•Opioid-dependent patients with chronic pain are more likely to misuse benzodiazepines.•Opioid-dependent patients with chronic pain are more likely to misuse cannabinoids.•Opioid-dependent patients with chronic pain are not more likely to misuse opioids.•This pattern of substance misuse is shown to persist in the longer term. Aims: To compare specific substance misuse in treatment-seeking, opioid-dependent patients with and without comorbid chronic pain, and to assess the respective value of urinalysis and patient reports in assessing substance misuse. Methods: Participants comprised a clinical population in a regional NHS Substance Misuse Service in the East of Scotland (N = 521). The Brief Pain Inventory – Short Form was used to assess pain, and the Maudsley Addiction Profile and urinalysis were used to assess substance misuse at study inception. Urinalysis was used to assess substance misuse during the 5-year follow-up period. Data were hosted, linked, anonymized and analyzed within a national Safe Haven. Results: Compared with opioid-dependent patients with no pain, a significantly higher proportion of treatment-seeking, opioid-dependent patients with chronic pain were engaged in non-medical benzodiazepine use (69% versus 58%; p = 0.016) and illicit cannabinoid use (84% versus 65%; p = 0.025) at study inception. Furthermore, a significantly higher proportion of this group was shown to continue non-medical benzodiazepine use (70% versus 42%; p = 0.037) and illicit cannabinoid use (100% versus 31%; p = 0.002) during the 5-year follow-up period. There were significant correlations between drug screen results and patient-reported use of opioids (Tetrachoric ϱ = 0.4944; p < 0.001), benzodiazepines (Tetrachoric ϱ = 0.2641; p = 0.001) and cannabinoids (Tetrachoric ϱ = 0.8384; p < 0.001). Conclusions: Whilst gaining control of illicit opioid use during treatment, opioid-dependent patients with comorbid chronic pain demonstrated persistent problematic use of benzodiazepines and cannabinoids. This pattern of misuse was shown to persist during the 5-year follow-up period.