relative 399.rel.002 Louis J. Sheehan, Esquire

Here we provide recommendations to improve predictions and support continued growth of this industry.


(Ng, Murray, and Venter, 2009). My thinking was that the industry-friendly bias affords reduced skepticism of the article’s criticisms of the industry. Additionally, the Ng et al. (2009) article references a “ … statement of conflicting interests … “ which adds to its persuasiveness. The bottom-line?

Agreement on risk predictions by DTC [direct to consumer] companies does not necessarily imply that the predictions are accurate or meaningful, and at this point in time [October 2009], we cannot determine who has the ‘best’ predictions. To effectively assess the clinical validity of these genetic tests the community needs more prospective studies with tens or hundreds of thousands of individuals that measure the predictive value of known markers. [Footnotes omitted.] Such studies are useful because they consider risk markers simultaneously, measure the interaction between different markers and do not assume a risk model. It may be practical to prioritize common diseases with significant health impact because of the large numbers of individuals and the expense associated with prospective studies.



(Ng et al., 2009). Apparently, then, the connection between the DTC predictions and expected realty is -- hopefully – an educated guess. Tellingly, when Ng et al. (2009) submitted the DNA of five people to two of the largest DTC companies, at best

… only two-thirds of relative risk predictions qualitatively agree between 23andMe and Navigenics.


Ethically, then, how does one assess handling this level of results’ ‘meaning’? In a nutshell, one of Wright and Keoese’s (2010) conclusions is:

However, to offer such tests to interested citizens who wish to investigate their own genomes or participate directly in the research process, is appropriate so long as providers are transparent about the evidence base for the test and offer appropriate levels of support.