Using routinely collected data to understand and predict adverse outcomes in opioid agonist treatment: Protocol for the Opioid Agonist Treatment Safety (OATS) Study

Effect of incarceration and opioid agonist treatment transitions on risk of hospitalisation with injection drug use-associated bacterial infections: A self-controlled case series in New South Wales, Australia

BACKGROUND

Transitional times in opioid use, such as release from prison and discontinuation of opioid agonist treatment (OAT), are associated with health harms due to changing drug consumption practices and limited access to health and social supports. Using a self-controlled (within-person) study design, we aimed to understand if these transitions increase risks of injection drug use-associated bacterial infections.

METHODS

We performed a self-controlled case series among a cohort of people with opioid use disorder (who had all previously accessed OAT) in New South Wales, Australia, 2001-2018. The outcome was hospitalisation with injecting-related bacterial infections. We divided participants' observed days into time windows related to incarceration and OAT receipt. We compared hospitalization rates during focal (exposure) windows and referent (control) windows (i.e., 5-52 weeks continuously not incarcerated or continuously receiving OAT). We estimated adjusted incidence rate ratios (aIRR) using conditional logistic regression, adjusted for time-varying confounders.

RESULTS

There were 7590 participants who experienced hospitalisation with injecting-related bacterial infections (35% female; median age 38 years; 78% hospitalised with skin and soft-tissue infections). Risk for injecting-related bacterial infections was elevated for two weeks following release from prison (aIRR 1.45; 95%CI 1.22-1.72). Risk was increased during two weeks before (aIRR 1.89; 95%CI 1.59-2.25) and after (aIRR 1.91; 95%CI 1.54-2.36) discontinuation of OAT, and during two weeks before (aIRR 3.63; 95%CI 3.13-4.22) and after (aIRR 2.52; 95%CI 2.09-3.04) OAT initiation.

CONCLUSION

Risk of injecting-related bacterial infections varies greatly within-individuals over time. Risk is raised immediately after prison release, and around initiation and discontinuation of OAT. Social contextual factors likely contribute to excess risks at transitions in incarceration and OAT exposure.

Prevalence of opioid dependence in New South Wales, Australia, 2014-16: Indirect estimation from multiple data sources using a Bayesian approach

AIMS

To estimate the prevalence of, and number of unobserved people with opioid dependence by sex and age group in New South Wales (NSW), Australia.

DESIGN

We applied a Bayesian statistical modelling approach to opioid agonist treatment records linked to adverse event rate data. We estimated prevalence from three types of adverse event separately: opioid mortality, opioid-poisoning hospitalizations and opioid-related charges. We extended the model and produced prevalence estimates from a 'multi-source' model based on all three types of adverse event data.

SETTING, PARTICIPANTS AND MEASUREMENTS

This study was conducted in NSW, Australia, 2014-16 using data from the Opioid Agonist Treatment and Safety (OATS) study, which included all people who had received treatment for opioid dependence in NSW. Aggregate data were obtained on numbers of adverse events in NSW. Rates of each adverse event type within the OATS cohort were modelled. Population data were provided by State and Commonwealth agencies.

FINDINGS

Prevalence of opioid dependence among those aged 15-64 years in 2016 was estimated to be 0.96% (95% credible interval [CrI] = 0.82%, 1.12%) from the mortality model, 0.75% (95% CrI = 0.70%, 0.83%) from hospitalizations, 0.95% (95% CrI = 0.90%, 0.99%) from charges and 0.92% (95% CrI = 0.88%, 0.96%) from the multi-source model. Of the estimated 46 460 (95% CrI = 44 680, 48 410) people with opioid dependence in 2016 from the multi-source model, approximately one-third (16 750, 95% CrI = 14 960, 18 690) had no record of opioid agonist treatment within the last 4 years. From the multi-source model, prevalence in 2016 was estimated to be 1.24% (95% CrI = 1.18%, 1.31%) in men aged 15-44, 1.22% (95% CrI = 1.14%, 1.31%) in men 45-64, 0.63% (95% CrI = 0.59%, 0.68%) in women aged 15-44 and 0.56% (95% CrI = 0.50%, 0.63%) in women aged 45-64.

