现在的位置: 首页临床话题, 基本知识>正文
[JAMA统计学与方法学指导]:病例对照研究:利用真实世界的证据评价相关性
2018年10月05日 临床话题, 基本知识 暂无评论

JAMA Guide to Statistics and Methods

August 23, 2018

Case-Control Studies: Using “Real-world” Evidence to Assess Association

Telba Z. Irony

JAMA. Published online August 23, 2018. doi:10.1001/jama.2018.12115

Associations between patient characteristics or treatments received and clinical outcomes are often first described using observational data, such as data arising through usual clinical care without the experimental assignment of treatments that occurs in a randomized clinical trial (RCT). These data based on usual clinical care are referred to by some as “real-world” data. A key strategy for efficiently finding such associations is to use a case-control study.1 In a recent issue of JAMA Internal Medicine, Wang et al2 assessed the association between cardiovascular disease (CVD) and use of inhaled long-acting β2-agonists (LABAs) or long-acting antimuscarinic antagonists (LAMAs) in chronic obstructive pulmonary disease (COPD), utilizing a nested case-control study.

Explanation of the Method

What Are Case-Control and Nested Case-Control Studies?

A case-control study compares individuals who had the outcome of interest (cases) vs individuals who did not have that outcome (controls) with respect to exposure to a potential “risk factor.” The goal is to determine if there is an association between the risk factor and the outcome. The risk factor may be a behavior such as tobacco use, a patient characteristic, or a treatment. The idea is to define a population or cohort, identify the cases and controls in the population, and retrospectively determine which patients in each group were exposed to the risk factor; the case-control study works backward from outcome to exposure (Figure). A higher proportion of individuals with exposure to the risk factor among cases than among controls suggests that the risk factor is associated with the outcome. The term control refers to an individual who did not have the outcome; in contrast, the same term in a clinical trial refers to a study participant who receives the standard (or placebo) treatment.

In a nested case-control study, the cases are identified in a large cohort and, for each case, a specified number of controls matching the case are selected from the cohort. The selected controls should match the cases with respect to characteristics, other than the risk factor, that are likely related to the outcome of interest. Because it is easier to find controls than cases when the outcome is rare, increasing the number of controls beyond the number of cases (eg, 2:1 or 3:1 matching) may be used to improve study precision.

The nested case-control study by Wang et al2 used data from 284 220 LABA-LAMA–naive patients with COPD retrieved from the Taiwan National Health Insurance Research Database with health care claims from 2007 to 2011. Cases (n = 37 719) were patients who had inpatient or emergency care visits for coronary artery disease, heart failure, ischemic stroke, or arrhythmia (CVD events). Each patient was matched to 4 controls (n = 146 139) without visits for these disorders.

In a case-control study, the most common measure of association between exposure and outcome is the odds ratio (OR), which aims to compare the occurrence of the outcome in the presence of the exposure vs in the absence of the exposure.3 In practice, the OR in a case-control study is the ratio of the odds of exposure among the cases to the odds of exposure among the controls, where the odds of exposure is the probability of exposure divided by the probability of no exposure. The prevalence of the exposure is compared between cases and controls and not the other way around. However, because the OR treats outcome and exposure symmetrically, it provides the desired measure of association. If the OR is greater than 1, the exposure is associated with the outcome, ie, having the exposure increases the odds of having an outcome (and vice versa). The OR is a measure of effect size; the larger the OR, the stronger the association.

In the study by Wang et al,2 new use of LABA occurred in 520 cases (1.4%) and 1186 controls (0.8%), resulting in an adjusted odds ratio of 1.50 (95% CI, 1.35-1.67). New use of LAMA occurred in 190 cases (0.5%) and 463 controls (0.3%), resulting in an adjusted odds ratio of 1.52 (95% CI, 1.28-1.80). An OR of 1.5 represents a modest association4 between outcome (CVD) and exposure (LABA and LAMA). Thus, the authors found that new use of LABAs or LAMAs was associated with a modest increase in cardiovascular risk in patients with COPD, within 30 days of therapy initiation.

