Given the high burden of pain in PWH and the need for effective pain management, it is important to understand how PWH use substances in the context of pain.^[14] There is some evidence that cannabis may be effective for alleviating pain, and the use of cannabis is likely to continue to increase as a result of legalization of recreational cannabis in California and other states.^[15] In our primary care-based sample, we found that PWH who reported moderate/high pain were more likely to report opioid and cannabis use compared to PWH with none/mild pain. These findings highlight the importance of understanding how these substances are used by PWH, to improve pain management and minimize risks to patient health.

Recent studies have found a decrease in pain medication use (both opioids and non-opioids) for Medicare Part D and Medicaid populations in states where medical cannabis laws were approved.^[16] Our study observed prescription opioid use in 1 region of California and found prevalent use of cannabis in PWH, approximately 40% in the none to mild pain group and the moderate to severe pain group. This study was conducted when medical cannabis laws were approved in the state of California, yet concerns were growing about opioid use risks, and prescribing policies were becoming more conservative.^[17] The legalization of recreational cannabis laws can affect the use of cannabis for pain^[17] and future studies should assess how PWH utilize substances other than prescription opioids to manage pain. One research team found that depression, alcohol consumption, and marijuana use were not significantly associated with current pain nor with severe pain.^[18] Our study found that depression and marijuana use were significantly associated with pain in the adjusted model. Our definition of unhealthy alcohol use was 4 or more drinks for women and 5 or more drinks for men in a day for 4+ days in the past 30 days, whereas this prior study defined alcohol consumption as either hazardous or nonhazardous as 14 standard drinks per week for men and 9 per week for women.^[18] Pain is one of the most common medical reasons for marijuana use,^[14,19] and although our study found a significant relationship, future studies are needed to better understand the use of marijuana for pain in PWH. In addition to understanding how PWH use substances to manage pain, it is important for future studies to address depression and to differentiate between chronic pain and neuropathic pain.

Several sociodemographic factors were associated with pain in our sample, including age, unknown/other race/ethnicity, education, and employment. These results are also of interest in understanding those most in need of assistance in managing pain. Lawson et al^[2] (2015) found that among PWH in the UK, increasing age was an independent risk factor of pain. Our study also found that older age was a risk for moderate to severe pain. Older PWH report poorer physical quality of life^[20] and thus our findings also suggest that pain should be an important consideration in the aging population. Joseph et al^[18] found that the odds of having any pain was 2.49 times greater in PWH who reported anxiety. In addition to anxiety, holding multiple stressors and marginalized identities, such as being African American, impact the mental health of PWH.^[21] Our study did not find a significant difference in Black versus White race/ethnicity in pain severity, which supports the inconclusive evidence that MSM who identify as African American have elevated HIV incidence or related symptoms.^[21] The role of race and structural racism in regards to HIV should not be overlooked.^[21] Graduate school and higher compared with less than high school and being employed were also predictors of pain. This relationship is different from prior studies where lower education and lower employment rates were associated with pain.^[22] Employment status could be associated with pain if patients are working inside or outside the home that can cause mental or physical strain. Usually, PWH who have substantial pain report impairments in work,^[22] and future research should explain substance use and pain in PWH who are employed.

There were limitations to this cross-sectional study. First, our sample was predominantly insured MSM from 1 city who were enrolled in an alcohol treatment study. Despite potential reduced generalizability, a benefit of this approach is enhanced internal validity with validated measures of exposure and outcomes for all study participants. An additional limitation was the use of a single-item to measure pain severity and there are other pain scales that could be used to measure pain in greater detail.^[19] Patients with none or mild pain may report as such given, they are getting pain properly managed with the prescription opioids. In this cross-sectional study, we were not able to determine if the use of opioids resulted in reduction in pain, such that patients using opioids were more likely to report mild or no pain. Further research is needed to understand the use of opioids and pain management in PWH, and the implications of opioid use by patients with relatively low pain levels. Another limitation is that we did not obtain information about why our patients used substances. We report correlations, but due to the cross-sectional nature of the study we are unable to determine causes of substance use.

The relationship of depression to pain and use of these substances is important to understand among PWH. Prior research has found that PWH with moderate to severe pain are more likely to have depressive symptoms and use prescription opioids compared to those without pain.^[23] Furthermore, structural or social determinants of health such as homelessness and discrimination potentially precipitate psychiatric comorbidities and substance use for all PWH.^[21] Our population was mainly White PWH and future research should consider causes of pain in more diverse populations as pain is not well-studied among non-White PWH.^[24] Although our study sample was similar with respect to demographics of the overall clinic population, the sample was enrolled from a clinical trial. This limits generalizability, for example, to PWH who do not use alcohol. In the last decade, the prevalence of opioid use has changed, with a reduction in opioid prescribing and much higher use of illicit, new synthetic, and non-prescribed opioids. Furthermore, social determinants of health and incarceration have led to more drug injections resulting in negative consequences for PWH and the need for medications for opioid use disorder need to be scaled up.^[21]