Pharmacogenomics (PGx) ? guiding drug therapy based on individuals' genetic make-up - offers significant potential to improve drug outcomes, as 97% of the US population carries at least one potentially actionable variant. Providing PGx test results using clinical decision support (PGx-CDS) alerts embedded in health records may be a useful way to guide drug therapy. Yet, Learning Health Systems (LHSs) resist adoption because they are unsure of the value of investing in CDS systems for PGx. Determining the value of providing population-specific PGx-CDS alerts would enable LHSs to invest in those alerts that have the greatest potential for improving health care quality and outcomes in their specific populations. The objective of this proposal is to help LHSs make informed decisions about the implementation of PGx-CDS alerts specific to their populations that consider trade-offs between the cost of implementation and the potential clinical benefits to patients. Using decision modeling, in Aim 1 we will create a framework for estimating the value of PGx-CDS alerts.
In Aim 2, we will adapt this framework to an online platform, creating a publically available, web-based tool that will enable customized estimates of the value of PGx-CDS alerts specific to each LHS. We will pilot and improve the tool by collaborating with stakeholder-colleagues in LHSs. Our work is responsive to the priorities outlined in AHRQ PA-17-246 in that it: 1) provides an innovative, evidence-based HIT solution to manage population health and improve quality and outcomes within LHSs in a way that makes the solution configurable across disparate LHSs, and 2) provides a platform to share and analyze practice data in a way that makes knowledge learned actionable and sharable, including tailoring messages to decision-makers. Our project is aligned with the priorities of AHRQ in that it will make healthcare safer and costs transparent, while simultaneously creating efficiencies.
Pharmacogenomics (PGx) is well-positioned to be the first genomics application to realize the promise of precision medicine. Active PGx clinical decision support (PGx-CDS) alerts are proposed as one preferred method for returning PGx test results; the value of providing PGx-CDS alerts has not been formally evaluated. We will create a framework (Aim 1) and publically available web-based tool (Aim 2) to assist Learning Health Systems in investing in PGx-CDS alerts of greatest value to their specific population. 1
