A medication safety program consisting of PGx-enriched comprehensive medication management (PGx + CMM) in a self-insured employee population resulted in favorable clinical outcomes in the year following the intervention. Program participation was associated with significantly fewer inpatient and emergency department visits in the 13-month follow-up period than those in the control group. In addition, the program showed potential economic benefits as measured by healthcare resource utilization (HRU) and costs in medical claims. These findings extend prior applications of real-world implementation of PGx + CMM into a broader population and offer cost savings potential for self-insured employers opting to provide a similar medication safety program to employees. Additionally, results further support attainment of a tipping point for population-level, large-scale pharmacogenomic testing.
The findings show a positive shift in HRU away from acute, expensive services and towards less costly outpatient care settings. Reducing hospital inpatient and emergency department admissions represent a favorable impact on HRU since emergency department visits and inpatient hospital admissions tax healthcare resources and often indicate missed proactive and preventive care opportunities. Costs of inpatient care are rising and account for ~ 27% of privately paid healthcare expenditures.15 Similarly, health spending attributable to emergency department visits is increasing in the U.S. and currently represents approximately 5% of total healthcare spending.16 Outpatient visits reflect healthcare engagement in primary and preventive care services,17 higher utilization of evidence-based preventive health measures,18,19 and an accepted strategy to prevent avoidable hospitalizations. In fact, observed increases in outpatient visits in the intervention group may be attributable to recommendations from program pharmacists for participants to follow up with their healthcare providers. Findings suggest that recommendations were successfully communicated by the pharmacist to the patient’s prescribers, resulting in more optimized medication management leading to significantly attenuated HRU. Combined, the evidence support that the PGx + CMM program favorably impacts HRU and offers a potential for cost savings at the population level.
The total pharmacy and medical costs were estimated to decrease by $128.31 PMPM in the intervention group when compared to the control group. Medical specific costs decreased by $172.24 PMPM for the intervention group. These savings estimates do not incorporate the cost of the program. While reductions in total medical costs were not statistically significant due to high variability in costs, observed savings track with the shift away from inpatient and emergency department services and are consistent with previous studies. 1,20 Additionally, pharmacy costs increased by $26.30 PMPM for the intervention group—not an unexpected result as healthcare providers may habitually prescribe less expensive medication before moving on to more costly alternatives. Prior to the intervention, the participants had lower total medical costs likely due to lower utilization of costly inpatient and emergency services compared to the control. However, even after matching for baseline differences in total medical costs between the groups, the PGx + CMM program was associated with a decrease in total medical costs for the intervention group.
These results build on the compelling evidence of the clinical and economic value of introducing PGx + CMM as a standard of care by expanding to a younger, employed population. While participants were invited into the program based on risk and utilization of medications with PGx implications, PGx interventions may have wider population impact as almost 65% of US adults may be exposed to at least one medication with an established pharmacogenomic association within a 5-year window.21 Further, it is estimated that 99% of individuals harbor a DNA variant known to impact medication safety and efficacy.22 This paper suggests that provisioning a PGx + CMM program across similar populations would yield positive clinical and economic impacts. Together, these outcomes provide extensive evidence that this program meets the requirements of a successful PGx + CMM implementation model.
At present, coverage for pharmacogenomics across most commercial plans is limited and reimbursement at the individual level continues to lag. This may be due—at least in part—
to the regulation of pharmacogenomic testing in the US. However, distinct from complicated insurance-based provisions is the opportunity for employers to implement a PGx + CMM program, thus no longer relying on third-party payers to reimburse for PGx testing and medication management services. Employers can thus remove third-party reimbursement constraints, while saving costs and improving the health of their employees. Furthermore, PGx + CMM implementations are being initiated by single-payer healthcare systems and countries23 lending additional credence to population-level PGx + CMM programs. These all suggest that, given no mechanism to fund this population-level activity within the established process of medical necessity decision-making at the patient level, self-insured employers, with more control over healthcare spend, might be interested in offering this program for employees.
Despite the positive results of the current report, findings should be interpreted in context and with an understanding of limitations. First, this paper reports findings of a real-world implementation as opposed to a randomized controlled trial. Although randomized controlled trials may provide the highest level of evidence, in clinical evaluations of pharmacogenomics ethical considerations arise on the assignment of individuals with pharmacogenetic risk to a control group.24 Specifically, such considerations may deem that it is not ethical to deny a person with known risk the access to an intervention known to be beneficial. It is also noted that real-life clinical evaluations carry the benefit of higher external validity with more transferable evidence to everyday clinical practice, despite any perceived lower level of evidence.24 Future study designs, along with larger samples of similar populations, which could be evaluated with a longer program timeframe, might provide additional scientific evidence to expand the conclusions of this paper. Additionally, this implementation was a voluntary employer-sponsored program, so selection bias may have contributed to the observed findings; however, this bias was addressed through statistical methodology and propensity score matching. Further research regarding why some employees, and not others, voluntarily participated in the intervention could enable more targeted engagement in future implementations. Moreover, several additional factors that have been associated with HRU such as demographic, socioeconomic, health services–related, health status–related, and health insurance, were not evaluated in the present study.25 In addition, individuals were invited to the program based on risk stratification and evidence of using ≥ 1 prescription medication. It may be that implementation in populations with lower medication utilization at baseline yield different results. However, potential differences can be mitigated by using a risk stratification process to identify individuals within the population most likely to benefit from the program.