Work

Data Resource Profile: The Opioid Agonist Treatment and Safety (OATS) Study, New South Wales, Australia

Year: 2020

Type: article

Abstract: Opioid agonist treatment (OAT) is the most effective treatment for opioid dependence.1,2 The main medications used for OAT are methadone, a full opioid agonist, and buprenorphine, a partial opioid agonist. Risk of death from any cause is halved while on OAT as compared to out-of-treatment opioid dependence; risk of overdose death is reduced by 70%.3 Other harms from opioid use and injecting are also reduced while on OAT.4 HIV and hepatitis C virus (HCV) acquisition are halved5,6 while on treatment. Health service use, particularly unplanned and crisis-driven hospitalizations and emergency department presentations, is significantly reduced.7–9 Criminal offending and associated costs are significantly reduced.10–12 As with any pharmacological treatment, there are risks associated with OAT. Periods of elevated risk of fatal overdose are observed during treatment induction (the first 4 weeks), particularly with methadone, and immediately following treatment cessation.13–16 Additional adverse outcomes during OAT may include non-fatal overdoses, accidents,17 cardiac problems resulting from QT interval prolongation18 or interactions with comorbidity.19 To date however, much of the post-marketing surveillance of methadone and buprenorphine has focused on diversion and illicit use of medications,20–22 with relatively less attention given to adverse outcomes of people in treatment or following treatment cessation. This is a significant knowledge gap that leaves OAT recipients at risk of harm. To address the need for greater knowledge of the frequency and magnitude of risk for adverse outcomes and to identify patients most at risk of adverse outcomes, the Opioid Agonist Treatment and Safety (OATS) Study was established to (i) determine the magnitude of risk for specific adverse outcomes; (ii) identify patient, treatment and provider factors associated with adverse outcomes; and (iii) develop risk prediction models to identify patients at greatest risk of adverse outcomes. The aim of this paper is to provide a profile of the study cohort by summarising available demographic, clinical and legal system indicators. This latest linkage is one in a series of data linkage projects that have been undertaken with OAT clients in New South Wales (NSW) over the past decade. Greater detail on outcomes from earlier linkage is provided in the Data Resource Use section. The OATS Study cohort has been established using linked administrative data from NSW, Australia. In order to prescribe OAT for a patient, a prescriber must submit an Authority to Prescribe form to the Pharmaceutical Regulatory Unit of the NSW Ministry of Health. This form includes the patient’s name and other identifying details (confirmed by the prescriber by viewing one or more forms of identification), basic demographic data (e.g. indigenous status, country of birth), drugs of concern to the patient, and opioid agonist to be prescribed (methadone or buprenorphine). Data from these forms are collated in the Electronic Reporting and Recording of Controlled Drugs (ERRCD). The ERRCD also includes information about the setting in which treatment will take place, including variables to identify the prescribing setting (e.g. primary care, public clinic, private clinic, correctional centre, hospital) and dispensing setting (e.g. public clinic, private clinic, community pharmacy, correctional centre, hospital inpatient). In the OATS study, ERRCD data have been linked to the following state-wide databases: Admitted Patient Data Collection, Emergency Department Data Collection, Mental Health Ambulatory Data Collection, Re-offending Database, Registry of Births, Deaths and Marriages, and the Cause of Death Unit Record File. Using the ERRCD, the OATS cohort is defined as any person prescribed methadone or buprenorphine (including buprenorphine–naloxone) for the treatment of opioid dependence (this being the terminology used in the NSW Opioid Treatment Program) between 1 August 2001 and 19 September 2018 (n = 48 148). This start date was selected to coincide with the introduction of buprenorphine for this indication. For cohort members with OAT episodes prior to 1 August 2001, OAT data prior to this date are also included in the dataset to allow for statistical adjustment for prior treatment. Data linkage will not continue beyond 2018 in this specific study. Missing or erroneous data is often raised as a concern with linked administrative data analyses. The OAT data were recently migrated into the ERRCD from the previous database. During this process, significant cleaning and correction of errors was undertaken, particularly in relation to dates relating to treatment entry and exit. Dates of treatment exit rely on timely reporting by prescribers, which may not always occur. Variables that have rates of missingness that may impact on analyses include main opioid of concern (not reported in 9% of new treatment episodes commenced in 2018) and other drugs of concern (where lack of information may indicate that the prescriber did not complete this section of the form, rather than the patient reporting no other drugs of