Cohort profile: The provincial substance use disorder cohort in British Columbia, Canada

Illicit drug use is associated with severe health-related harms and an elevated risk of mortality. In Canada and globally, people with substance use disorder (PWSUD) experience high rates of co-occurring mental illness, infectious disease and chronic conditions that are further exacerbated by high rates of poverty and homelessness.1,2 However, despite their complex health challenges, PWSUD face substantial barriers to care.3–6 Historically, primary care, substance use services and mental health care services have developed and operated separately in North America.7,8 As a result, PWSUD often access care from multiple sources9 and rely heavily on acute and emergency services.10–14 This kind of fragmented care delivery has been associated with poor health outcomes.2,15 Nonetheless, despite a growing recognition of the importance of integrated care for this population, significant barriers to health care integration remain.2,16,17 Improved integration across the continuum of health care services for PWSUD, including improved linkages between acute, primary and community-based services, is urgently needed. In British Columbia (BC), Canada, a public health emergency was declared in April 2016 following a rapid increase in drug-related deaths due to contamination of the illicit drug supply with fentanyl and its analogues.18,19 In 2017, there were 1495 illicit drug overdose deaths in the province (rate 30.4 per 100 000), representing an increase of 51% from 2016 and 182% from 2015.19 The announcement prompted a coordinated and multifaceted public health response across the province, including the establishment of emergency harm reduction services,20–22 numerous policy and practice initiatives to improve treatment availability for PWSUD23–27 and a newly formed Ministry of Mental Health and Addictions to coordinate these efforts.28 Much of this response was concentrated in Vancouver’s Downtown Eastside (DTES) neighbourhood, where a large proportion of the province’s most vulnerable PWSUD reside and access health care services. The DTES has historically experienced high rates of poverty, homelessness, communicable disease and complex concurrent substance use and mental health disorders,29–31 presenting unique challenges to the delivery of health care services in the neighbourhood. Before the public health emergency declaration, Vancouver Coastal Health Authority (VCH) planned a systemic redesign of health care services in the DTES. A key focus of this redesign was to improve access to coordinated, integrated health care.32 The Downtown Eastside Second Generation Strategy (DTES-2GS) included: the establishment of a low-barrier addiction clinic providing rapid access to opioid agonist treatment (OAT); a new community health clinic for people with multiple concurrent disorders; integrated care teams within three existing community health clinics; intensive case management teams; a new drop-in centre; partnerships with private clinics to improve care coordination with public health care services; staff competency training including Indigenous cultural safety, trauma-informed care, harm reduction and pain management; and the integration of peer navigators into health care teams.32–35 Within this context, we developed the provincial cohort of PWSUD and a key subgroup DTES-2GS to monitor service use, health care integration and patient outcomes in British Columbia on a continuing basis and also, to provide the foundation for evaluative studies on the costs and benefits of the DTES-2GS. A unique and unprecedented linkage of local and provincial health administrative data sources was conducted to develop a set of quantitative indicators to achieve these evaluative aims. British Columbia operates as a single-payer health care system, offering some of the world’s most comprehensive health administrative datasets. However, provincial data sources do not capture all services specific to the DTES-2GS, which are necessary to effectively evaluate its impact on health care integration across the continuum of health care services available in the DTES. To address this limitation, we linked provincial-level data with VCH-specific data sources capturing community health care services and acute care contacts in the DTES. The resulting database offers a critical opportunity to evaluate the impact of these health system changes amid an unprecedented public health emergency. We secured funding for this cohort from the Canadian Institutes of Health Research, with additional funding support from VCH. The cohort was defined based on an individual-level linkage of seven health administrative databases held by Population Data BC (provincial data holdings) and VCH data stewards. The Discharge Abstract Database (DAD),36 the Medical Services Plan (MSP) database,37 the PharmaNet database,38 the National Ambulatory