Limitations of applying summary results of clinical trials to individual patients: The need for risk stratification
ABSTRACT There is growing awareness that the results of randomized clinical trials might not apply in a straightforward way to individual patients, even those within the trial. Although randomization theoretically ensures the comparability of treatment groups overall, there remain important differences between individuals in each treatment group that can dramatically affect the likelihood of benefiting from or being harmed by a therapy.1- 4 Averaging effects across such different patients can give misleading results to physicians who care for individual, not average, patients.
- SourceAvailable from: Egbert R te Velde
[Show abstract] [Hide abstract]
- "). Investigating how the prognostic score from a model affects the treatment response, rather than the individual treatment predictive factors which made up the score, is called a risk-stratified analysis (Kent, 2007). Due to heterogeneity in the treatment protocols of the included trials in the Van den Boogaard study it was not possible to combine the individual patient data from each trial to conduct a meta-analysis. "
ABSTRACT: Infertility is defined as failure to conceive after 1 year of unprotected intercourse. This dichotomization into fertile versus infertile, based on lack of conception over 12-month period, is fundamentally flawed. Time to conception is strongly influenced by factors such as female age and whilst a minority of couples have absolute infertility (sterility), many are able to conceive without intervention but may take longer to do so, reflecting the degree of subfertility. This natural variability in time to conception means that subfertility reflects a prognosis rather than a diagnosis. Current clinical prediction models in fertility only provide individualized estimates of the probability of either treatment-independent pregnancy or treatment-dependent pregnancy, but do not take account of both. Together, prognostic factors which are able to predict natural pregnancy and predictive factors of response to treatment would be required to estimate the absolute increase in pregnancy chances with treatment. This stratified medicine approach would be appropriate for facilitating personalized decision-making concerning whether or not to treat subfertile patients. Published models are thus far of little value for decisions regarding when to initiate treatment in patients who undergo a period of, ultimately unsuccessful, expectant management. We submit that a dynamic prediction approach, which estimates the change in subfertility prognosis over the course of follow-up, would be ideally suited to inform when the commencement of treatment would be most beneficial in those undergoing expectant management. Further research needs to be undertaken to identify treatment predictive factors and to identify or create databases to allow these approaches to be explored. In the interim, the most feasible approach is to use a combination of previously published clinical prediction models.Human Reproduction 07/2014; 29(9). DOI:10.1093/humrep/deu173 · 4.59 Impact Factor
Theory & Psychology 04/2014; 24:256-277. DOI:10.1177/0959354314525282 · 0.70 Impact Factor
- "Sample means from a group of individuals permit inferences about the population average, but these means do not permit inferences to individuals unless it is demonstrated that the mean is, in fact, representative of individuals. Surprisingly, it is rare in psychology to see the issue of representativeness of an average even mentioned, although recently, in the domain of randomized clinical trials in medicine, the limitations attendant to group averages have been gaining increased mention (e.g., Goodman, 1999; Kent & Hayward, 2007a, 2007b; Morgan & Morgan, 2001; Penston, 2005; Williams, 2010). "
[Show abstract] [Hide abstract]
- "Also, conventional subgroup analyses with one-variable-at-a-time approach would easily fail to identify the subgroup that should be described simultaneously with multiple characteristics (Hayward, Kent, Vijan, & Hofer, 2006). As a way to cope with the problem of the univariate approach, Hayward et al. (2006) and Kent and Hayward (2007) advocated multivariate risk-stratified subgroup analysis, which builds a risk score by combining multiple patient characteristics and compares subgroups based on the risk score along with the treatment effect. Although this approach has the advantage of increasing the statistical power of detecting treatment heterogeneity across subgroups, its limitations include: (a) it requires the independent development of risk-prediction tools prior to the particular study (Kent et al., 2002) and those tools should be adapted and validated for the specific RCTs and (b) unlike the conventional subgroup analysis, it has no ability to examine individual factors that directly modify the treatment effect (Hayward et al., 2006). "
ABSTRACT: In randomized controlled trials (RCTs), the most compelling need is to determine whether the treatment condition was more effective than control. However, it is generally recognized that not all participants in the treatment group of most clinical trials benefit equally. While subgroup analyses are often used to compare treatment effectiveness across pre-determined subgroups categorized by patient characteristics, methods to empirically identify naturally occurring clusters of persons who benefit most from the treatment group have rarely been implemented. This article provides a modeling framework to accomplish this important task. Utilizing information about individuals from the treatment group who had poor outcomes, the present study proposes an a priori clustering strategy that classifies the individuals with initially good outcomes in the treatment group into: (a) group GE (good outcome, effective), the latent subgroup of individuals for whom the treatment is likely to be effective and (b) group GI (good outcome, ineffective), the latent subgroup of individuals for whom the treatment is not likely to be effective. The method is illustrated through a re-analysis of a publically available data set from the National Institute on Drug Abuse. The RCT examines the effectiveness of motivational enhancement therapy from 461 outpatients with substance abuse problems. The proposed method identified latent subgroups GE and GI, and the comparison between the two groups revealed several significantly different and informative characteristics even though both subgroups had good outcomes during the immediate post-therapy period. As a diagnostic means utilizing out-of-sample forecasting performance, the present study compared the relapse rates during the long-term follow-up period for the two subgroups. As expected, group GI, composed of individuals for whom the treatment was hypothesized to be ineffective, had a significantly higher relapse rate than group GE (63% vs. 27%; χ (2) = 9.99, p-value = .002).