concern). As mentioned previously, OAT is the most effective treatment for opioid dependence and a World Health Organization (WHO) Essential Medicine for this indication.23 It reduces illicit opioid use and injecting, injecting risk behaviour, risk of HIV and HCV acquisition, increases engagement in the HIV and HCV cascade of care, reduces criminal activity and reduces mortality.24 There is strong evidence that longer-term treatment is required.24,25 To enrol a participant on the NSW Opioid Treatment Program, authorization is required from NSW Ministry of Health. When a participant completes an episode of OAT the prescriber is required to notify the NSW Ministry of Health, along with the reason for exiting the program. Participants moving out of NSW are lost to follow-up. There were 132 402 OAT episodes during follow-up, ranging in duration from 1 day to a maximum of 6363 days, with a median of 209 days (mean of 840 days). The median number of episodes per participant was 2, ranging between 1 and 46 episodes. At the end of the follow-up period there were 21 180 participants in treatment. The OATS data have been probabilistically linked to the following state-wide datasets. Admitted Patient Data Collection: records all admitted patient services provided by public hospitals (including psychiatric hospitals), private hospitals and private day procedures centres. Emergency Department Data Collection: records patient presentations to emergency departments (EDs) of public hospitals. The number of contributing EDs has increased over time such that data for more recent years capture a greater proportion of presentations. Previous analyses of ED use in this population has established that data from 2012 onwards provide reasonably complete capture.26 Mental Health Ambulatory Data Collection: records mental health care contacts with non-admitted patients, including mental health day programs, outreach services, community health service contacts, and outpatient psychiatric contacts. Re-offending Database (including custody data): contains finalized legal actions including Local, District, Supreme, and Children’s Criminal Court actions, as well as adult and juvenile custody data. Registry of Births, Deaths and Marriages/Cause of Death Unit Record File: records fact and cause of death for deaths recorded in NSW. The dates for which data from each dataset are available are shown in Figure 1. Datasets linked in the OATS Study, including date ranges of available data Datasets linked in the OATS Study, including date ranges of available data Linkage was undertaken by the Centre for Health Record Linkage, a dedicated data linkage unit within the Ministry of Health.27 Data were linked probabilistically, based on full name and any recorded aliases, date of birth, sex, and date of last known contact. Ethical approval for this study was provided by the NSW Population and Health Services Research Ethics Committee (2018/HRE0205) and the Aboriginal Health and Medical Research Council Ethics Committee (1400/18). Community involvement in the study has included consultation with a Community Reference Panel of people prescribed OAT. This panel provided feedback on study design and preferred methods for dissemination of results. Additionally, an Aboriginal Advisory Group has been established to review all study outputs to ensure that results relating to Aboriginal peoples are reported in a culturally appropriate manner. Demographic and clinical data on OATS cohort members are presented in Tables 1 and 2 and Figures 2–4. Of the 48 148 cohort members, 68% are male and the median age at cohort entry was 32 years (Table 1). Just over one in five (21%) is of Aboriginal and/or Torres Strait Islander descent. Cohort members are largely located in major cities of NSW (73%) and just under half (43%) live in the most or second-most socio-economically disadvantaged quintiles. A total of 5176 cohort members died during the observation period. Number of the OATS cohort receiving OAT on 1 January annually, 2001–2018, by medicine prescribed Number of the OATS cohort receiving OAT on 1 January annually, 2001–2018, by medicine prescribed Proportion of the OATS cohort reporting heroin or pharmaceutical opioids as a primary concern, 2012–2018 Proportion of the OATS cohort reporting heroin or pharmaceutical opioids as a primary concern, 2012–2018 Other recorded drugs of concern among OATS cohort participants, by year Other recorded drugs of concern among OATS cohort participants, by year Demographic characteristics, deaths and exposure to OAT in the OATS cohort Most disadvantaged Least disadvantaged Most disadvantaged Least disadvantaged Remoteness categorized by applying the Accessibility/Remoteness Index of Australia to postcode at last OAT enrolment. Socio-economic disadvantage categorized by applying the Socio-Economic Indexes for Areas to postcode at last OAT enrolment. Demographic characteristics, deaths and exposure to OAT in the OATS cohort Most disadvantaged Least disadvantaged Most disadvantaged Least disadvantaged Remoteness categorized by applying the Accessibility/Remoteness Index of Australia to postcode at last OAT enrolment. Socio-economic disadvantage categorized by applying the Socio-Economic Indexes for Areas to postcode at last