Abstract
Objective To describe team-based care use among a cohort of people who use drugs (PWUD) and to determine factors associated with receipt of team-based care.
Design A cohort study using survey data collected between March and December 2013. These data were then linked to provincial-level health administrative databases to assess patterns of primary care among PWUD in the 2 years before survey completion.
Setting Ottawa, Ont.
Participants Marginalized PWUD 16 years of age or older.
Main outcome measures Patients were assigned to primary care models based on survey responses and then were categorized as attached to team-based medical homes, attached to non–team-based medical homes, not attached to a medical home, and no primary care. Descriptive statistics and multinomial logistic regression were used to determine associations between PWUD and medical home models.
Results Of 663 total participants, only 162 (24.4%) received team-based care, which was associated with high school level of education (adjusted odds ratio [AOR] = 2.18; 95% CI 1.13 to 4.20), receipt of disability benefits (AOR = 2.47; 95% CI 1.22 to 5.02), and HIV infection (AOR = 2.88; 95% CI 1.28 to 6.52), and was inversely associated with recent overdose (AOR = 0.49; 95% CI 0.25 to 0.94). In comparison, 125 (18.8%) received non–team-based medical care, which was associated with university or college education (AOR = 2.31; 95% CI 1.04 to 5.15) and mental health comorbidity (AOR = 4.18; 95% CI 2.33 to 7.50), and was inversely associated with being detained in jail in the previous 12 months (AOR = 0.51; 95% CI 0.28 to 0.90).
Conclusion Although team-based, integrated models of care will benefit disadvantaged groups the most, few PWUD receive such care. Policy makers should mitigate barriers to physician care and improve integration across health and social services.
People who use drugs (PWUD) experience substantial comorbidity, disability,1,2 and premature mortality.3 Most PWUD report having unmet health needs,4,5 with high associated rates of emergency department visits and hospital admissions for mental health and substance use diagnoses, soft tissue infections, pneumonia, and other issues.6-8 We previously found that multiple emergency department visits were about 50% less likely among a cohort of PWUD if they had a regular family physician6; however, only 56.2% of the cohort accessed regular primary care. Although policy makers have emphasized the importance of medical home models in ensuring timely access to the right services9 for people with substance use disorder, the literature about actual access is sparse.10,11 In general, disadvantaged patients in Canada are less likely to access medical homes, particularly homes with team-based care.12,13 A recent study in Ontario has demonstrated that people receiving opioid agonist therapy (OAT) are less likely to receive chronic disease prevention and management than other Ontarians, but are more likely to receive such care if they receive primary care within medical home models, particularly in models with team-based care.14
The aim of this study was to describe the association of attachment to a medical home among a cohort of PWUD enrolled in the community-based Participatory Research in Ottawa: Understanding Drugs (PROUD) study.15 We created a unique data set by linking the rich survey responses of the PROUD cohort to provincial health services administrative databases in a single-payer system with first dollar universal coverage for physician services. The objectives were to describe team-based care use among PWUD and to determine factors associated with receipt of team-based care. The main outcome was attachment to a medical home.
METHODS
Study design
We conducted a cohort study using survey data collected between March and December 2013, and linked those data to provincial-level health administrative databases to assess patterns of primary care in the 2 years before survey completion.
Setting
Ontario has several models of primary care: reimbursement based on fee-for-service, capitation payments based on patient enrolment to a physician, and a combination of both. Capitation models vary organizationally, with some including interprofessional teams (family health teams [FHTs]).16 Community health centres (CHCs) are a distinct model of interprofessional team care with salaried physician remuneration. For the purposes of this study, we distinguished between interdisciplinary team-based medical homes (FHTs and CHCs) and non–team-based medical homes (family health groups, family health networks, non–FHT family health organizations, and comprehensive care models).
Data sources
As described previously,15 the cross-sectional PROUD study used a street-based peer recruitment and snowball sampling approach to recruit and enrol participants, focusing on socially and economically marginalized PWUD. Eligibility criteria included being 16 years of age or older and having injected or smoked drugs other than marijuana in the 12 months before enrolment (March to December 2013). Participants completed a peer- or medical student–administered survey with questions about sociodemographic information, substance use, environmental-structural factors, and health and social services use. The PROUD study activities are governed by a community advisory committee of PWUD and allies.
We linked consenting participants’ survey responses with health administrative databases held at ICES, an independent non-profit research institute in Toronto, Ont, whose legal status under Ontario’s health information privacy laws allows it to collect and analyze health care and demographic data without consent for the purposes of health system evaluation and improvement (Supplemental Table 1, available from CFPlus*). The ICES data sets employ unique encoded identifiers, which were used to link PROUD participants deterministically based on their reported Ontario Health Insurance Plan (OHIP) numbers if available, or probabilistically based on their names, dates of birth, and postal codes. Following linkage, we identified participants with duplicate enrolment and retained responses with the most complete data. Data sets were analyzed at ICES.