Care Reporting System (NACRS)39 and the Vital Statistics database were provided to the study team by Population Data BC.40 These datasets were linked with two local VCH datasets: the EDMart database and the CommunityMart database. The datasets were deterministically linked based on unique personal health numbers used to track individual contacts with the health care system which are recorded in each database. We describe each of the component databases included in our analyses in Figure 1. We also present key features of the databases in Supplementary Table S1, available as Supplementary data at IJE online, a list of hospitals reporting to NACRS in Supplementary Table S2, available as Supplementary data at IJE online, and community-based services captured in the CommunityMart database in Supplementary Table S3, available as Supplementary data at IJE online. Data linkage and key variables of the provincial and local databases. Acronyms: BC: British Columbia, VCH: Vancouver Coast Health Authority, DTES-2GS: Downtown Eastside Second Generation Strategy, ICD-9/10-CA: International Statistical Classification of Diseases and Related Health Problems (ICD), Ninth and Tenth Revisions, Canada. Data linkage and key variables of the provincial and local databases. Acronyms: BC: British Columbia, VCH: Vancouver Coast Health Authority, DTES-2GS: Downtown Eastside Second Generation Strategy, ICD-9/10-CA: International Statistical Classification of Diseases and Related Health Problems (ICD), Ninth and Tenth Revisions, Canada. The provincial cohort includes all people with any indication of substance use disorder (SUD) in any of the linked health administrative databases between 1 April 2009 and 31 March 2017. The case-finding algorithm used to identify this population is described in Supplementary Table S4, available as Supplementary data at IJE online. Two key subgroups comprise the provincial substance use disorder cohort: the DTES-2GS cohort and non-DTES-2GS PWSUD. The DTES-2GS cohort is the subset of the provincial cohort selected based on an indication of DTES residency in the DAD or EDMart, or receiving at least one community-based service in the DTES within the study period. This cohort will primarily inform the evaluation of the DTES-2GS through the analysis of care pathways of people exposed to the DTES-2GS policy changes. Other members of the provincial SUD cohort will provide a basis of comparison to evaluate the effects of specific services and policies introduced as part of the DTES-2GS. As of 31 March 2017, there were 162 099 individuals in the provincial cohort, with 22 579 (13.9%) included in the DTES-2GS cohort (Table 1). Demographic characteristics and health service use rates of cohort participants DTES, Downtown Eastside; DTES-2GS, Downtown Eastside Second Generation Strategy; PWSUD, people with substance use disorder; IQR: inter-quartile range; ER: emergency room; ICD-9/10-CA, International Statistical Classification of Diseases and Related Health Problems (ICD), Ninth and Tenth Revisions, Canada. Individuals having DTES indication, determined by known residential postal code, or homeless user of DTES-2GS services in at least one health care record during follow-up. Determined by ICD-9-CA code V60.0, V60.1 or ICD-10-CA code Z59.0, Z59.1 in at least one health care record during follow-up, or homeless user of DTES-2GS services. At least one PharmaNet record of enrolment in Pharmacare Plan C during follow-up. DTES-2GS cohort: emergency department admission was available from 1 April 2009 onwards. Control group: emergency department admission was available from 1 April 2012 onwards. No health care record 66 months preceding end of study follow-up date 31 March 2017. Demographic characteristics and health service use rates of cohort participants DTES, Downtown Eastside; DTES-2GS, Downtown Eastside Second Generation Strategy; PWSUD, people with substance use disorder; IQR: inter-quartile range; ER: emergency room; ICD-9/10-CA, International Statistical Classification of Diseases and Related Health Problems (ICD), Ninth and Tenth Revisions, Canada. Individuals having DTES indication, determined by known residential postal code, or homeless user of DTES-2GS services in at least one health care record during follow-up. Determined by ICD-9-CA code V60.0, V60.1 or ICD-10-CA code Z59.0, Z59.1 in at least one health care record during follow-up, or homeless user of DTES-2GS services. At least one PharmaNet record of enrolment in Pharmacare Plan C during follow-up. DTES-2GS cohort: emergency department admission was available from 1 April 2009 onwards. Control group: emergency department admission was available from 1 April 2012 onwards. No health care record 66 months preceding end of study follow-up date 31 March 2017. Table 1 provides demographic and health service use characteristics and Table 2 provides health conditions of each of the three cohort