OAT enrolment. OATS Study cohort summary characteristicsa Hospital diagnoses are based on appearance of the specific diagnosis codes appearing in the principal or contributing diagnosis fields. ICD-10 codes for each category are provided in the Appendix, available as Supplementary data at IJE online. Mental health contacts 1 January 2001–31 December 2017, emergency department presentations 1 January 2005–30 June 2018, selected cause-specific hospital admissions 1 July 2001–30 June 2018, and contact with the criminal justice system July 2001–31 December 2017 among the OATS Study cohort. OATS Study cohort summary characteristicsa Hospital diagnoses are based on appearance of the specific diagnosis codes appearing in the principal or contributing diagnosis fields. ICD-10 codes for each category are provided in the Appendix, available as Supplementary data at IJE online. Mental health contacts 1 January 2001–31 December 2017, emergency department presentations 1 January 2005–30 June 2018, selected cause-specific hospital admissions 1 July 2001–30 June 2018, and contact with the criminal justice system July 2001–31 December 2017 among the OATS Study cohort. Many cohort members have long histories of OAT, including one in five having first started OAT prior to 1995. As shown in Figure 2, although the total number of people in OAT has increased over time, that increase is largely due to increasing numbers of people prescribed buprenorphine, as numbers prescribed methadone have remained relatively steady since 2001. Data on patients’ primary opioid of concern have been collected since 2012, and similar proportions of patients report heroin or a pharmaceutical opioid as their primary opioid of concern (Figure 3). Cannabinoids and methamphetamine are the non-opioid drugs most frequently nominated by OAT entrants as being additional drugs of concern (Figure 4). Care should be taken in interpreting the findings in Figure 4, as there is likely variation between prescribers in how these data are collected (e.g. asking patients about specific drugs of concern vs asking patients if they have any other drugs of concern). Around half of the participants had at least one contact with an ambulatory mental health service (52%; Table 2). Most participants have at least one ED presentation (84%) and hospitalization (82%) during the observation period. Diagnoses are coded using International Classification of Diseases (ICD) 10 codes; the coding scheme used is provided in the Appendix, available as Supplementary data at IJE online. Nearly two-thirds have a hospitalization where mental and behavioural disorders due to use of opioids is included in the diagnostic code fields, making up more than one-quarter of all hospitalizations in the cohort. Hospitalizations with diagnostic codes for accidental injuries and liver disease are also widespread, with 41 and 38% of cohort members having at least one of these. Two-thirds of the cohort had at least one criminal conviction during the observation period (Table 2). There were >500 000 convictions in total, of which 22% were property crimes. Two in every five of the cohort were incarcerated at least once. Participants spent 45 392 person-years, or ∼11.5% of their time, in custody over the observation period. In work published from this latest cohort linkage, we have found that OAT was associated with reductions in ED presentations and associated costs among people with opioid dependence. Low acuity ED presentations were reduced in OAT among people residing in major cities of NSW, however low acuity presentations increased in regional and remote NSW, suggesting poorer access to primary care.26 As noted earlier, this current project is one in a series of linkage projects using the cohort of OAT clients in NSW. Previous publications based on the OAT cohort have found that opioid-dependent people who enter OAT have high rates of contact with the criminal justice system, but most contacts are made with a small minority of this cohort,28 and that there has been substantial variation in causes of death over time,29 with some causes of death becoming increasingly important such as cancer.30 Further, we have found that time spent in treatment is associated with around a halving of mortality risk;31 that in high-risk periods for mortality, such as the immediate period after release from prison, being in OAT is associated with a 74% reduction in mortality risk;32 and that time spent in OAT is associated with reductions in criminal offending, especially if people are retained in treatment (rather than cycling in and out of treatment).33 For further publications please see: https://ndarc.med.unsw.edu.au/project/determining-impact-opioid-substitution-therapy-upon-mortality-and-recidivism-among-prisoners,https://ndarc.med.unsw.edu.au/project/opioid-agonist-treatment-and-safety-oats-study and https://ndarc.med.unsw.edu.au/search/our-projects/. A key strength of this cohort is its inclusion of all OAT prescribed state-wide, in any setting, including primary care, specialist clinics and prisons. Dispensing setting is also available to allow for comparisons of adverse outcomes in specific treatment contexts, such as methadone prescribing in primary care with pharmacy dispensing compared to