Variables
We categorized gender using self-reported gender in the PROUD survey except when gender was missing or when participants reported gender as “2-spirit” or “other”; in that case, we used data from health cards. We excluded transgender individuals because of the risk of re-identification (< 6 participants). We used postal codes to assign neighbourhood income into quintiles. We classified comorbidity using the Johns Hopkins Adjusted Clinical Group System case-mix assignment software, version 10.17 Comorbidity was classified as low (≤ 5 aggregated diagnosis groups [ADGs]), medium (6 to 9 ADGs), or high (≥ 10 ADGs). We used validated ICES algorithms to classify the prevalence of mental health conditions (Supplemental Table 2a, available from CFPlus*)18 and HIV infection, defined by having 3 or more physician claims in 3 years with OHIP diagnosis codes of 042, 043, or 044.19 We calculated the number of primary care visits from the OHIP and CHC databases.
Outcomes
We used enrolment tables provided by the Ontario Ministry of Health and Long-Term Care to assign family physicians to patients who were formally rostered in a primary care model on their PROUD survey date. Unrostered patients were assigned in 2 steps: first, we virtually rostered patients to the family physician responsible for most of each patient’s costs of primary care services in the previous 2 years.20 We excluded visits that were only for opioid substitution therapy, as there is debate among our community advisory committee and in the literature about the extent of primary care services provided within specialized OAT clinics14,21 (Supplemental Table 2b, available from CFPlus*). Second, for patients who received any care at CHCs, which do not submit billing codes to OHIP, we determined whether the virtually rostered physician or the CHC physician provided most of the visits, and then assigned the patient to the appropriate physician. We then categorized patients based on the practice model of their family physician: attached to a team-based medical home (FHTs and CHCs); attached to a non–team-based medical home; not attached to a medical home; and no primary care.22
Analyses
We used descriptive statistics to summarize our cohort, stratified by primary care model, and included measures of central tendencies and dispersion. We compared patient characteristics between primary care models using Wilcoxon rank sum tests for continuous variables and χ2 tests or Fisher exact tests as appropriate for categorical variables. We used multinomial logistic regression to analyze variables associated with being in a team-based or a non–team-based medical home and included all potential covariates, guided by our understanding of access to care. We used a nonparsimonious approach based on conceptual understanding, as guided by our community advisory committee, but excluded those that we judged likely to be collinear (such as “sex work ever” and “received drugs, money, or gifts for sex in the past 12 months”). For several PROUD variables, participants could respond with “no answer” or “do not know or unsure”; some variables had missing values because of human error. Our primary analyses used a complete case approach; we also conducted a sensitivity analysis in which response categories were dichotomized (yes vs no), with the “no” category including any response other than “yes.” We reported associations as odds ratios with 95% CIs. Cell sizes of 6 or less were reported in aggregate. We used SAS statistical software, version 9.4, to conduct all analyses.
This study received approval from the Ottawa Health Science Network Research Ethics Board. The use of administrative data was authorized under section 45 of Ontario’s Personal Health Information Protection Act, which does not require review by a research ethics board.
RESULTS
Between March and December 2013, 858 PROUD participants completed the survey and 798 agreed to be linked to ICES. After duplicate enrolment data and participants without OHIP were excluded, 663 participants were successfully linked. Demographic characteristics have been reported previously and are shown in Supplemental Table 3, available from CFPlus.6* We compared PROUD participants by medical home model based on their primary care assignment in the 2 years before completion: 162 (24.4%) were assigned to a team-based medical home, 125 (18.8%) to a non–team-based medical home, 252 (38.0%) to a non–medical home, and 124 (18.7%) received no care. Descriptive comparisons by medical home model are shown in Table 1.23
PROUD participant descriptive characteristics by medical home model: N = 663.
In our complete case multivariable analysis (n = 533; Table 2),23 after adjustment we found the following: when compared with no attachment to a medical home (ie, non–medical home and no care combined), attachment to a team-based medical home was associated with high school education (adjusted odds ratio [AOR] = 2.18; 95% CI 1.13 to 4.20), receipt of disability benefits (AOR = 2.47; 95% CI 1.22 to 5.02), and HIV infection (AOR = 2.88; 95% CI 1.28 to 6.52), and was inversely associated with recent overdose (AOR = 0.49; 95% CI 0.25 to 0.94). Attachment to a non–team-based model, when compared with attachment to a non–medical home and no care combined, was associated with university or college education (AOR = 2.31; 95% CI 1.04 to 5.15) and mental health comorbidity (AOR = 4.18; 95% CI 2.33 to 7.50), and was inversely associated with being detained in jail in the previous 12 months (AOR = 0.51; 95% CI 0.28 to 0.90).
Multinomial regression complete case analysis of PROUD participant characteristics associated with primary care models (n = 533): Excluded from medical home was used as the reference variable*
In the sensitivity analysis where we collapsed all “no answer,” “do not know,” and missing responses with “no” responses, effect sizes between models were similar (Supplemental Table 4, available from CFPlus*).
