6 May 2014
One of the most significant barriers to the successful introduction of genomic medicine in health systems around the world is the difficulty clinicians and scientists face in determining which of the millions of variants in our genomes are harmless and which might be contributing to a risk (high or low) of disease now or in the future. This problem is most simply expressed as the question: how do we decide whether or not genotype X is implicated in producing phenotype (disease) Y? A new set of guidelines published by an influential group of US-based genomics experts from both clinical and academic settings aims to address the role that researchers play in providing high quality, reliable evidence to answer this question.
The authors touch on a number of issues in their report, and I would encourage those with an interest in this area to read it in full (it’s open access). I will focus my comments on the area in which policy makers can probably make the biggest difference, and that is around access to and quality control of research-generated genomic knowledge that is used in clinical care.
Unlike other areas of biomedical research, where there are multiple intervening steps in the bench to bedside translation process during which erroneous results may get weeded out, a newly published research finding of the association between a genetic variant and a rare disease may be contributing to the clinical diagnosis of a patient within days of publication without any validation beyond a peer review process (that is itself not without flaws). Whilst clinical scientists and geneticists draw on many sources of evidence when making diagnoses, it is clear that evidence from the research literature and research focused databases form a key part of this decision making. Given that many published associations between genetic variation and disease are wrong, but remain uncorrected in databases and the published literature, there is a pressing need to improve standards of research in this field to reduce as far as possible serious cases of harm to patients from incorrect diagnoses.
The authors of the guidelines emphasised that underlying genomic variant data and structured descriptions of the strength of the evidence base for any interpretation of medical significance should be made freely available through existing databases. From a policy perspective, this point is crucial for two reasons:
A number of databases are in place (or under development) around the world that should provide the platform researchers and clinicians need to facilitate the types of genomic data sharing, access and curation that will be the lifeblood of genomic medicine. The challenge is to ensure that the quality and quantity of data in these databases is maximised, and this is where health policy makers have a role to play. Whilst the current incomplete and sometimes inaccurate nature of many genomic databases would suggest that existing incentives and enforcement of standards of research practice, largely undertaken by funders, are far from optimal, it may be unreasonable to expect researchers to fulfil the significantly more onerous requirements for accuracy placed on clinicians.
Policy makers will therefore have to judge whether or not the type of community derived best practice guidelines proposed by MacArthur et al. to improve the quality of the research-derived evidence base for genomic medicine will be sufficient to deliver the quality required for its increasingly widespread use in clinical care. Currently, in the UK, it is the job of our clinical genetic scientists to effectively compensate for the acknowledged weaknesses of the evidence base they rely upon by careful curation and filtering in each case they investigate. Whilst this model is sustainable against a background of low volume genetic testing, it would appear incompatible with the envisaged future of higher test volumes that will come with widespread use of genomic testing across medical specialities.
Here at the PHG Foundation we are exploring these issues as part of our Clinical Genome Analysis project with an expert group of clinicians and scientists involved in the delivery of genomic medicine in the UK. In our next briefing note from the project we will focus on analysing current standards of best practice in genomic analysis for clinical care and considering how they may need to adapt to be fit for the genomic era. We will also consider the steps UK policy makers could take to improve the quality of the evidence on which this analysis is based.