Skip to main content
NCBI home page
BMJ Open logoLink to BMJ Open
. 2022 Dec 14;12(12):e065362. doi: 10.1136/bmjopen-2022-065362

Who gets access to an interprofessional team-based primary care programme for patients with complex health and social needs? A cross-sectional analysis

Sydney Jopling 1, Walter P Wodchis 1, Jennifer Rayner 1,2,3, David Rudoler 1,4,5,✉
PMCID: PMC9756166  PMID: 36517102

Abstract

Objectives

To determine whether a voluntary referral-based interprofessional team-based primary care programme reached its target population and to assess the representativeness of referring primary care physicians.

Design

Cross-sectional analysis of administrative health data.

Setting

Ontario, Canada.

Intervention

TeamCare provides access to Community Health Centre services for patients of non-team physicians with complex health and social needs.

Participants

All adult patients who participated in TeamCare between 1 April 2015 and 31 March 2017 (n=1148), and as comparators, all non-referred adult patients of the primary care providers who shared patients in TeamCare (n=546 989), and a 1% random sample of the adult Ontario population (n=117 753).

Results

TeamCare patients were more likely to live in lower income neighbourhoods with a higher degree of marginalisation relative to comparison groups. TeamCare patients had a higher mean number of diagnoses, higher prevalence of all chronic conditions and had more frequent encounters with the healthcare system in the year prior to participation.

Conclusions

TeamCare reached a target population and fills an important gap in the Ontario primary care landscape, serving a population of patients with complex needs that did not previously have access to interprofessional team-based care.

Strengths and limitations

This study used population-level administrative health data. Data constraints limited the ability to identify patients referred to the programme but did not receive services, and data could not capture all relevant patient characteristics.

Keywords: PRIMARY CARE, Health policy, HEALTH SERVICES ADMINISTRATION & MANAGEMENT

STRENGTHS AND LIMITATIONS OF THIS STUDY

  • This study accessed administrative health data that captures all patients who participated in TeamCare, and enabled comparison with other patient populations.

  • This study compared these populations using nearly complete and validated information on patient diagnostic characteristics and healthcare utilisation.

  • It was not possible to identify patients who were referred to TeamCare but did not receive services from the programme.

  • Administrative health data do not capture all patient characteristics that may have instigated referral to TeamCare.

Introduction

With an ageing population, growing prevalence of chronic disease and increasing social disparities, health systems and healthcare professionals are grappling with the challenge of caring for people with complex health and social needs. People with complex needs are a heterogeneous population, defined by the interaction of multiple biological, socioeconomic, cultural, environmental and behavioural challenges; these people can experience co-occurring chronic conditions, psychosocial vulnerabilities and/or behavioural health issues.1–5 People with complex needs are at risk for poor outcomes and frequent interactions with the healthcare system.6

The level of support required by people with complex health and social needs is often beyond the capacity of primary care physicians working alone.1 7 Data from the Commonwealth Fund suggest that people with high needs often do not have access to the services they need, such as care coordination, emotional counselling and assistance with managing functional limitations; this is despite having a regular doctor or place of care.8 Patients with unmet needs are likely to report difficulties in accessing care primary care, and are therefore less likely to participate in preventative care and more likely to visit the emergency department (ED).8

Interprofessional primary care teams are suited to address the needs of people with medical and social complexity and have been implemented for that purpose. A review by the Commonwealth Fund identified 28 interventions designed for high needs patients, 25 of which included interprofessional teams.9 Interprofessional teams are groups of professionals from different disciplines collaborating and working toward a common goal of providing comprehensive care for patients and/or populations in partnership with patients and families.10–12 By improving access to comprehensive and appropriate primary care, interprofessional team-based care is expected to reduce inequities in access to healthcare and reduce unmet need.13 14

Evidence suggests that interprofessional team-based interventions are effective in improving health outcomes and reducing acute care utilisation.15–26 Team-based care has been shown to improve care quality,27–29 particularly in the management of chronic illnesses.28 30–32 However, models that link general or family practice to an interprofessional team targeted for complex patients are less represented in the literature on team-based care. Existing models tend to target people with mental illness19 33 or specific conditions like diabetes34 and dementia,35 rather than people with general complex needs.

