Background: Access to drug substances with mood rewarding effects such as opioids, central nervous system stimulants, sedatives, hypnotics and intravenous anaesthetics is tightly controlled by law globally. Those which are clinically approved have strict prescription and accountability requirements throughout the supply chain, and are collectively called prescription drugs or controlled prescription drugs (CPDs), while some such as heroin, cocaine, marijuana and methylenedioxymethamphetamine (MDMA) are prohibited from clinical use and are called illicit drugs. Despite their essential uses in clinical medicine, access to CPDs is tightly controlled because nonmedical use elicits drug use disorders, namely, drug misuse, drug abuse, and drug dependence. Nonmedical use of CPDs and drug use disorders has a reciprocal relationship with neurological impairment and severe mental disorders that undermines effective treatment of mental disorders. Therefore, mitigating drug use disorders among mental health patients is important. This requires understanding the burden and factors associated with CPD nonmedical use among mental health patients. This dataset was compiled from a cross-sectional survey to determine the prevalence of CPD and illicit drug nonmedical use and associated factors among patients attending mental health services in Uganda.

Data collection: A cross-sectional survey of 1275 patients was performed at three referral mental health facilities in Uganda from November to December, 2018 using interviewer-administered semi-structured questionnaires, desk review guide, and urine drug assays. The questionnaire recorded CPD nonmedical and illicit drug use history from patients’ files, CPD lifetime use, and risk factors. We defined CPD nonmedical use as either a positive urine assay for CPDs without current prescription for the drug(s) in patient’s file or a documented clinician’s diagnosis of CPD drug use disorder in a patient’s lifetime in their hospital files. On the other hand, lifetime use was defined as having used at least one CPD from a list of 22 commonly prescribed CPD products indicated on the questionnaire.

Data processing: Numerical data such as age and duration of ailment was collected in absolute figures, not ranges. In order to involve such as variables in logistic regression, binary categories for numerical variables were generated using the median as cut-off. Additionally, some multi-category data such as education level were also further processed into binary categories before regression analysis. Furthermore, urine positive results were categorized into CPD nonmedical and medical use after collaboration with patients’ file data on recently prescribed medications.

Dataset contents: The dataset contains the following data; a) background characteristics of study participants redacted for some details like questionnaire number, name of health facility, variables like nationality, native home region, religion, previous mental disorder, professional background, and specific education level; b) history of clinician diagnosed CPD and illicit drug use disorders; c) current mental health disorder; d) drivers of CPD nonmedical use, including consumption of alcohol and tobacco; e) history of lifetime use of individual CPDs; f) history of treatment against CPD use; g) history of CPD withdrawal effects; i) whether participant provided a urine sample; and j) the urine assay result.

Ethical considerations: The study protocol upon which this data was collected was approved by the Institutional Review Board (IRB) of the School of Health Sciences at Makerere University and the Uganda National Council of Science and Technology. Data was also collected after obtaining informed consent from the next of kin or caregiver of study participants (mental health patients) with the assent of the participants.

Reuse potential: The dataset contains significant amounts of data that has not been utilized in the analyses we conducted for the manuscript that was submitted for publication. These include; amounts of alcohol and tobacco consumption, experiences with CPD withdrawal symptoms, how participants obtain CPDs for nonmedical use, and history of care for CPD drug use disorders, among others. Therefore, there is significant reuse potential in the dataset.

The dataset was generated from three data sources, namely; semi-structured questionnaire; prescribed drug history data extraction from hospital records using a desk review guide; urine drug immunoassay data. The study outcome "nonmedical use" was derived from urine drug assay data by ruling out authorized use from the patients prescribed drug history. Binary predictor variables used in the logistic regression analysis are labeled in the dataset. 

The total dataset has 1275 study participants. However, urine data is only for 988 participants. Also, data on historical controlled drug use disorder is from 1267 participants only as eight files were not accessible. The dataset has been redacted of potential participant identifiers such as name of hospital, questionaire number, age (only categorization is in the dataset), sex, marital status, current mental illness, year of onset of mental illness, past mental illness, and date of recent admission to hospital for inpatients.