CONCLUSIONS

A Bayesian statistical approach to estimate prevalence from multiple adverse event types simultaneously calculates that the estimated prevalence of opioid dependence in NSW, Australia in 2016 was 0.92%, higher than previous estimates.

Does opioid agonist treatment reduce overdose mortality risk in people who are older or have physical comorbidities? Cohort study using linked administrative health data in New South Wales, Australia, 2002-17

AIMS

To quantify the association between opioid agonist treatment (OAT) and overdose death by age group; test the hypothesis that across different age groups, opioid overdose mortality is lowest during OAT with buprenorphine compared with time out of treatment or OAT with methadone; and test associations between OAT and opioid overdose mortality in the presence of chronic circulatory, respiratory, liver and kidney diseases.

DESIGN

Retrospective observational cohort study using linked administrative data.

SETTING

New South Wales, Australia.

PARTICIPANTS

A total of 37 764 people prescribed OAT, 1 August 2002 and 31 December 2017.

MEASUREMENTS

OAT exposure, opioid overdose mortality and key confounders were measured using linked population data sets on OAT entry and exit, hospitalization, mental health care, incarceration and mortality. ICD-10 codes were used to define opioid overdose mortality and chronic disease groups of interest.

FINDINGS

Relative to time out of treatment, time in OAT was associated with a lower risk of opioid overdose death across all age groups and chronic diseases. Among people aged 50 years and older, there was weak evidence that buprenorphine may be associated with greater protection against opioid overdose death than methadone [generalized estimating equation (GEE) adjusted incident rate ratio (aIRR) = 0.47; 95% confidence interval (CI) = 0.21, 1.02; marginal structural models (MSM) aIRR = 0.49; 95% CI = 0.17, 1.41]. Buprenorphine was associated with greater protection against overdose death than methadone for clients with circulatory (MSM aIRR = 0.27; 95% CI = 0.11, 0.67) or respiratory (MSM aIRR = 0.26; 95% CI = 0.07, 0.94) diseases, but not liver (MSM aIRR = 0.59; 95% CI = 0.14, 2.43) or kidney (MSM aIRR = 1.16; 95% CI = 0.31, 4.36) diseases.

CONCLUSIONS

Opioid agonist treatment (OAT) appears to reduce mortality risk in people with opioid use disorder who are older or who have physical comorbidities. Opioid overdose mortality during OAT with buprenorphine appears to be lower and reduced in clients with circulatory and respiratory diseases compared with OAT with methadone.

Incidence of suicide and self-harm among people with opioid use disorder and the impact of opioid agonist treatment: A retrospective data linkage study

BACKGROUND

Rates of suicide and self-harm are elevated among people with opioid use disorder (OUD). This study examined incidence of self-harm and suicide among people who have entered OAT and assessed the impact of different OAT exposure periods on these events.

METHOD

We conducted a retrospective population-based cohort study of all OAT recipients (N = 45,664) in New South Wales, Australia (2002-2017), using linked administrative data. Incidence rates of self-harm hospitalisations and suicide deaths were estimated per 1000 person-years (PY). The first 28 days of an OAT episode, ≥ 29 days on OAT, the first 28 days off OAT, and ≥ 29 days off OAT (maximum four years post-OAT) were exposure periods. Poisson regression models with generalised estimating equations estimated the adjusted incidence rate ratios (ARR) of self-harm and suicide by OAT exposure periods, adjusting for covariates.

