ENRICHMENT STRATEGIES FOR CLINICAL TRIALS
Author: Athira Sudhakaran – Biostatistics Lead at Genpro Research.
Enrichment trials use patient characteristics to select a study population in which detection of a drug effect is more likely than it would be in an unselected population.
There are three categories of enrichment strategies for Clinical Trial:
- Strategies to decrease heterogeneity.
- Prognostic enrichment strategies in clinical trial.
- Predictive enrichment strategies in clinical trial.
Strategies to Decrease Heterogeneity
There are many commonly used strategies to decrease subject heterogeneity including carefully defining inclusion and exclusion criteria, using placebo lead-in periods to eliminate patients that improve spontaneously or excluding patients taking pharmacologically similar drugs. Excluding patients with a comorbid illness that would make completing the treatment period unlikely will also help to reduce variability.
Prognostic Enrichment Strategies in Clinical Trial
Prognostic enrichment is the selection of patients that are more likely to have a specific, disease‑related endpoint. The guidance points out that to properly power a study, the sample size needs to be based on the “effect size and event rate in the placebo group”. Prognostic enrichment will allow an increase in the absolute effect size while keeping the relative effect the same. Examples of prognostic enrichment strategies include clinical and laboratory measures and gene expression profiles. These measurements predict the prognosis of the patients, but not the predictive value of drug efficacy.
Predictive Enrichment Strategies in Clinical Trial
Predictive enrichment in clinical trial is the identification and enrolment of patients that are likely to respond to the study drug. Predictive enrichment increases both the absolute and relative effect of the study drug. Pre-randomization screening for responders and screening for proteomic or genomic markers related to a drug’s mechanism of action are an example of predictive enrichment strategies.
In one of the device studies, we have worked to evaluate the performance of an automated device for the detection of diabetic retinopathy. At the start of the study, all participants with diabetes who met inclusion and exclusion criteria were enrolled. Then later to avoid excessive enrolment in any one stratum (no/mild, moderate DR, or vision-threatening DR), the totals monitored monthly and the study population enriched to ensure sufficient numbers of subjects with more than mild DR. Patient enrichment was targeted based on elevated HbA1C or fasting blood glucose (FBG), factors which have been shown to be correlated with higher rates of vision-threatening DR among patients with diabetes.
Wish to know more? Feel free to write to us at email@example.com.