OBJECTIVES: To assess the relationship between DGIR and HRU in elderly patients to evaluate potential benefit from pharmacogenetic testing.  METHODS: A retrospective cohort of patients age ≥65 years was identified through Inovalon’s MORE2® registry with continuous enrollment, taking ≥3 prescription medications (July 1, 2012 - March 31, 2013) and on ≥1 drug metabolized by a polymorphic drug metabolizing enzyme. Patients were stratified into zero, low, medium, and high DGIR groups via a diagnostic test (Genelex Youscript®).  Counts of HRU during 9 months follow-up post index-date (date of first claim for ≥1 drugs with pharmacogenetic implications) included all-cause hospitalizations, emergency-room and clinic visits. Poisson regression was used to test the association between DGIR and HRU counts. The model was adjusted for age, gender, race, Charlson Comorbidity Index (CCI) and number of known drug-drug interactions. RESULTS: A total of 252,184 patients were included and the mean age was 74 ± 6 and 60% were female. The median DGIR score was 8.7% [IQR 2.3%-44%]. There were 59,559 (23.6%) with a DGIR score of zero, 82,224 (32.6%) with low risk (0-20%), 33,439 (13.3%) with medium risk (20-40%), and 76,962 (30.5%) with high risk (>40%).  Regression analysis revealed that the low and medium DGIR groups were associated with a 9% (95%CI 8.4 to 9.7%, p <0.0001), and 8% (95%CI 7.7 to 9.2%, p <0.0001) increase in the rate of HRU, compared to zero risk.  The high DGIR group was associated with a 5% (95%CI 4.2 to 5.3%, p <0.0001) decrease in HRU, compared to zero-risk.  CONCLUSIONS: Among elderly patients, low and medium DGIR groups were associated with increased rates of HRU. In contrast, high DGIR was associated with lower HRU rates. This may be explained by a time-dependent effect of changing DGIR as a result of medication changes over time.