Like any medical test genetic tests have to have analytical and clinical validity. Often medical geneticists and laboratory specialists have stopped there; satisfied that if a new genetic test leads to improved diagnostic rates that it is worth using for patients. That view is based on a strong intuition that getting to a molecular diagnosis leads to patient and family benefit. We haven’t developed a large body of evidence or conducted many clinical trials that show those anticipated benefits. Other branches of medicine have faced similar questions. How can we measure Clinical Utility of genetic tests? What can we learn from other fields of medicine?
The first challenge is that there are thousands of rare and ultrarare genetic diseases. We can study a few of the more common ones individually but then we get stuck. There are too many diseases with too few patients to conduct clinical trials one by one. Clinical Utility (CU) refers to configurations of diseases, gene variants, and therapeutic responses. To deal with thousands of rare diseases we need aggregation strategies. Measures that can be applied across many diseases among the categories of clinical presentations that doctors are asked to diagnose. Some of those categories, not exhaustive, are intellectual disability, autism & delayed development, birth defects, epilepsy, heritable connective tissue disease, inborn errors of metabolism, cardiovascular disorders. The list cuts across every discipline of medicine.
Robin Hayeems introduced me and the other participants in the Medical Genome Initiative to the Fryback and Thornbury framework for measuring Clinical Utility. It’s not the only way to approach genetic testing, but is very appealing to me as a practitioner because it focuses on the practical changes that can be enabled by doing genetic tests. There are others who put greater emphasis on the Personal Utilities. Those are real too and a major aspect of the goals that both doctor and patient share.
Diagnostic Thinking is how the doctor works through all the information about a patient to arrive at a diagnosis. Having the right diagnosis should trigger the best possible treatment. Doctors memorize hundreds of patterns of disease and when a common one comes along they can immediately recognize the pattern. Students and trainees spend years learning these patterns and the basic science that supports them. But the expert practitioner does not have to remember all that basic science or invoke it to accurately recognize the disease pattern and efficiently act on that conclusion. The huge clinical challenge with genetic disorders is that even the most experienced doctors will not have encountered many rare diseases and most ultrarare diseases. That basic science knowledge of biochemical and developmental pathways has to be invoked for a large fraction of patients. Because of limitations in current knowledge, it turns out that we aren’t very good at it. It is common for people with rare diseases to have a long diagnostic odyssey, visiting many doctors, with many misdiagnoses and false-negative diagnoses. Genetic tests and especially Genome Sequencing can radically change this situation. They take the doctor off the hook for accurately recognizing each genetic disease. A good genetic test can lead to a molecular diagnosis even if the doctor has never seen the condition before. How can we measure an improvement in diagnostic thinking? One element of diagnostic thinking is forming a differential diagnosis. Sometimes the clinical presentation is so specific that the differential diagnosis includes only a handful of possible explanations. But for many clinical disorders, the genetic differential diagnosis can include hundreds of conditions. Narrowing a differential diagnosis is really useful. Confirmation of a strongly suspected diagnosis is also useful. Refuting a diagnosis or correcting a misdiagnosis is useful. Genetic test results can trigger additional lab and imaging tests, both to nail down the diagnosis and to assess for brewing complications that have not yet shown themselves. Another aspect of diagnostic thinking is the ability to predict prognosis. If a doctor can be more certain about the prognosis that can lead to much better clinical decision making – a shared endeavor with the patient and family.
Change in management is another big dimension of Clinical Utility. If the doctor has only a descriptive diagnosis, then general interventions can be used. Sometimes it is possible to use a genetic and molecular diagnosis to motivate much better tailoring of therapy. One of the common changes in management is a referral to another specialist to help monitor and perhaps avoid additional complications. Some medications used to treat symptoms might be added or taken away. The holy grail is the situation where a new treatment can be added that directly attacks the mechanism of disease. The easiest to understand is gene correction therapy, just at its earliest stages of clinical investigation.
Genetic tests can have personal and social utilities. Getting a specific genetic diagnosis sometimes leads to changes that avoid mortality or reduce long-term disability. These outcomes are the hardest to measure because they take large numbers of patients studied over a long period of time. It might be that target trial emulation strategies and quasi-experimental approaches to large cohort databases will be a more feasible way to address such questions.
belmontgenetics.org 