The practice of medicine can be roughly divided into two specialties: the diagnosis of disease and the treatment of disease. Historically, healthcare investing has been focused on treatment, with pharmaceutical and biotech companies commanding the majority of attention. However, diagnosis remains a significant opportunity for the informed investor. Diagnostic products are pervasive in the healthcare system and as varied as the diseases the humans encounter. This report provides an overview of both the diversity and the commonalities in this market.
Sensitivity and specificity are the statistics that are most commonly invoked when describing a diagnostic test. They are the rates of true positives and true negatives respectively and are frequently cited because in theory they do not depend on the prevalence of the disease being examined. However, in practice these values are highly contingent on the circumstances under which they are looked at. Moreover, they only tell half of the story, and the value of a test is inseparable from how it fits into the real world environment of a disease. Therefore, it is important to understand how these data were gathered and how these circumstances reflect the environment in which the disease is encountered.
In the US, most diagnostics are regulated as medical devices. As part of this, they are stratified based on their level of risk, with three independent approval pathways for standalone tests: premarket approval (PMA) for high risk tests, De Novo request for low risk tests, and 510(k) for tests with a previously approved predicate (which is the vast majority of applications). Review times are generally six months to a year. Tests attached to a single laboratory can also be regulated as laboratory developed tests (LDTs). In Europe, the process is both more lenient and more complicated with marketing approval primarily through the CE marking system, but with numerous state-level controls and different regulations for devices and in vitro diagnostics. Also, the majority of in vitro diagnostics do not get CE marking before entering the market. The EU is in the middle of a multi-year harmonization process to require all new tests to receive a CE mark and oversight.
In the early stages, companies may employ retrospective studies to examine the parameters of their tests. Although these studies are useful, low-cost methods of validating a test, they lack the rigor of prospective studies and are susceptible to ‘overfitting’, where the data perfectly fits an individual trial but cannot be replicated. Therefore, prospective trials are the gold standard for evaluating clinical validity. However, there is a standard beyond validity, where the value of test is evaluated. Clinical utility is the measure of how the diagnostic changes clinical practice. These trials are not always required for regulatory approval but can be essential for inclusion into medical recommendations and for reimbursement.