RESULTS

There were 7482 hospitalisations (4148 individuals) for self-harm and 556 suicides, equating to incidence rates of 19.2 (95% confidence intervals [CI]=18.8-19.7) and 1.0 (95%CI=0.9-1.1) per 1000 PY, respectively. Opioid overdose was implicated in 9.6% of suicides and 28% of self-harm hospitalisations. Compared to ≥ 29 days on OAT, the incidence rate of suicide was elevated in the 28 days following OAT cessation (ARR=17.4 [95%CI=11.7-25.9]), and the rate of self-harm hospitalisations was elevated during the first 28 days of OAT (ARR=2.2 [95%CI=1.9-2.6]) and the 28 days after leaving OAT (ARR=2.7 [95%CI=2.3-3.2]).

CONCLUSIONS

OAT may reduce suicide and self-harm risk among people with OUD; however, OAT initiation and cessation are critical periods for targeting self-harm and suicide prevention interventions.

The impact of opioid agonist treatment on fatal and non-fatal drug overdose among people with a history of opioid dependence in NSW, Australia, 2001-2018: Findings from the OATS retrospective linkage study

BACKGROUND

There are critical periods of mortality risk at onset and cessation of opioid agonist treatment. We aim to determine whether non-fatal overdose followed the same pattern as fatal overdose, comparing the first 4 weeks of treatment and treatment cessation and the remainder time off treatment, with the remainder treatment time, to determine intervention markers.

METHODS

Retrospective cohort study of people with a history of opioid agonist treatment using linked New South Wales data. The incidence of non-fatal overdose hospitalization; emergency department presentation; and fatal overdose from national death records were compared. Rates were calculated using generalized estimating equations adjusting for demographics, year, and recent health and incarceration events.

RESULTS

The remainder time in OAT had the lowest incidence of overdose for all outcomes and is the reference level for the adjusted incident rate ratios (aIRR). Fatal overdose was lowest in treatment and highest in the first four weeks out of treatment, aIRR of 12.83 (95% CI 10.0-16.4). Whereas the highest overdose rate for non-fatal opioid overdose was in the first four weeks in treatment, aIRR of 3.11 (95% CI 2.19-4.42).

CONCLUSIONS

Retention on opioid agonist treatment is protective against drug related overdose. There is elevated risk of non-fatal overdose at treatment initiation that is not evident for fatal overdose, but the first month of treatment cessation is a critical period for both non-fatal and fatal overdose. These findings emphasize the importance of treatment retention and interventions for polysubstance overdose at cessation.

Opioid agonist treatment and risk of death or rehospitalization following injection drug use-associated bacterial and fungal infections: A cohort study in New South Wales, Australia

BACKGROUND

Injecting-related bacterial and fungal infections are associated with significant morbidity and mortality among people who inject drugs (PWID), and they are increasing in incidence. Following hospitalization with an injecting-related infection, use of opioid agonist treatment (OAT; methadone or buprenorphine) may be associated with reduced risk of death or rehospitalization with an injecting-related infection.

METHODS AND FINDINGS

Data came from the Opioid Agonist Treatment Safety (OATS) study, an administrative linkage cohort including all people in New South Wales, Australia, who accessed OAT between July 1, 2001 and June 28, 2018. Included participants survived a hospitalization with injecting-related infections (i.e., skin and soft-tissue infection, sepsis/bacteremia, endocarditis, osteomyelitis, septic arthritis, or epidural/brain abscess). Outcomes were all-cause death and rehospitalization for injecting-related infections. OAT exposure was classified as time varying by days on or off treatment, following hospital discharge. We used separate Cox proportional hazards models to assess associations between each outcome and OAT exposure. The study included 8,943 participants (mean age 39 years, standard deviation [SD] 11 years; 34% women). The most common infections during participants' index hospitalizations were skin and soft tissue (7,021; 79%), sepsis/bacteremia (1,207; 14%), and endocarditis (431; 5%). During median 6.56 years follow-up, 1,481 (17%) participants died; use of OAT was associated with lower hazard of death (adjusted hazard ratio [aHR] 0.63, 95% confidence interval [CI] 0.57 to 0.70). During median 3.41 years follow-up, 3,653 (41%) were rehospitalized for injecting-related infections; use of OAT was associated with lower hazard of these rehospitalizations (aHR 0.89, 95% CI 0.84 to 0.96). Study limitations include the use of routinely collected administrative data, which lacks information on other risk factors for injecting-related infections including injecting practices, injection stimulant use, housing status, and access to harm reduction services (e.g., needle exchange and supervised injecting sites); we also lacked information on OAT medication dosages.