subgroups. The provincial cohort was predominantly male (60.0%), with a median age of 39.9 [interquartile range (IQR): 27.7, 51.3] years at the first health service use record. Just over half (53.1%) of the cohort had a history of receiving income assistance and 8.1% were homeless at least once during the study period. The majority (68.5%) of the provincial cohort had at least one emergency department (ED) visit, 82.4% had at least one hospitalization and nearly everyone had at least one physician billing and drug dispensation record (99.2% and 95.2%, respectively). We identified 57 176 (35.3%) individuals with opioid use disorder (OUD), 151 325 (93.4%) individuals with non-opioid SUD and 138 990 (85.7%) individuals with a mental health condition. Physician billing records exclusively captured approximately half of all non-opioid SUD (49.9%) and mental health conditions (55.0%) and a small number (12.6%) of OUD. A further 21.0% of non-opioid SUD were identified only in acute care (Figure 2). Identification of substance use disorders and mental health conditions among cohort participants DTES-2GS, Downtown Eastside Second Generation Strategy; PWSUD, people with substance use disorder. Numerator: number of diagnosed individuals; denominator: number of cohort participants. Non-opioid or non-alcohol substance use disorder. Identification of substance use disorders and mental health conditions among cohort participants DTES-2GS, Downtown Eastside Second Generation Strategy; PWSUD, people with substance use disorder. Numerator: number of diagnosed individuals; denominator: number of cohort participants. Non-opioid or non-alcohol substance use disorder. Identification of substance use disorders and mental health conditions. Acronyms: AUD: alcohol use disorder; OUD: opioid use disorder; SUD: non-opioid, non-alcohol substance use disorder; MH: mental health, PWSUD: people with substance use disorder; MSP: Medical Services Plan; DAD: District Abstract Database; ED: Emergency Department. Identification of substance use disorders and mental health conditions. Acronyms: AUD: alcohol use disorder; OUD: opioid use disorder; SUD: non-opioid, non-alcohol substance use disorder; MH: mental health, PWSUD: people with substance use disorder; MSP: Medical Services Plan; DAD: District Abstract Database; ED: Emergency Department. Compared with non-DTES-2GS PWSUD, a higher percentage of DTES-2GS cohort participants were homeless (33.8% vs 4.0%) and received income assistance (86.2% vs. 47.7%) (Table 1). At the end of follow-up, more DTES-2GS cohort members visited the ED (90.0% vs 65.1%) and were hospitalized (86.8% vs 81.6%) compared with non-DTES-2GS PWSUD. In the DTES-2GS cohort, we identified 21 649 (96.0%) people with non-opioid SUD, 12 515 (55.5%) people with opioid use disorder (PWOUD) and 19 943 (88.3%) people with a mental health condition (Table 2). This cohort captures health care use longitudinally. Data capture in the initial extract ran from 1 April 2009 to 31 March 2017. Since follow-up depends on health system contacts, individuals are lost to follow-up if they move out of British Columbia or die. Individuals with no death record at end of the study period, and no records in the health administrative databases between 1 September 2011 and 31 March 2017, were classified as lost to follow-up. This 66-month cut-off period was determined empirically based on the 97.5th percentile in gap times between points of contact with health care services captured in our database. The detailed information regarding the determination of the cut-off period is described elsewhere.41 The median duration of follow-up of the provincial cohort was 6.8 (IQR: 3.8, 7.8) years from first health care service contact until the end of study follow-up or death; 11.2% of the cohort died during follow-up and another 6.5% were lost to follow-up; 2.2% of the individuals who were lost to follow-up and 14.1% of the individuals who died were in the DTES-2GS cohort. Of the total 133 437 individuals alive and not administratively censored at the end of the study period, 14.8% were in the DTES-2GS cohort. The cohort will be updated annually as new data become available. We obtained information on individuals’ demographic characteristics including age, gender, location, DTES residency, receipt of income assistance, homelessness status at the time of service use, and death through the provincial vital statistics database. The DAD contains information on all admitted patient services at all hospitals.42 For acute care, the resource intensity weight, comorbidity code, Major Clinical Category code43 and the Case Mix Group Plus code43,44 enable researchers to estimate daily costs per hospital stay. The PharmaNet database includes information on cost of medications dispensed by the pharmacies in the province; it provides a unique opportunity to estimate the province-wide spending on prescription medications for SUD and