Impact of pain severity and opioid use on health care resource utilization and costs among patients with knee and hip osteoarthritis

Abstract

BACKGROUND: Osteoarthritis (OA) is the most common form of arthritis. The primary symptom of OA—pain—increases the disease burden by negatively affecting daily activities and quality of life. Opioids are often prescribed for treating pain in patients with OA but have questionable benefit-risk profiles. There is limited evidence on the economic impact of pain severity and opioid use among patients with OA. OBJECTIVES: To (a) evaluate the association of pain severity with health care resource utilization (HRU) and costs among patients with knee/hip OA and (b) characterize the association of opioid use with HRU and costs while controlling for pain severity. METHODS: Using deterministically linked health care claims data and electronic health records from the Optum Research Database, this retrospective cohort study included commercial and Medicare Advantage Part D enrollees who were diagnosed with knee/hip OA during the January 1, 2010-December 31, 2016 time period and had ≥1 pain score (11-point Likert scale: 0=no pain, 10=worst possible pain) between the first OA diagnosis date and October 2016. The index date was the date of the evaluated pain score; HRU and costs were observed over the 3-month post-index period. For patients with multiple pain scores, each episode required a 3-month post-index follow-up period. Generalized estimating equation models, adjusted for multiple observation panels per patient and baseline variables that may contribute to HRU and costs (age, sex, race/ethnicity, region, insurance type, integrated delivery network, body mass index, pain medication use, provider specialty, Charlson Comorbidity Index score, and other select comorbid conditions), were used to estimate all-cause and OA-related costs expressed as per patient per month (PPPM). Comparisons were performed for moderate (score 4-6) and severe (score 7-10) pain episodes versus no/mild pain episodes (score 0-3) and those with baseline opioid use versus those without. RESULTS: Included were 35,861 patients with knee/hip OA (mean age 66.5 years; 64.7% women) who had 70,716 pain episodes (58% mild, 23% moderate, 19% severe, and 37.0% with baseline opioid use). When controlling for other potential confounding factors, moderate/severe pain episodes were associated with higher all-cause and OA-related HRU than mild pain episodes. Relative to mild pain episodes, moderate/severe pain episodes were also associated with significantly higher adjusted average all-cause PPPM costs ($1,876/$1,840 vs. $1,602), and OA-related PPPM costs ($550/$577 vs. $394; all P<0.05). Baseline opioid use was associated with significantly higher all-cause PPPM costs versus no opioid use (mild, $1,735 vs. $1,492; moderate, $2,034 vs. $1,755; severe, $2,100 vs. $1,643, all P<0.001), and a higher likelihood of incurring OA-related costs relative to those without (P<0.001). CONCLUSIONS: These results provide evidence of the economic impact of opioid use and inadequate pain control by demonstrating that increased pain severity and opioid use in patients with OA were independently associated with higher HRU and costs. Additional studies should confirm causality between opioid use and HRU and costs, taking into consideration underlying OA characteristics.