I’m currently working on a project which is identifying the key barriers which are slowing down the translation of whole genome sequencing into clinical practice, and as a result I’ve been digging into the literature on priority setting and genomics (on the basis that one barrier might be resource constraints). To be honest, this hasn’t taken a lot of time, as it’s not a particularly well-researched area. That said, there were two specific papers that have informed the development of our work in this area, and I thought it might be interesting to bring these to the attention of a wider audience.
The first paper by Severin and colleagues (“Value judgments for priority setting criteria in genetic testing: a discrete choice experiment”) used a discrete choice experiment (DCE) to investigate the factors which inform prioritisation decisions between different genetic tests. DCE respondents around the world (mainly clinical geneticists, lab scientists and patient reps) were asked to assume that a decision maker from a health care organisation had been asked to decide which of two tests should be provided if resources were only sufficient to provide one of them. They were then asked to indicate which test they thought the decision maker should choose.
The tests differed in terms of disease severity, disease risk, test benefit, test cost and aim of the test, and respondents were willing to trade between all of these attributes. However, the most important attributes were judged to be proven medical benefit from the test, high risk of having the condition in question and a low test cost. Interestingly, interactions between these latter two attributes and a dummy for clinical geneticist respondents were significant, suggesting some heterogeneity amongst respondents.
Preferences for prioritisation amongst geneticists and patients are, however, only half of the picture. If their desired tests are commissioned, we then need to know something about the preferences of those who will be ordering the tests if we want to be able to say something about the likely health consequences. After all, clinician preferences could be an even bigger barrier to translation than resource constraints.
Blumenschein and colleagues took a small step towards addressing this uncertainty by conducting the Severin DCE (with some slight variations) in Canada and recruiting respondents from several medical specialties (including neurologists, endocrinologists and haematologists), as well as medical geneticists and genetic counsellors. Their results indicated that proven medical benefit was again the most important factor to respondents, followed by the detection rate of the test (which was included as an attribute instead of disease risk due to the slightly different sample population). Test cost was the least important attribute in this survey. So, in a population which contains more of those respondents who will actually be ordering these tests, attributes which capture the benefits of testing feature even more prominently in decision-making.
What does this tell us? Probably not that much to be honest, as only around a third of the 190 respondents were actually clinicians. In addition, it’s likely that the preferences of clinicians regarding which tests should be available are probably quite different from their preferences for specific tests when they have a patient sat in their consultation room. This is a much more complex decision scenario. For example, clinicians are likely to be faced with more than two options, one of which could well be not ordering a test. So I’m not sure that we’re any closer to establishing which potential barrier is more significant (resource constraints or clinician preferences). However, this is an issue that we’re hoping to address in our current project, and we’ll hopefully have some results to share in the coming months.