Sensitivity and specificity of self-reported psychiatric diagnoses amongst patients treated for opioid use disorder

Abstract

Background Patients with opioid use disorder (OUD) frequently present with comorbid psychiatric illnesses which have significant implications for their treatment outcomes. Notably, these are often identified by self-report. Our study examined the sensitivity and specificity of self-reported psychiatric diagnoses against a structured diagnostic interview in a cohort of patients receiving outpatient pharmacological treatment for OUD. Methods Using cross-sectional data from adults receiving outpatient opioid agonist treatment for OUD in clinics across Ontario, Canada, we compared participants’ self-reported psychiatric diagnoses with those identified by the Mini Neuropsychiatric Interview (MINI) Version 6.0 administered at the time of study entry. Sensitivity and specificity were calculated for self-report of psychiatric diagnoses. Results Amongst a sample of 683 participants, 24% (n = 162) reported having a comorbid psychiatric disorder. Only 104 of these 162 individuals (64%) reporting a comorbidity met criteria for a psychiatric disorder as per the MINI; meanwhile, 304 (75%) participants who self-reported no psychiatric comorbidity were in fact identified to meet MINI criteria for a psychiatric disorder. The sensitivity and specificity for any self-reported psychiatric diagnoses were 25.5% (95% CI 21.3, 30.0) and 78.9% (95% CI 73.6, 83.6), respectively. Conclusions Our findings raise questions about the utility of self-reported psychiatric comorbidity in patients with OUD, particularly in the context of low sensitivity of self-reported diagnoses. Several factors may contribute to this including remittance and relapse of some psychiatric illnesses, underdiagnosis, and the challenge of differentiating psychiatric and substance-induced disorders. These findings highlight that other methods should be considered in order to identify comorbid psychiatric disorders in patients with OUD.

An Artificial Intelligence–Based Smartphone App for Assessing the Risk of Opioid Misuse in Working Populations Using Synthetic Data: Pilot Development Study

“…That means if any participant responds higher scale in questions 1-6, they are most likely to be at a high risk of OUD. Compared with the moderate and normal OUD risk described by the bar chart in Figures 4 and 5, the participants responded higher in the mid (7-10) and low (10)(11)(12)(13)(14)(15) part of the survey questionnaires order.…”

“…Some other psychosocial and psychiatric disturbances, including cultural influences, social support, comorbid mood disorder, and drug abuse, were found to be associated with higher levels of chronic pain, which could significantly enhance the risk of opioid overuse [12,13]. Additionally, there are some behavioral risk factors for OUD, such as chronic substance use, current or past substance abuse, family history of substance abuse, overconsumption of alcohol, posttraumatic stress disorder, and physical abuse [14][15][16][17].…”

Exploring dual diagnosis in opioid agonist treatment patients: A registry-linkage study in Czechia and Norway

“…Previous studies have found co-occurring mental disorders in 65% to 87% of patients in OAT (Aas et al, 2021;Morin et al, 2020;Roncero et al, 2016). The heterogeneity in prevalence estimates across existing studies is presumed to be partially attributed to methodological differences (Leung et al, 2021;Santo et al, 2022) and socio-cultural aspects (Carrà et al, 2015). Nevertheless, the consistent observation of prevalence rates well exceeding 50% elsewhere across studies indicates a considerable burden of cooccurring mental disorders in this population.…”

Exploring dual diagnosis in opioid agonist treatment patients: a registry-linkage study in Czechia and Norway

“…Previous studies have found co-occurring mental disorders in 65–87% of patients in OAT [ 4 , 13 , 14 ]. The heterogeneity in prevalence estimates across existing studies is presumed to be partially attributed to methodological differences [ 1 , 42 ] and socio-cultural aspects [ 28 ]. Nevertheless, the consistent observation of prevalence rates well exceeding 50% elsewhere across studies indicates a considerable burden of co-occurring mental disorders in this population.…”