Associations of Mental Health and Family Background With Opioid Analgesic Therapy

A Nationwide Swedish Register-Based Study

Patrick D. Quinn; Martin E. Rickert; Johan Franck; Amir Sariaslan; Katja Boersma; Paul Lichtenstein; Henrik Larsson; Brian M. D'Onofrio


Pain. 2019;160(11):2464-2472. 

In This Article

Abstract and Introduction


There is evidence of greater opioid prescription to individuals in the United States with mental health conditions. Whether these associations generalize beyond the US prescription environment and to familial mental health and socioeconomic status (SES) has not been examined comprehensively. This study estimated associations of diverse preexisting mental health diagnoses, parental mental health history, and SES in childhood with opioid analgesic prescription patterns nationwide in Sweden. Using register-based data, we identified 5,071,193 (48.4% female) adolescents and adults who were naive to prescription opioid analgesics and followed them from 2007 to 2014. The cumulative incidence of any dispensed opioid analgesic within 3 years was 11.4% (95% CI, 11.3%–11.4%). Individuals with preexisting self-injurious behavior, as well as opioid and other substance use, attention-deficit/hyperactivity, depressive, anxiety, and bipolar disorders had greater opioid therapy initiation rates than did individuals without the respective conditions (hazard ratios from 1.24 [1.20–1.27] for bipolar disorder to 2.12 [2.04–2.21] for opioid use disorder). Among 1,298,083 opioid recipients, the cumulative incidence of long-term opioid therapy (LTOT) was 7.6% (7.6%–7.7%) within 3 years of initiation. All mental health conditions were associated with greater LTOT rates (hazard ratios from 1.66 [1.56–1.77] for bipolar disorder to 3.82 [3.51–4.15] for opioid use disorder) and were similarly associated with concurrent benzodiazepine-opioid therapy. Among 1,482,462 adolescents and young adults, initiation and LTOT rates were greater for those with parental mental health history or lower childhood SES. Efforts to understand and ameliorate potential adverse effects of opioid analgesics must account for these patterns.


Opioid analgesic prescription has increased and remains quite common, especially in North America,[25,33] and adverse public health outcomes have raised serious concerns about opioid safety. Patients receiving greater-quantity opioid regimens are at increased risk of morbidity and mortality.[7,8,17,22,47,55] Questions remain, however, about the contribution of prescribed opioid use—vs comorbid conditions, diversion, and illicit opioid use[64]—to substance use disorder (SUD) and related harms.[11,12,54] Evaluating real-world opioid prescription patterns is therefore necessary to understanding their impact on health.

Studies of prescription in the United States, often using commercial health care insurance records, have identified a number of predictors of opioid receipt. Critically, patients with preexisting SUD and other mental health conditions are substantially more likely to receive long-term opioid therapy (LTOT) than are patients without those conditions.[9,13,15,26,51,52] Because such individualsmay be at greater risk of harmful opioid outcomes, this phenomenon has been termed "adverse selection."[73] A major question is whether these findings are specific to the US prescription environment or whether they reflect processes more fundamental to patients with chronic pain—and those receiving LTOT in particular. Limited large-scale health care–record evidence supports a similar pattern among primary care,[19] postsurgical,[32] and other patients outside the United States. For example, benzodiazepine and other psychiatric medication use is overrepresented among prescription opioid recipients.[10,24,45] Moreover, Nordic and United Kingdom studies combining health care records with surveys and other clinical and population data have linked various indices of mental health (eg, depression, smoking, SUD) to LTOT.[5,21,23,27] At present, conclusions from this literature are constrained by assessment of few mental health factors, often with self-report surveys or prescription records. To the best of our knowledge, no study has yet comprehensively examined the course of opioid therapy among patients with mental health conditions outside the United States.

The current study helped address this gap using nationwide register data from Sweden, where, although opioid prescription has historically been relatively conservative, use and accompanying concerns have increased.[33,58] Specifically, this study characterized opioid initiation and LTOT for up to 8 years, across analgesia indications, for patients with a wide range of mental health conditions (eg, SUD, self-injurious behavior). In addition, we examined associations with concurrent benzodiazepine-opioid therapy, a prescription pattern that may contribute to risk of opioid overdose[76] and has therefore been discouraged in Sweden[46,56,65] and elsewhere.[14] Finally, given clearly established genetic influences on mental health,[50] as well as socioeconomic disparities in opioid prescription,[44,77,78] research is needed to more fully examine family backgrounds for these patient populations.[39] We therefore incorporated register-based family mental health history and socioeconomic status (SES).