CONCLUSIONS

Following hospitalizations with injection drug use-associated bacterial and fungal infections, use of OAT is associated with lower risks of death and recurrent injecting-related infections among people with opioid use disorder.

The impact of opioid agonist treatment on hospitalisations for injecting-related diseases among an opioid dependent population: A retrospective data linkage study

BACKGROUND

Injecting-related bacterial and fungal infections cause substantial illness and disability among people who use illicit drugs. Opioid agonist treatment (OAT) reduces injecting frequency and the transmission of blood borne viruses. We estimated the impact of OAT on hospitalisations for non-viral infections and examine trends in incidence over time.

METHODS

We conducted a retrospective cohort study using linked administrative data. The cohort included 47 163 individuals starting OAT between August 2001 and December 2017 in New South Wales, Australia, with 454 951 person-years of follow-up. The primary outcome was hospitalisation for an injecting-related disease. The primary exposure was OAT status (out of OAT, first four weeks of OAT, and OAT retention [i.e., more than four weeks in treatment]). Covariates included demographic characteristics, year of hospitalisation, and recent clinical treatment.

RESULTS

9122 participants (19.3%) had at least one hospitalisation for any injecting-related disease. Compared to time out of treatment, retention on OAT was associated with a reduced rate of injecting-related diseases (adjusted rate ratio[ARR]=0.92; 95%CI 0.87-0.97). The first four weeks of treatment was associated with an increased rate (ARR 1.53, 95%CI 1.38-1.70), which we believe is explained by referral pathways between hospital and community OAT services. The age-adjusted incidence rates of hospitalisations for any injecting-related disease increased from 34.8 (95% CI =30.2-40.0) per 1000 person-years in 2001 to 54.9 (95%CI=51.3-58.8) in 2017.

INTERPRETATION

Stable OAT is associated with reduced hospitalisations for injecting-related bacterial infections; however, OAT appears insufficient to prevent these harms as the rate of these infections is increasing in Australia.

Modeling the population-level impact of opioid agonist treatment on mortality among people accessing treatment between 2001 and 2020 in New South Wales, Australia

BACKGROUND AND AIMS

The individual-level effectiveness of opioid agonist treatment (OAT) in reducing mortality is well established, but there is less evidence on population-level benefits. We use modeling informed with linked data from the OAT program in New South Wales (NSW), Australia, to estimate the impact of OAT provision in the community and prisons on mortality and the impact of eliminating excess mortality during OAT initiation/discontinuation.

DESIGN

Dynamic modeling.

SETTING AND PARTICIPANTS

A cohort of 49 359 individuals who ever received OAT in NSW from 2001 to 2018.

MEASUREMENTS

Receipt of OAT was represented through five stages: (i) first month on OAT, (ii) short (1-9 months) and (iii) longer (9+ months) duration on OAT, (iv) first month following OAT discontinuation and (v) rest of time following OAT discontinuation. Incarceration was represented as four strata: (i) never or not incarcerated in the past year, (ii) currently incarcerated, (iii) released from prison within the past month and (iv) released from prison 1-12 months ago. The model incorporated elevated mortality post-release from prison and OAT impact on reducing mortality and incarceration.

