Regulatory and reimbursement decisions for precision medicine tests are based in large part upon a body of evidence around three performance characteristics of the test: analytical validity, clinical validity, and clinical utility. We will first review these concepts and then discuss the current regulatory and reimbursement environment in the United States.
Analytical validity refers to the accuracy of the laboratory test, that is, how well the assay detects what it is intended to detect. In the case of genetic testing, it refers to how well the testing platform accurately measures and reports a given genetic variant, for example. Analytical validity is usually measured in terms of the test's sensitivity and specificity, which capture information about the likelihood of true versus false positive or true versus false-negative results.
Clinical validity addresses if the test is clinically meaningful, for example, how well a test can predict a clinical outcome (e.g., disease development, disease progression, or efficacy/toxicity of a drug). These measures are usually derived from statistically robust observational studies. Typical metrics derived from these studies include relative risk, hazard ratios, sensitivity and specificity (sometimes referred to as clinical sensitivity and specificity in this case to distinguish it from analytical sensitivity and specificity discussed above), positive predictive value (PPV), and negative predictive value (NPV) of the test. The strongest evidence comes from synthesizing a body of evidence, as is done for clinical practice guidelines and systematic reviews.
In terms of the measures of effect, the clinical PPV indicates how often someone with a specific genotype will experience the outcome. It is effectively the same as penetrance, a genetic term that denotes the probability that a person who tests positive for the genetic marker will develop the outcomes (disease in this case). Clinical PPV is highly dependent on how common the outcome is. A test for a rare disease or adverse event will never have a high PPV for that outcome. In these cases, NPV may be the more relevant metric. A high NPV usually indicates that most if not all of the people who develop the disease or adverse event will test positive for the specific biomarker. Therefore, a negative test is used to rule out the possible occurrence of the event. This is the case for a genetic test for HLA type and celiac disease. Celiac has a population prevalence of about 1% and virtually all patients with celiac have one of two common HLA variants, DQ2 or DQ8. For this relatively uncommon disorder, this translates into a high NPV, or the ability to rule out celiac in patients who lack these HLA variants.
High clinical PPV is difficult to achieve for common, complex diseases, where the genetic variant associated with the disease is common as well, in large part ...