Background/Purpose: Analgesics are the cornerstone of medical treatment of osteoarthritis (OA) but are associated with serious side effects. A more optimal use of analgesics in OA is one strategy to reduce the population health burden of OA. The purpose of this simulation study was to assess the impact of hypothetical changes in the use of analgesics on the quality-adjusted life years (QALYs) among persons with OA in the general population.

Methods: We used a previously validated population-based microsimulation model of OA in Canada, POHEM-OA. Model parameters were derived from a large administrative database in British Columbia, Canadian national surveys, and the scientific literature. The model included data on OA incidence, frequency of use of 4 classes of medication (acetaminophen, traditional non-steroidal anti-inflammatory drugs [NSAIDs], COX-2 inhibitors [coxibs], and opioids), treatment benefits in terms of pain reduction, frequency and quality-of-life (QOL) impact of adverse effects associated with each medication type (dyspepsia, ulcer, cardiovascular disease [CVD], stroke, and “other”) and overall QOL measured by the Health Utilities Index 3. We assessed the impact of changes in medication use on overall QALYs. In the base-case scenario we assumed that the relative risk of adverse effects due to medication is doubled among persons aged 70+ compared to those < 70 years of age. Alternative assumptions were tested in sensitivity analyses. The scenarios involved increases and decreases in the use of all medications and each medication type individually.

Results: Under the base-case scenario, an average person with OA accumulated about 13.3 QALYs from OA diagnosis till death. Increasing the odds of using analgesics by 50%, 100% and 200% resulted in incrementally greater improvements in overall QALYs. However, the improvements were small (0.05-0.15 QALYs) and were observed only for opioids and acetaminophen. Increasing NSAIDs had virtually no effect on overall QALYs. Targeting persons with moderate or severe pain ( >2 or >3 on a 5-point scale) for aggressive treatment (increasing the odds of treatment 3-fold) achieved a greater overall benefit than targeting all OA patients. Reducing the odds of using all analgesics by 50% among older (70+) patients with OA did not affect QALYs, whereas eliminating NSAIDs in this group had a slight beneficial effect at the population level. These results were sensitive to the assumptions about the relative risk of adverse effects among the elderly.

Conclusion: We found a beneficial but relatively small impact on QALYs of more aggressive treatment of OA with acetaminophen and opioids, and no benefit from increasing NSAIDs. For coxibs, even though gastrointestinal complications were reduced compared to traditional NSAIDs, the benefits of pain reduction were virtually nullified by the increased risk of CVD and stroke. Overall, the results seem plausible and support the validity of the simulation model. A limitation of the current model is that it does not consider the long-term harmful effects of opioids, including addiction and overdose.

« Back to 2019 ACR/ARP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/impact-of-hypothetical-changes-in-the-use-of-analgesics-on-the-burden-of-osteoarthritis-a-population-based-microsimulation-study/