FINDINGS

Among the cohort, mortality was 0.9 per 100 person-years, OAT coverage and retention remained high (> 50%, 1.74 years/episode). During 2001-20, we estimate that OAT provision reduced overdose and other cause mortality among the cohort by 52.8% [95% credible interval (CrI) = 49.4-56.9%] and 26.6% (95% CrI =22.1-30.5%), respectively. We estimate 1.2 deaths averted and 9.7 life-years gained per 100 person-years on OAT. Prison OAT with post-release OAT-linkage accounted for 12.4% (95% CrI = 11.5-13.5%) of all deaths averted by the OAT program, primarily through preventing deaths in the first month post-release. Preventing elevated mortality during OAT initiation and discontinuation could have averted up to 1.4% (95% CrI =  0.8-2.0%) and 3.0% (95% CrI = 2.1-5.3%) of deaths, respectively.

CONCLUSION

The community and prison opioid agonist treatment program in New South Wales, Australia appears to have substantially reduced population-level overdose and all-cause mortality in the past 20 years, partially due to high retention.

Using administrative data to predict cessation risk and identify novel predictors among new entrants to opioid agonist treatment

BACKGROUND

Longer retention in opioid agonist treatment (OAT) is associated with improved treatment outcomes but 12-month retention rates are often low. Innovative approaches are needed to strengthen retention in OAT. We develop and compare traditional and deep learning-extensions of Cox regression to examine the potential for predicting time in OAT at individuals' first episode entry.

METHODS

Retrospective cohort study in New South Wales, Australia including 16,576 people entering OAT for the first time between January 2006 and December 2017. We develop 12-month OAT cessation prediction models using traditional and deep learning-extensions of the Cox regression algorithm with predictors evaluated from linked administrative datasets. Proportion of explained variation, calibration, and discrimination are compared using 5 × 2 cross-validation.

RESULTS

Twelve-month cessation rate was 58.4%. The largest hazard ratios for earlier cessation from the deep learning model were observed for treatment factors, including private dosing points (HR=1.54, 95% CI=1.49-1.60) and buprenorphine medication (HR=1.43, 95% CI=1.39-1.46). Diagnostic codes for homelessness (HR=1.09, 95% CI=1.04-1.13), outpatient treatment for drug use disorders (HR=1.10, 95% CI=1.06-1.15), and occupant of vehicle accident (HR=1.04, 95% CI=1.01-1.07) from past-year health service presentations were identified as significant predictors of retention. We observed no improvement in performance of the deep learning model over traditional Cox regression.

CONCLUSIONS

Deep learning may be more useful in identifying novel risk factors of OAT retention from administrative data than evaluating individual-level risk. An increased focus on addressing structural issues at the population level and considering alternate models of care may be more effective at improving retention than delivering fully personalised OAT.

The effect of person, treatment and prescriber characteristics on retention in opioid agonist treatment: a 15-year retrospective cohort study

BACKGROUND AND AIMS

There is limited evidence on the relationship between retention in opioid agonist treatment for opioid dependence and characteristics of treatment prescribers. This study estimated retention in buprenorphine and methadone treatment and its relationship with person, treatment and prescriber characteristics.

DESIGN

Retrospective longitudinal study.

SETTING

New South Wales, Australia.

PARTICIPANTS

People entering the opioid agonist treatment programme for the first time between August 2001 and December 2015.

MEASUREMENTS

Time in opioid agonist treatment (primary outcome) was modelled using a generalized estimating equation model to estimate associations with person, treatment and prescriber characteristics.

FINDINGS

The impact of medication type on opioid agonist treatment retention reduced over time; the risk of leaving treatment when on buprenorphine compared with methadone was higher among those who entered treatment earlier [e.g. 2001-03: odds ratio (OR) = 1.59, 95% confidence interval (CI) = 1.45-1.75] and lowest among those who entered most recently (2013-15: OR = 1.23, 95% CI = 1.11-1.36). In adjusted analyses, risk of leaving was reduced among people whose prescriber had longer tenure of prescribing (e.g. 3 versus 8 years: OR = 0.94, 95% CI = 0.93-0.95) compared with prescribers with shorter tenure. Aboriginal and Torres Strait Islander people, being of younger age, past-year psychosis disorder and having been convicted of more criminal charges in the year prior to treatment entry were associated with increased risk of leaving treatment.

