2.1 Data source
All data were register data and were drawn from the Norwegian Prescription Database (NorPD). From January 1, 2004, all pharmacies are obliged to submit data for all dispensed drugs electronically to NorPD underlying the Norwegian Institute of Public Health (www.norpd.no). The NorPD contains information on all drugs dispensed from pharmacies, except for drugs administrated at hospitals, nursing homes, and outpatient clinics [23]. The Anatomical Therapeutic Chemical (ATC) classification system was employed in accordance with the World Health Organization (WHO) standards per October 2018 [24].
2.2 Study population
All patients above 18 years of age who received at least one dispensation of OAT opioids per calendar year, including methadone, levomethadone, buprenorphine, and buprenorphine-naloxone from January 1, 2015, to December 31, 2017, were included. In addition, some patients in palliative care use methadone tablets to achieve pain relief. These patients were excluded by identifying those who only were dispensed methadone tablets without any dispensations of other OAT opioids or methadone mixture in the period from January 1, 2004, to December 31, 2017.
2.3 Analysis strategy and statistical analyses
2.3.1 Definitions of drugs, including ADHD medications and opioid agonist therapy opioids, the number of dispensations of OAT opioids, and diagnoses.
ADHD medications were defined as all CAS that had marketing authorizations in Norway in the period 2015 to 2017, including racemic amphetamine, dexamphetamine, methylphenidate, and lisdexamphetamine. In addition, we included the non-stimulant atomoxetine, which is also authorized and recommended in the treatment of ADHD according to guidelines and reviews [6-10, 18, 25]. All included OAT opioids, ADHD medications, non-OAT opioids, benzodiazepines, z-hypnotics, gabapentinoids, including gabapentin and pregabalin, and DAAs were categorized according to their ATC codes (Additional file 1). The type of OAT opioid that patients were dispended was defined as the last type of OAT opioid that was dispensed per calendar year.
The number of dispensed OAT opioids was defined as the number of dispensations of any OAT opioid per patient per calendar year. The number of dispensations was stratified according to four categories: 1-6, 7-12, 13-51, and ≥ 52 dispensations per calendar year. Age was defined according to the patient’s age in the calendar year and categorized into five groups: ≤ 25, 26-35, 36-45, 46-55, and ≥ 56 years.
All reimbursable and non-reimbursable ADHD medication dispensations were included. The prescribing physician needs to specify the medical condition that is treated by the particular drug, using codes from the 10th revision of International Classification of Diseases (ICD-10) or International Classification of Primary Care 2 (ICPC-2) to get public drug expenses reimbursed in Norway. The diagnostic codes of reimbursed drugs are recorded in the NorPD. Only two medical indications are approved for ADHD medication expense reimbursements in Norway: Hyperkinetic disorder/ADHD (ICD-10: F90 and ICPC-2: P81) or narcolepsy (ICD-10: G47 and ICPC-2: P81). For narcolepsy, only CAS are reimbursed. The information on diagnostic codes for non-reimbursable dispensations are not collected in the NorPD.
2.3.2 Analysis strategy according to the aims
One-year’s dispensation rates of ADHD medication during the study period were assessed by summing all patients who received at least one dispensation of an ADHD medication per the calendar year. Furthermore, patients were divided into two groups “all medical indications” and “ADHD” for the years in the study period. The group named “ADHD” only included patients who were dispensed ADHD medications with reimbursement codes for ADHD. The group named “all medical indications” included all patients who received dispensations of ADHD medications, either they were reimbursed or not. Less than five patients were dispensed CAS on the reimbursement code for narcolepsy in the study period.
The association with being dispensed ADHD medication, or not adjusted for age, gender, type of dispended OAT opioids, the number of dispensed OAT opioids, being dispensed benzodiazepines, z-hypnotics, gabapentinoids, or non-OAT opioids were calculated per calendar year in the study period. Age and the number of dispensed OAT opioids were categorized according to the predefined categories or groups per year. All dispensed ADHD medication, and potentially addictive drugs were identified and categorized into four drug groups: “benzodiazepines or z-hypnotics,” “gabapentinoids,” “non-OAT opioids,” and “ADHD medication” per year. For each group, categorical variables were created by whether patients were dispensed one or more of the drugs in the drug groups or not. Dispensations of DAA were also included due to the frequent use of illicit stimulant drugs in the OAT population and the fact that DAA against hepatitis C infection has made treatment more applicable for these patients. Patients were defined to be dispensed treatment with DAA if they had at least one dispensation of DAA from 2011 and until the end of 2015, 2016, or 2017, respectively.
The mean daily dose of each ADHD medication and the dispensation rates of benzodiazepines or z-hypnotics, gabapentinoids, and non-OAT opioids in 2017, were calculated among patients with at least one dispensation of ADHD medication and OAT, respectively, for each calendar year in the study period. These patients were assumed to have achieved medical continuity in their ADHD treatment and follow up treatment according to national guidelines. The mean daily doses of ADHD medication were calculated by summarizing the total volume of defined daily doses (DDD) of each drug that was dispensed for each patient per year [26]. Further, the number of DDDs dispensed per patient was converted to milligrams according to WHO’s standards (Additional file 2). The mean daily dose for each ADHD medication was calculated by dividing the total dose (in milligram) of each drug per year by 365.25 days. The drug groups of each potentially addictive drugs were used to calculate dispensation rates. Each drug group was defined as categorical variables according to whether patients were dispensed at least one drug defined into the drug group or not during 2017.
2.3.3 Statistical analyses
Means, medians, percentiles, percentages, 95 % confidence intervals (CI), odds ratios (OR), and p-values are presented when appropriate. Multivariable analyses for categorical variables were performed by binary logistic regression. Being dispensed ADHD medication, as well as OAT at least once, respectively, during a calendar year, were defined as a dependent variable in the logistic regression model. Independent variables were age, gender, ‘the number of dispensations of OAT opioids,’ ‘benzodiazepines or z-hypnotics,’ ‘gabapentinoids,’ ‘non-OAT opioids,’ and ‘DAA.’ All these variables were defined categorically. The level of statistical significance was defined as p < 0.05. The Chi-square test and paired sample t-test were used to estimate differences in dispensed mean daily doses of ADHD medication in 2015 compared to 2017 among patients with at least one dispensation of ADHD medication and OAT, respectively, during a calendar year throughout the study period. All patients were censored from the calendar year they died. SPSS version 24 was used for all analyses.
2.4. Ethical considerations
The Regional Committee for Medical and Health Research Ethics, REC vest, Norway, has approved the use of registry data for this study (approval number 2018/939/REK Vest, June 19, 2018). No informed consent from included patients was necessary. The STROBE checklist was applied in the preparation of the study (Additional file 3).