Why Are Case-Control Studies Used?

Case-control studies are time-efficient and less costly than RCTs, particularly when the outcome of interest is rare or takes a long time to occur, because the cases are identified at study onset and the outcomes have already occurred with no need for a long-term follow-up. The case-control design is useful in exploratory studies to assess a possible association between an exposure and outcome. Nested case-control studies are less expensive than full cohort studies because the exposure is only assessed for the cases and for the selected controls, not for the full cohort.

Limitations of Case-Control Studies

Case-control studies are retrospective and data quality must be carefully evaluated to avoid bias. For instance, because individuals included in the study and evaluators need to consider exposures and outcomes that happened in the past, these studies may be subject to recall bias and observer bias. Because the controls are selected retrospectively, such studies are also subject to selection bias, which may make the case and control groups not comparable. For a valid comparison, appropriate controls must be used, ie, selected controls must be representative of the population that produced the cases. The ideal control group would be generated by a random sample from the general population that generated the cases. If controls are not representative of the population, selection bias may occur.

Case-control studies provide less compelling evidence than RCTs. Due to randomization, treatment and control groups in RCTs tend to be similar with respect to baseline variables, including unmeasured ones.5 Because the only difference between treatment and control groups is the treatment, RCTs can demonstrate causation between treatment and outcome. In case-control studies, case and control groups are similar with respect to the matching variables, but are not necessarily similar with respect to unmeasured variables. Such studies are susceptible to confounding, which occurs when the exposure and the outcome are both associated with a third unmeasured variable.6 Unlike RCTs, case-control studies demonstrate association between exposure and outcome but do not demonstrate causation.

The objective of case-control studies is to compare the occurrence of an outcome with and without an exposure. The relative risk (RR), which is the ratio between the probability of the outcome when exposed and the probability of the outcome when not exposed, provides a straightforward comparison measure but, because the case-control study design does not allow for the estimation of the occurrence of the outcome in the population (ie, incidence or prevalence), the RR cannot be determined from a case-control study. A case-control study can only estimate the OR, which is the ratio of odds and not the ratio of probabilities. The OR approximates the RR for rare outcomes, but differs substantially when the outcome of interest is common. In addition, case-control studies are limited to the examination of one outcome, and it is difficult to examine the temporal sequence between exposure and outcome.

Despite these limitations, case-control studies and other “real-world” evidence can provide valuable empirical evidence to complement RCTs. Additionally, case-control studies may be able to address questions for which an RCT is either not feasible or not ethical.7

How Was the Method Applied in This Case?

In the case-control study by Wang et al,2 the exposure to LABA and LAMA use for both cases and controls in the year preceding the occurrence of the CVD event was measured and stratified by duration since initiation of LABA or LAMA into 4 groups: current (≤30 days), recent (31-90 days), past (91-180 days), and remote (>180 days). Additional stratification on concomitant COPD medications and other factors was also conducted. The data source used in the study (Taiwan National Health Insurance Research Database) mitigates data quality concerns because it is national, universal, compulsory, and subject to periodic audits. Overall, the authors found that new use of LABAs or LAMAs was associated with a modest increase in cardiovascular risk in patients with COPD, within 30 days of therapy initiation, and this finding was strengthened by the steps taken to ensure data quality and comparability of cases and controls.

How Does the Case-Control Design Affect the Interpretation of the Study?

Causality cannot be established in a case-control study because there is no way to control for unmeasured confounders. In the study by Wang et al,2 the use of the disease risk score for predicting CVD events was helpful to control for measured confounders but could not adjust for unmeasured confounders. The authors mitigated further possible confounding effects by conducting extensive sensitivity analyses.

给我留言

您必须 [ 登录 ] 才能发表留言!

×
腾讯微博