DISCUSSION
Our community-based participatory research approach allowed us to gather survey data from a marginalized population of PWUD, which we then linked to population-level data in a setting with universal health insurance and a variety of primary care organizational models. While we found that those with HIV infection, with mental health comorbidity, and on disability support were more likely to be in medical home models, overall attachment to any medical home was low (43.3%) compared with more than 80% for the Ontario population as a whole in 2011.12,20 We also found that some substantially disadvantaged groups were omitted, such as people with less formal education, those who have experienced imprisonment, and those with recent overdose. We also found that only one-quarter of PWUD were receiving care in team-based, integrated models, despite these models being explicitly designed for people with comprehensive needs, such as FHTs and CHCs.
Medical homes are intended to improve access to comprehensive care for a rostered patient population, and some are enhanced by interdisciplinary teams, including mental health workers.13,24,25 Patients in medical homes have improved prevention and management of chronic conditions,12,20,22,26 as well as improved access to mental health services.27 Those not belonging to medical homes are also at greater risk of turning to the emergency department for health care services or being admitted to hospital.22,28-30 Our findings are consistent with general population studies demonstrating that both team-based and non–team-based models, despite their intended outcomes, are less likely to care for, and thus benefit, those who need them most, such as people in low-income neighbourhoods, new immigrants, and those with physical and mental health comorbidity,12,22,31 even in our setting with universal access to physician services. Our results are also aligned with a recent study in Ontario that found that people receiving OAT are less likely to receive prevention and management of chronic conditions compared with matched controls.14 While rates of medical home attachment were similar to those found in our study, that population also found that medical home attachment, in particular to team-based care, improved the primary care received.
The reasons for suboptimal medical home attachment despite the many health care needs of PWUD are likely multifold. Stigma remains a serious barrier to receipt of primary care among people with substance use disorders.32 Specifically, discrimination directly corresponds to unmet care needs by this population.33 We have previously found that the one-third of the PROUD cohort who had ever received OAT had improved primary care engagement (unpublished data), and that addictions treatment had the potential to link PWUD to primary care. However, integration of addictions and primary care services remains suboptimal.34,35 Given the specialized nature of these services, which are often limited to high-volume OAT providers, the integration of OAT care within medical homes has not had adequate reach in Canada.14,21
Limitations
Our study has some limitations. Survey results comprised self-reported data about highly stigmatized or illegal practices, which may have contributed to reporting biases. We used street-based peer recruitment to improve response and representativeness compared with standard recruitment methods,36 which means that our findings may not be generalizable to non–street-based populations. Finally, ICES data do not capture most visits to non–physician primary care providers, including nurse practitioners, nurses, and other allied health professionals, or visits by people without active health cards, a situation which is not uncommon among very disadvantaged groups.
Conclusion
These findings should spur policy makers to extend the reach of team-based care to mitigate barriers to physician care and improve integration across health and social service needs to help patients with health and social complexity.32,37-39 Models to integrate medication-assisted therapy for substance use disorder within primary care should incorporate pharmacologic therapy, psychosocial services, service integration, and education and outreach.40
Acknowledgment
Parts of this material are based on data and/or information compiled and provided by CIHI. However, the analyses, conclusions, opinions and statements expressed in the material are those of the author(s), and not necessarily those of CIHI. This study was supported by ICES, which is funded by an annual grant from the Ontario Ministry of Health (MOH) and the Ministry of Long-Term Care (MLTC). The opinions, results and conclusions reported in this paper are those of the authors and are independent from the funding sources. No endorsement by ICES, the MOH or MLTC is intended or should be inferred. The authors acknowledge the participants who agreed to share their information with the study team and the contributions of the full Participatory Research in Ottawa: Understanding Drugs Community Advisory Committee.
Notes
Editor’s key points
▸ Overall attachment to medical homes among marginalized people who use drugs (PWUD) in Ottawa, Ont, is substantially lower than among the general Ontario population. Only one-quarter of PWUD receive care in team-based medical home models.
▸ Attachment to a team-based medical home is associated with high school level of education, receipt of disability benefits, and HIV infection, and is inversely associated with those who have experienced a recent overdose.
▸ Stigma and discrimination are both likely to be reasons for suboptimal medical home attachment among PWUD.
Points de repère du rédacteur
▸ Dans l’ensemble, le rattachement à un centre de médecine de famille chez les personnes marginalisées qui consomment des drogues (PQCD) à Ottawa (Ontario) est considérablement moins élevé que dans la population ontarienne en général. Seulement le quart des PQCD reçoivent des soins dans des modèles semblables à celui du centre de médecine de famille en équipe.
▸ Le rattachement à un centre de médecine en équipe est associé à une scolarité du niveau secondaire, à la réception de prestations pour incapacité et à une infection au VIH, et est inversement associé aux personnes ayant vécu une surdose récente.
▸ La stigmatisation et la discrimination sont 2 des raisons probables expliquant un rattachement sous-optimal à un centre de médecine de famille chez les PQCD.
Footnotes
↵* Supplementary tables 1 to 4 are available from https://www.cfp.ca. Go to the full text of the article online and click on the CFPlus tab.
Contributors
All authors contributed to the concept and design of the study; data gathering, analysis, and interpretation; and preparing the manuscript for submission.
Competing interests
None declared
This article has been peer reviewed.
Cet article a fait l’objet d’une révision par des pairs.
- Copyright © 2022 the College of Family Physicians of Canada
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.