co-existing conditions. Finally, the CommunityMart database documents individual contacts with community-based services within VCH. Similar data are not available for other regions of the province. Together, these datasets allow for a robust examination of health care integration across the near-full spectrum of health care services available to cohort participants. This linkage makes the identification and quantification of care pathways between these services possible over time, providing critical time-to-event data for a wide variety of health care service endpoints and patient outcomes. Key novel measures of health care integration include: the receipt of community-based (e.g. housing, detoxification, sobering, support recovery, discharge planning) and primary care services following discharge from ED and hospital; receipt of primary and acute care services following discharge from community-based addiction services; and patient movement between community-based services. These indicators will enable the study team to better identify barriers to care and health care integration for a patient population with complex care needs. An initial draw of VCH data was retrieved in April 2017 and the linkage to the provincial administrative data was made available in August 2019. Our initial steps have been to assess the validity of the under-researched components of the database, provide descriptive analyses of these data to understand and communicate its strengths and weaknesses to our collaborators, and generate measures of health system performance. Studies assessing the validity and comprehensiveness of regionalized databases providing community-based and acute care services are limited.45 Using a subset of the DTES-2GS cohort data, we examined the concordance between ED and hospital diagnostic codes for identifying mental health and SUDs and conducted analyses to identify patient- and ED visit-related factors independently associated with discordance.45 Among 48 116 pairs of ED and hospital discharge diagnoses codes, we found a high (overall agreement = 0.89, positive agreement = 0.74, kappa = 0.67) level of concordance for broad categories of mental health conditions, and a fair (overall agreement = 0.89, positive agreement = 0.31, kappa = 0.27) level of concordance for SUD. A comparable proportion of visits were classified as mental health-related in ED and hospital (21.7% vs and SUD was to be as the primary in ED as to in hospital vs In multiple ED visits during and were associated with of in identifying mental health These support the use of ED primary codes in to hospital records for case of these conditions, and provide the use of ED data for of and of mental health and The CommunityMart database, capturing records of community-based services provided acute care in the DTES, will as a to evaluate low-barrier addiction care, mental health care and We of community-based services captured by the CommunityMart database, and examined the selected of community-based service between 2009 and 2016 (Figure The database captured service provided by teams Table providing primary care, other health care, addiction care (e.g. and support and community care, mental health care services and discharge in the DTES. A total of DTES-2GS cohort participants community-based services in the DTES during this period, new service this period, to of all new were for services, to were to and to were for mental health care services. A small number of diagnosed in the DTES-2GS cohort received new for opioid agonist treatment in community-based services. members of the DTES-2GS cohort community-based services as the first of contact for substance health and were captured only in the CommunityMart database. To support the response to the opioid overdose public health emergency in British Columbia, a is to develop a comprehensive set of health system measures for in British Columbia in and Using a defined cohort from provincial administrative data, unique measures were developed to the of care for with health care across the spectrum of available health care services and for treatment of concurrent disorders; and health care the unique availability of data on community-based services in the DTES-2GS cohort was to develop of these on health care integration between acute and community services. For we ED within of discharge among the provincial cohort, ED an first date of The percentage of ED from in to in 2017. We this for provincial cohort participants with and mental health conditions. In 2017, ED rates were among cohort participants with by mental health disorders (Figure We also the percentage of community-based service within of ED discharge among DTES-2GS cohort participants with OUD. The percentage from in to in 2017 (Figure of substance use mental health and community-based service among the DTES-2GS cohort: April to March of substance use mental health and community-based