Posted on by Scott D. Grosse, National Center on Birth Defects and Developmental Disabilities, Centers for Disease Control and Prevention, Atlanta, Georgia
Genetic testing in children has traditionally focused on conditions with clinical actionability or utility. However, parents may want to know whether their child is at high risk of a rare disease even if a treatment doesn’t exist. A newly published article reports on a study conducted by researchers at RTI International and the University of North Carolina at Chapel Hill that assessed parents’ perceptions about genomic testing on their children. Researchers queried parents about factors that would make them want to know genomic sequencing results for their children that relate to disabling or life-threatening disorders for which effective treatments are not currently available. The researchers used a discrete choice experiment (DCE) method, also referred to as conjoint [considered jointly] analysis, to quantify the strength of preferences over seven different characteristics of a test or a hypothetical condition. This is the first peer-reviewed study that has jointly considered these various aspects, although some of these attributes were considered in a formative test of a pilot conjoint survey developed by RTI in 2010.
The study asked 1,289 parents of young children to compare profiles of hypothetical genetic tests with different characteristics. The authors found the most important attribute to be the predictive value of a genotype for the development of clinical disease, i.e., disease penetrance. Tests for conditions where a positive result was associated with only a 10% chance of developing clinical disease were found to be distressing to parents whereas there was a positive preference for tests with 75% or 90% predictive value of disease. Assuming an accurate test for a highly penetrant genotype, parents favored knowing results for a condition with early onset by 5 years of age that have moderate to rapid progression of symptoms and that cause mental or physical disabilities or result in a shortened lifespan. The more severe the disabilities, the greater the parents’ interest in knowing whether their child had the condition. The one characteristic that parents did not care about very much was whether options were available to improve or ameliorate the impact of the condition on a child’s quality of life. Finally, the authors found no significant differences by demographic strata – gender and race/ethnicity.
The study findings have implications for genomic testing or screening for highly penetrant childhood-onset rare diseases that cause serious physical and mental problems or early death, e.g., spinal muscular atrophy (SMA). The study referred specifically to the return of genomic sequencing information for children in either a clinical or public health context. However, the same considerations may apply to a broader range of genetic testing applications in various settings – preconception, prenatal, neonatal, or early childhood. Decisions on which conditions should be prioritized for genetic testing in any of those settings have rarely been informed by information on parental preferences.
For example, at one point there were conflicting professional society recommendations on whether carrier testing for SMA should be routinely offered during prenatal care to all women (the American College of Obstetrics and Gynecology in 2017 endorsed routine screening for SMA). In the context of that debate, I wrote an editorial in 2010 that called for DCE/conjoint studies of parental preferences to inform prioritization of conditions for carrier screening. That editorial referred to the 2010 RTI pilot study, noting “This method also has been applied to voluntary newborn screening for rare disorders lacking effective treatments to elicit parental valuations on the relative importance of disorder prevalence, test sensitivity, developmental disability, physical disability, and life expectancy.” It concluded more broadly, “Research is needed on the perceived value to stakeholders, including the public, of carrier screening for SMA in particular and for rare disorders in general.” I am gratified by the appearance of this new study which addresses that research need and does so in an elegant and statistically robust fashion.
Knowledge of average preferences in one sample of people who were asked to choose among hypothetical tests for hypothetical conditions is informative but must be combined with many other factors in informing decisions about the appropriateness and prioritization of genomic testing. From a public health perspective, newborn screening targets conditions for which there is good evidence not only of medical actionability but evidence of the health impact of timely identification – prevention of death or disability. What this study shows is that parents may still wish to know if their child is affected by a serious rare disease even if public health criteria on impact are not satisfied.
Examination of preference heterogeneity could provide additional insights. Latent class analysis methods, when applied to DCEs, allow for the identification of groups of people who care differently about specific attributes. In particular, some people may place no value on an attribute that others value positively. For example, a recent DCE study of preferences on mammography screening found that many women place a positive value on getting a diagnosis of breast cancer, regardless of whether it improves outcomes, whereas others place no value on a diagnosis independently of health outcomes. How people value different aspects of genomic test results can affect their uptake of genetic tests. Therefore, it could be useful for decision makers to have information on the distribution of preferences.