CONCLUSION

In New South Wales, Australia, retention in buprenorphine treatment for opioid dependence, compared with methadone, has improved over time since its introduction in 2001. Opioid agonist treatment retention is affected not only by characteristics of the person and his or her treatment, but also of the prescriber, with those of longer prescribing tenure associated with increased retention of people in opioid agonist treatment.

Big data and predictive modelling for the opioid crisis: existing research and future potential

A need exists to accurately estimate overdose risk and improve understanding of how to deliver treatments and interventions in people with opioid use disorder in a way that reduces such risk. We consider opportunities for predictive analytics and routinely collected administrative data to evaluate how overdose could be reduced among people with opioid use disorder. Specifically, we summarise global trends in opioid use and overdoses; describe the use of big data in research into opioid overdose; consider the potential for predictive modelling, including machine learning, for prevention and monitoring of opioid overdoses; and outline the challenges and risks relating to the use of big data and machine learning in reducing harms that are related to opioid use. Future research for improving the coverage and provision of existing interventions, treatments, and resources for opioid use disorder requires collaboration of multiple agencies. Predictive modelling could transport the concept of stratified medicine to public health through novel methods, such as predictive modelling and emulated trials for evaluating diagnoses and prognoses of opioid use disorder, predicting treatment response, and providing targeted treatment recommendations.

Scotland's 2009-2015 methadone-prescription cohort: Quintiles for daily dose of prescribed methadone and risk of methadone-specific death

AIMS

As methadone clients age, their drug-related death (DRD) risks increase, more than doubling at 45+ years for methadone-specific DRDs.

METHODS

Using Community Health Index (CHI) numbers, mortality to 31 December 2015 was ascertained for 36 347 methadone-prescription clients in Scotland during 2009-2015. Cohort entry, quantity of prescribed methadone and daily dose (actual or recovered by effective, simple rules) were defined by clients' first CHI-identified methadone prescription after 30 June 2009 and used in proportional hazards analysis. As custodian of death records, National Records of Scotland identified non-DRDs from DRDs. Methadone-specific DRD means methadone was implicated but neither heroin nor buprenorphine.

RESULTS

The cohort's 192 928 person-years included 1857 non-DRDs and 1323 DRDs (42%), 546 of which were methadone specific. Actual/recovered daily dose was available for 26 533 (73%) clients who experienced 420 methadone-specific DRDs. Top quintile for daily dose at first CHI-identified methadone prescription was >90 mg. Age 45+ years at cohort-entry (hazard ratio vs 25-34 years: 3.1, 95% confidence interval: 2.4-4.2), top quintile for baseline daily dose of prescribed methadone (vs 50-70 mg: 1.9, 1.1-3.1) and being female (1.3, 1.0-1.6) significantly increased clients' risk of methadone-specific DRD.

CONCLUSION

Extra care is needed when methadone daily dose exceeds 90 mg. Females' higher risk for methadone-specific DRD is new and needs validation. Further analyses of prescribed daily dose linked to mortality for large cohorts of methadone clients are needed internationally, together with greater pharmacodynamic and pharmacokinetic understanding of methadone by age and sex. Balancing age-related risks is challenging for prescribers who manage chronic opiate dependency against additional uncertainty about the nature, strength and pharmacological characteristics of drugs from illegal markets.

Retention of opioid agonist treatment prescribers across New South Wales, Australia, 2001-2018: Implications for treatment systems and potential impact on client outcomes

BACKGROUND

There has been much research on the efficacy and effectiveness of opioid agonist treatment (OAT), but less on its implementation and sustainability. A challenge internationally has been recruiting and retaining prescribers. This paper aims to characterise the prescribers in terms of OAT prescribing behaviours.

METHODS

Retrospective cohort study in New South Wales, Australia. Participants were 2199 OAT prescribers between 1 st August 2001-19th September 2018.We examined trends in initiation and cessation of OAT prescribers. Adjusted hazard ratios were calculated to estimate prescriber retention, adjusting for year of initiation, practice type, client load and treatment prescribed.

