target Population and the Study Pool
The target population is the largest group of people to whom the results of your study may apply. It may be very general, for example simply “being human” for studies of basic physiological mechanisms, or very narrow, such as a group with a carefully defined and very homogeneous disease. Within the target population we define the study pool, which is the group of individuals who could actually be observed in the study. The study pool could be the same as the target population but is usually a subset of the target population observed in one or more specific locations while the study is being done. Sometimes the study pool is not representative of the target population, or you may not have access to enough individuals in the target population in your area to test your hypothesis. The study pool may be expanded to resemble the target population more closely by designing a multi-center study, in which several centers or hospitals all follow the same protocol and use the same techniques.
In the study described in Example 11C on the use of a nutritional supplement during pregnancy, the treatment was to be started in the fifth month of pregnancy. Thus, the study pool would be limited to women in prenatal care by the fifth month of their pregnancy. This means that the study pool may underrepresent certain groups of women who are at risk for early delivery, such as intravenous drug users, because they are less likely to have prenatal care. The investigators plan to take extra steps, such as working with community outreach programs, to try to recruit more of these women.
In a physiological study of the impact of an insulin clamp, the goal was to determine what was abnormal in the response of Type 1 diabetics compared to a normal physiological response. This requires a group of “normal healthy individuals.” Although it is straightforward to define Type 1 diabetics, it is much harder to define a “normal healthy individual.”
12.2 The Eligible Group
Within the study pool, specific criteria for inclusion and exclusion will identify a subset of individuals who are eligible for the study. The eligibility standards defined in the study protocol are a set of physical, demographic, and medical standards that the individual must meet to be included (inclusion criteria) and another set of standards that they must not meet or else they are excluded (exclusion criteria). The specification and implementation of these criteria are critical to the success of the study. They must be spelled out precisely and without ambiguity, or the results of the study may not be interpretable or credible.
The first and most important inclusion criterion is that a participant be able to understand and give informed consent, usually a signed document. Chapter 3 describes the details of creating an informed consent document, presenting it, and discussing the details with a participant in an interview. Most of the time this process is straightforward, but if your study involves special populations (such as pregnant women, children, and individuals with impairments), then special procedures apply (Section 3.4). Your Institutional Review Board (IRB) will be able to provide advice and assistance for these special populations. Sometimes informed consent is waived (Section 3.6), most usually for record reviews (Chapter 9). Your IRB can provide information about waivers of consent as well, and will have to authorize it, if appropriate for your study, before your study begins.
Table 12.1 is a list of some participant characteristics in addition to informed consent that are often specified in inclusion or exclusion criteria. This list cannot be exhaustive, as each study will have specific requirements. Also, not all categories on the list will be important for every study. Items in most categories can be either inclusion or exclusion criteria. Usually it is easier to understand a requirement when it is presented positively, so that the exclusion criterion “allergic to drug” is easier to understand than the inclusion criterion “not allergic to drug,” but the investigator decides which way to present it.
|
The protocol must specify methods for assessing the variables that comprise the inclusion and exclusion criteria. Many criteria include lists of items, such as other diagnoses or medications. In general, if a list in the inclusion criteria is all wanted items, others are not of interest unless they are in the exclusion criteria. If a list is in the exclusion criteria, then it is assumed that items not on this list are allowed.
In the study in Example 10C to assess whether patients presenting with an AIDS-defining opportunistic infection (OI) should be treated first for the OI and defer antiretroviral treatment (ART) until the OI was resolved, or should be treated concurrently with ART, detailed criterion for each specific OI was needed. For some OIs the specific criteria for diagnosis varied by site, as in some parts of the United States certain OIs are frequently seen and thus are treated empirically without definitive diagnosis, while at other sites the specific OI is sufficiently uncommon that physicians routinely do a definitive workup prior to initiating treatment for the OI.
Diagnosis is probably the most common criteria, and frequently the focus of the study. A study protocol might specify what diagnosis the participant must have and what diagnoses would make the participant ineligible for the study. The method of ascertaining the diagnosis must be spelled out in detail. If laboratory tests are used, what are the ranges for pathology? May prior or current results from the participant’s usual laboratory tests be used, or must the participant be tested at a central laboratory? If it is a physical limitation, how will it be tested and what range of results is considered diagnostic? If the diagnosis is derived according to a manual of diagnosis such as the Diagnostic and Statistical Manual of Mental Disorders (DSM), how is the information required for the diagnosis obtained? Similarly, if information is from a current history, what sources are considered reliable?
In a study of treatment for depression in an outpatient population, the primary inclusion criteria were that participants be diagnosed with major depression. This was defined according to the current DSM, using data from the Structural Clinical Interview for the DSM (SCID). The SCID includes criteria for deriving DSM diagnoses, and the two together are accepted standards of diagnosis in the field of depression. Patients with major depression frequently have other psychiatric disorders. In this study the investigators decided to allow the presence of comorbid generalized anxiety disorder if major depression was also present. This was stated specifically in the inclusion criteria. All other DSM comorbid disorders, such as obsessive-compulsive disorder or posttraumatic stress disorder, were exclusion criteria, since there was evidence that these disorders affected the response to one of the drugs being studied.
Diabetes mellitus is included in the eligibility criteria in many studies, either as an inclusion requirement, when it is being studied, or as an exclusion criterion. In some cases the medical record may be sufficient to ascertain the presence or absence of diabetes. If not, the criteria for diagnosing Type 1 diabetes are well defined, but different methods may be used for diagnosing Type 2. For some studies a fasting blood glucose measurement is sufficient, while others may require that the participant undergo an oral glucose tolerance test.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
