[Figure 1 here]
Description of included studies (Table 1) : Most studies were conducted in the United States (103 studies, 82%), followed by Europe (16 studies, 13%), six which were conducted in Norway. The number of studies escalated over time, with 86 studies (68%) published between 2015 and 2019, of which 26 (20.6%) were published in 2019 (Jan-Jul). Cohort studies were the most common (92 studies, 73%), followed by cross-sectional studies (32 studies, 25%).
Data sources (Table 1): Data on opioid use was retrieved from pharmacy dispensing records for 108 studies (86%), with almost half (58 studies, 46%) using data from nationwide administrative systems. Pharmacy dispensing claims or prescription records were used with one (70 studies, 56%) or multiple data sources (42 studies, 33%), such as medical health records, medical claims, hospitals separations, surveys, census, and surveillance registries (e.g., mortality, cancer).
Study Populations (Table 1): One-quarter of studies did not impose age restrictions, 64% included adults only and 9% focused exclusively on the elderly (aged 65+ years);[20-30] two studies focused on adolescents and young adults (13-29 years old). [31, 32]
Study populations varied substantially in size, ranging from 121 [33] to 48 million people. [34] One-fifth of studies included only opioid-naïve individuals (27 studies), defined as the absence of opioid use in the six or 12 months prior to the index date. Studies investigated opioid use among patients with various specific health conditions, such as CNCP (42 studies, 33%), musculoskeletal conditions (15 studies, 12%), injuries or trauma (13 studies, 10%) or the infectious diseases HIV or hepatitis C (8 studies, 6%). Six studies (5%) included patients with both cancer pain and CNCP without reporting stratified results [35-40] and six studies (5%) evaluated patients with cancer [26, 41, 42] or cancer survivors. [27, 43, 44]