methadone prescribing and dispensing in a specialized clinic setting. The OAT cohort data do not, however, include event-level dose data, nor is it possible to distinguish between mono-buprenorphine and buprenorphine–naloxone. Additionally, it is not possible to determine when people have access to unsupervised (take-home) dosing. It should not be assumed that buprenorphine prescriptions are for unsupervised dosing as is routinely the case in the USA. NSW Opioid Treatment Program clinical guidelines recommend that all methadone and buprenorphine doses be supervised during treatment induction and stabilization.34 After this, unsupervised dosing may be considered, with the number and timing of take-home doses determined based on risk assessment (e.g. a patient considered ‘moderate risk’ may receive up to two take-home doses of methadone per week, on consecutive or non-consecutive days).34 The cohort is formed using linked administrative health and criminal justice data only. As such, data are limited to those variables that are collected for administrative purposes. We do not have data to confirm e.g. HIV or hepatitis C infection; we can only infer these from hospitalization data, which almost certainly underestimates true prevalence. Similarly, we do not have data on indicators of social determinants of health, such as homelessness. Socio-economic status can only be inferred indirectly, by applying the Socio-Economic Indexes for Areas to the postcode listed for each participant.35 The linkage does not include data on visits to primary health care providers other than OAT prescribing by general practitioners, nor data on medicines prescribed other than OAT. To address missing data and data entry errors across datasets, we have undertaken substantial data cleaning efforts including logic checks, correction of obvious typographical errors, and where variables are recorded across datasets, comparison of these to determine ‘correct’ values. For example, indigenous status is known to be poorly recorded in many health datasets,36 we have elected to code an individual as Aboriginal and/or Torres Strait Islander if they are identified as such in any of the included datasets. Despite these limitations, the size and duration of the cohort will permit analyses of rare adverse outcomes and changes over time. We have linked multiple health and criminal justice datasets that provide a rich set of covariates for inclusion in predictive models. These data will enable detailed exploration of OAT trajectories that account for individual and treatment setting variables and how these may influence outcomes. We expect findings from these analyses to inform OAT clinical guidelines and the expansion of diverse treatment models that increase access to OAT. To protect privacy and confidentiality, approval for the linkage of health data in NSW is provided under strict conditions for the storage, retention and use of the data. The current approval permits storage of the data at one site, University of New South Wales Sydney, for up to 7 years following the date of publication of results. Data may only be supplied for analysis within Australia. We encourage interested parties to contact us to discuss potential secondary data analyses, noting that legislation requires that the data can only be stored and analysed within NSW. Virtual Private Network and other virtual access is not permitted under current approvals. Requests for data access can be submitted to Professor Louisa Degenhardt ([email protected]) and will be reviewed by the OATS investigator team. Potential collaborators will be required to gain approval for data access and specific secondary analyses from the NSW Population and Health Services Research Ethics Committee. Collaborators proposing to examine research questions relating specifically to Aboriginal peoples will also be required to seek approval from the Aboriginal Health and Medical Research Council. Supplementary data are available at IJE online. This work was supported by the National Institutes of Health [R01 DA144740 to L.D.]. S.L., S.N. and L.D. are supported by Australian National Health and Medical Research Council Research Fellowships [1140938, 1163961, 1135991]. The National Drug and Alcohol Research Centre is supported by funding from the Australian Government Department of Health under the Drug and Alcohol Program. We acknowledge the NSW Ministry of Health, Centre for Health Record Linkage, and Bureau of Crime Statistics and Research for their support in data provision and linkage. We also acknowledge the support and expertise of Doug James and Craig Vaughan from the OATS Study Aboriginal Advisory Group in reviewing this paper for publication. S.L., L.D., M.H., S.N., D.A.F., R.A., and T.D. developed the study design and and data analyses. the reviewed the and provided This study is by the NSW Population and Health Services Research Ethics Committee (2018/HRE0205) and Aboriginal Health and Medical Research Council Research Ethics Committee (1400/18). A of was under the NSW has funding from L.D. has funding from and S.N. has from and has from and has funding from and from of from of from from University of from the submitted

Source: International Journal of Epidemiology

DOI: 10.1093/ije/dyaa125