PCN153 - Quantifying the Association between Health-Related Quality of Life Scores and Subsequent Health Resource Utilization in Oncology

Abstract

Although HRQL scores have been linked to HCRU in the general population, this link has not been clearly demonstrated in oncology. We quantified the association between differences in scores from the SF-12v2 Health Survey (SF-12v2) and subsequent HCRU among noninstitutionalized US adults with cancer. We analyzed the Medical Expenditure Panel Survey (MEPS) data, including patients 18 years of age or older with a completed SF-12v2, 6 months of subsequent HCRU data, and a reported diagnosis of cancer not in remission (n=647). HCRU included total medical expenditures (MEs) and number of medical events (EVs). We modeled the effect of physical (PCS) and mental (MCS) component summary scores on MEs using a gamma-log generalized linear model (GLM) controlling for demographics and comorbidities. A Poisson-log GLM was used for EVs. Estimates were obtained for the total sample and major cancer subgroups: breast, prostate, skin and lung. A one-point higher (better) PCS score was associated with 2% lower MEs (P<0.001); a one-point higher MCS score was associated with 2.5% lower MEs (P=0.015). A 3-point higher PCS or MCS score (proposed MID for these scores) was associated with 6% and 7.5% lower MEs respectively. Within the cancer subgroups, a 1-point higher PCS score was most strongly associated with MEs for patients with prostate cancer (7% lower MEs; P=0.009) followed by patients with skin cancer (5% lower MEs; P=0.019). We found less association between PCS and EVs, and no evidence that this association varied by cancer group. We showed that differences in PCS and MCS scores from the SF-12v2 are significantly associated with differences in subsequent HCRU in a national oncology sample. The key role of HRQL in oncology makes these findings a potential guide in health policy and decisions on the evaluation of outcomes.