RESULTS

The rate of prescribers ceasing OAT prescribing has been increasing over time: a prescriber who initiated between 2016-2017 had over four times the risk of cessation compared with one who initiated before 2001, AHR: 4.77; [3.67-6.21]. The highest prescriber cessation rate was in prescribers who had prescribed for shorter time periods. The annual percentage of prescribers who ceased prescribing among those who prescribed for ≤5 years increased from 3% in 2001 to 20 % in 2017. By 2017 more prescribers were discontinuing prescribing than new prescribers were starting. Approximately 87 % (n = 25,167) of OAT clients were under the care of 20 % of OAT prescribers (n = 202); half had been prescribing OAT for 17+ years.

CONCLUSIONS

OAT prescribing is increasingly concentrated in a small group of mature prescribers, and new prescribers are not being retained. There is a need to identify and respond to the reasons that contribute to newer prescribers to cease prescribing and put in place strategies to increase retention and broaden the base of doctors involved in such prescribing.

Intersections of machine learning and epidemiological methods for health services research

The field of health services research is broad and seeks to answer questions about the health care system. It is inherently interdisciplinary, and epidemiologists have made crucial contributions. Parametric regression techniques remain standard practice in health services research with machine learning techniques currently having low penetrance in comparison. However, studies in several prominent areas, including health care spending, outcomes and quality, have begun deploying machine learning tools for these applications. Nevertheless, major advances in epidemiological methods are also as yet underleveraged in health services research. This article summarizes the current state of machine learning in key areas of health services research, and discusses important future directions at the intersection of machine learning and epidemiological methods for health services research.

Hospitalisations for non-fatal overdose among people with a history of opioid dependence in New South Wales, Australia, 2001-2018: Findings from the OATS retrospective cohort study

BACKGROUND

To examine, among a cohort of opioid dependent people with a history of opioid agonist treatment (OAT), the frequency and incidence rates of non-fatal overdose (NFOD) hospital separations over time, by age and sex.

METHODS

Retrospective cohort study of people with a history of OAT using state-wide linked New South Wales (NSW) data. The incidence of NFOD hospital separations involving an opioid, sedative, stimulant or alcohol was defined according to the singular or combination of poisoning/toxic effect using ICD-10-AM codes. Crude incidence rates were calculated by gender, age group and calendar year.

RESULTS

There were 31.8 (31.3-32.3) NFOD per 1,000 person-years (PY). Opioid NFOD incidence was higher in women than men: incidence rate ratio (IRR) 1.11 per 1,000PY; 95 %CI: [1.06-1.17]; women had higher sedative NFOD rates than men, IRR 1.27 per 1,000PY [1.21-1.34]. Participants ≤25 years, 26-30yrs, and 31-35yrs had higher incidence of opioid NFOD compared to 46+yrs, with IRRs of: 1.45 per 1,000PY; [1.32-1.59]; 1.20 per 1,000PY; [1.11-1.30] and 1.22 per 1,000PY; [1.13-1.32], respectively. Between 2006-7 and 2016-17, the cohort accounted for 19 % of NSW opioid NFOD episodes, 12 % of sedative, 14 % of stimulant and 5 % of acute alcohol-related NFOD.

CONCLUSIONS

Hospital stays due to NFOD are a relatively frequent occurrence among opioid-dependent people. There are clear differences in rates and substances involved by sex, age and over time. Evidence-based interventions that prevent overdose among people who are opioid dependent need to be delivered to scale, including widespread community provision of naloxone.

Risk of discharge against medical advice among hospital inpatients with a history of opioid agonist therapy in New South Wales, Australia: A cohort study and nested crossover-cohort analysis

BACKGROUND

People who use illicit opioids have high rates of hospital admission. We aimed to measure the risk of discharge against medical advice among inpatients with a history of opioid agonist therapy (OAT), and test whether OAT is associated with lower risk of discharge against medical advice.