service among the DTES-2GS cohort: April to March emergency department within of cohort participants with opioid and other substance use disorder cohort participants with opioid use disorder; cohort participants with any mental health = Emergency department (ED) among cohort with an ED for any within = ED among cohort emergency department to emergency department to hospital as as from emergency ED from first of diagnoses to 31 March 2017. emergency department within of cohort participants with opioid and other substance use disorder cohort participants with opioid use disorder; cohort participants with any mental health = Emergency department (ED) among cohort with an ED for any within = ED among cohort emergency department to emergency department to hospital as as from emergency ED from first of diagnoses to 31 March 2017. service within of emergency department Numerator: Emergency department among DTES-2GS cohort participants with opioid use with a community-based within Emergency department among DTES-2GS cohort participants with opioid use emergency department to case addiction care, primary care, care, and other health to emergency department or hospital admission service within of emergency department Numerator: Emergency department among DTES-2GS cohort participants with opioid use with a community-based within Emergency department among DTES-2GS cohort participants with opioid use emergency department to case addiction care, primary care, care, and other health to emergency department or hospital admission We identified 12 within VCH in 2017. Of were in the DTES-2GS cohort. Our initial significant in treatment in VCH. In 2017, of in VCH had opioid agonist of were on treatment and of on treatment were for at least 12 The DTES-2GS cohort is based on a novel linkage between provincial health administrative data and local data capturing community-based service A key of this cohort is that it provides a record of health system in the DTES and across VCH which is unprecedented in the province. to the VCH CommunityMart database for the examination of a wide range of community-based services including detoxification, support and which are not captured the linkage a robust examination of the integration between acute and community-based services which is not possible with provincial health administrative datasets Health care integration is critical for providing care and health outcomes of with complex health needs. In the DTES, a high proportion of the population concurrent and an to access primary This population heavily on community-based or health care services 2 and often to acute and emergency care as the first of This cohort will enable researchers to better identify barriers to care among To the capture of health care services in this we to linkage to the British Columbia for Control and C and the British Columbia for in and drug treatment with additional and treatment data will allow for an of and a unique opportunity to assess the health care use of people with and with the existing cohort data, these databases will nearly the spectrum of health care services provided in some of which have been linked to the provincial health data We the of reporting to CommunityMart between community-based services. to the DTES-2GS, reporting was based on Reporting is and is to increase as the DTES-2GS is harm reduction services, including and overdose an part of the continuum of care for PWSUD, during the public health emergency in opioid However, these services do not individuals to provide Identification as personal health number is for linkage across datasets and evaluation other health care services. The estimate of the PWSUD among the health service in DTES is the PWSUD do not use any other health services harm reduction services, the of individual-level data also in an of the PWSUD in the DTES. Finally, there are some in the provincial health administrative datasets. The PharmaNet database medications dispensed in by the Canadian and and medications dispensed from the British Columbia for in The database some services by in community-based these are or not they a billing the NACRS database was in its data This database was available 1 April and only hospitals in the province were reporting to NACRS at the end of follow-up. However, these hospitals provide care to the majority of the provincial population Table and for analyses have from all database stewards. will be at the of the British Columbia Ministry of Health and Vancouver Coastal information be found at and be by to Supplementary data are available at IJE online. and the and conducted data and and the first and provided critical and to the and secured funding for the have the We Population Data BC and Vancouver Coastal Health Authority staff in data access and and This was by Vancouver Coastal Health Authority and the Canadian Institutes of Health

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.