METHODS

We conducted a cohort study of patients admitted to hospital in an emergency between 1 August 2001 and 30 April 2018 in New South Wales, Australia. All patients had a previous episode of OAT in the community. The main outcome was discharge against medical advice, and the main exposure was whether patients had an active OAT permit at the time of admission.

RESULTS

14,035/116,957 admissions (12 %) ended in discharge against medical advice. Admissions during periods of OAT had 0.79 (0.76-0.83; p < 0.001) times the risk of discharge against medical advice, corresponding to an absolute risk reduction of 3.0 percentage points. Risk of discharge against medical advice was higher among patients who were younger, male, identified as Aboriginal and/or Torres Strait Islander, and those admitted for accidents, drug-related reasons, or injecting-related injuries (such as cutaneous abscesses). In a subsample of 7793 patients included in a crossover-cohort analysis, OAT was associated with 0.84 (95 % CI 0.76-0.93; p < 0.001) times the risk of discharge against medical advice.

CONCLUSIONS

Among patients with a history of OAT, one in eight emergency hospital admissions ends in discharge against medical advice. OAT enrolment at the time of admission is associated with a reduction of this risk.

Life expectancy of people who are dependent on opioids: A cohort study in New South Wales, Australia

BACKGROUND

People who are dependent on opioids have increased risk of premature death, but there are few estimates of life expectancy.

METHODS

We calculated age-specific mortality rates in a cohort of people who had at least one prescription of an opioid agonist (methadone or buprenorphine) between 2001 and 2018 in New South Wales, Australia. We used life tables to estimate life expectancy at age 18. We also estimated the potential years of life lost before age 75, decomposed by cause of death.

RESULTS

The cohort included 47,197 people, with a median of 9.8 years of follow-up. 5097 participants died, and the standardised mortality ratio (compared to the general population of New South Wales) was 6.06 (95% CI 5.90-6.23). Life expectancy at age 18 was an additional 47.5 years (95% CI 42.9-50.5) for men and 50.7 years (95% CI 45.4-54.8) for women; deficits of 14.7 and 15.8 years respectively when compared to the general population. The largest cause of death was non-communicable physical diseases, which accounted for 47% of deaths in life tables for men and 42% for women. Drug-related deaths accounted for 16% of deaths for men and 19% for women, but due to the young age at which these deaths occur, they contributed approximately one third of potential years of life lost.

CONCLUSION

In common with people with serious mental illnesses, people who are dependent on opioids have substantially reduced life expectancy. In both populations most excess deaths relate to non-communicable physical diseases.

Artificial Intelligence in Pharmacoepidemiology: A Systematic Review. Part 1-Overview of Knowledge Discovery Techniques in Artificial Intelligence

Aim

To perform a systematic review on the application of artificial intelligence (AI) based knowledge discovery techniques in pharmacoepidemiology.

Study Eligibility Criteria

Clinical trials, meta-analyses, narrative/systematic review, and observational studies using (or mentioning articles using) artificial intelligence techniques were eligible. Articles without a full text available in the English language were excluded.

Data Sources

Articles recorded from 1950/01/01 to 2019/05/06 in Ovid MEDLINE were screened.

Participants

Studies including humans (real or simulated) exposed to a drug.

Results

In total, 72 original articles and 5 reviews were identified via Ovid MEDLINE. Twenty different knowledge discovery methods were identified, mainly from the area of machine learning (66/72; 91.7%). Classification/regression (44/72; 61.1%), classification/regression + model optimization (13/72; 18.0%), and classification/regression + features selection (12/72; 16.7%) were the three most frequent tasks in reviewed literature that machine learning methods has been applied to solve. The top three used techniques were artificial neural networks, random forest, and support vector machines models.

Conclusions

The use of knowledge discovery techniques of artificial intelligence techniques has increased exponentially over the years covering numerous sub-topics of pharmacoepidemiology.

Systematic Review Registration

Systematic review registration number in PROSPERO: CRD42019136552.