In Ontario, Canada, two interprofessional primary care models—the Family Health Team (FHT) and the Community Health Centre (CHC), serve about 30% of the population.36 FHTs are a physician-led primary care model that include teams of physicians and other health professions working to provide care to a rostered patient population.37 As of 2019, approximately 3.5 million people were rostered with an FHT.38 CHCs are a community-governed interprofessional model that are mandated to serve vulnerable, marginalised and complex patients. In their mandate to serve vulnerable and complex populations, CHCs may offer a scope of services (eg, community outreach and social services) not available in other models of care. CHCs provided primary care to approximately 250 000 people in 2017.39 The remaining population receives primary care from group or solophysician practices.40 41

Thus, a large proportion of the population does not receive care from interprofessional practices, and there is a reason to believe that these people are disproportionately those with complex health and social needs.42 Evidence suggests that FHTs are less likely to serve patients who require complex care, are low income, are newcomers to the province or live in urban centres.42–44 One study found that 6.1% of the population of Ontario—approximately 725 500 people—had high comorbidity, but that only 15% of these people were rostered to practices offering interprofessional team-based care.45

Recognising that people with complex health and social needs were not accessing interprofessional care, a programme called TeamCare was implemented by several CHCs and some FHTs in Ontario. TeamCare allows patients of non-team physicians to access non-physician (interprofessional) services. These services include, but are not limited to, counselling, community health work, health promotion, dietitian services and chiropody. TeamCare is intended to support patients with complex health and social needs and their physicians by improving the connection between non-team physicians and participating TeamCare sites. The programme model is based on voluntary referral; patients are referred by their own primary care physician, who had the discretion to identify patients with complex needs and did not have the means to access teams elsewhere (eg, through private insurance). Patients did not have to meet a specific set of eligibility criteria to be referred. While receiving services through TeamCare, patients maintain their relationship with their existing primary care physician. A key question is whether voluntary physician participation and referral led to improved access for the target population of people with complex health and social needs who do not already have access to team-based primary care.

The purpose of this study is to address two specific aims: (1) to characterise the patients and physicians participating in TeamCare, and (2) to determine whether TeamCare reached individuals with more conditions and higher complexity health and social needs than the general adult population. The results of this study have implications for the implementation and expansion of TeamCare, and more generally for programmes that rely on voluntary participation and referral.

Methods

Study setting and design

At the time of this study (2015–2017), three distinct programmes existed in Ontario under the umbrella of TeamCare: Primary Care Outreach (PCO), Solo Practitioners in Need (SPiN) and TeamCare (previously People in Need of Teams (PINOT)). PCO operated in Ottawa delivering team-based care services to frail seniors. During the study period, SPiN operated through a network of CHCs in Toronto delivering care to medically complex and socially vulnerable patients. Both PCO and SPiN are referral-based programmes. TeamCare (PINOT) is the most recent iteration of the programme and aimed to move beyond the referral model by emphasising ongoing communication between the referring primary care physician and the interprofessional team. TeamCare (PINOT) has been implemented in several CHCs and a few FHTs in various regions of the province. However, no FHTs were participating in TeamCare during the period of this study; thus, our analysis is limited to CHC participants. In this cross-sectional study, the patients who participated in TeamCare between 1 April 2015 and 31 March 2017, and their primary care physicians were identified and described. The TeamCare exposure group was compared on characteristics related to medical and social complexity, including emergency, primary and specialist care in the year prior to the exposure date. Characteristics of the most responsible physicians of TeamCare participants were compared with those of all other practising primary care physicians in Ontario.

Data

We used administrative health data accessed at ICES (formally known as the Institute for Clinical Evaluative Sciences), a research institute in Toronto, Ontario. All databases used in this study are listed in online supplemental table S1. A database of electronic health record data collected by CHCs was also available as an ICES data holding linked to the administrative databases. The CHC data included a special programme variable that flagged participation in TeamCare, allowing for the identification of TeamCare patients, as well as data on patient encounters, including the date of each contact with the TeamCare programme. Additional administrative data sources provided information on inpatient admissions and ED use, patient and physician sociodemographic, geographic, and socioeconomic characteristics, clinical conditions, and prescription drug use for people 65 years and older or on social assistance. These datasets were linked using unique encoded identifiers and analysed at ICES. Missing data are reported where >1% of patients had missing data for any variable.

Supplementary data

bmjopen-2022-065362supp001.pdf (178.5KB, pdf)

Study population

Patients

All patients who participated in TeamCare between 1 April 2015 and 31 March 2017 were included. Each patient was assigned an index date based on their date of first encounter at a CHC.

Two comparison groups were created to determine whether patients who received TeamCare services reflect the target population of people with complex health and social needs. The first comparison group included non-TeamCare patients who had the same responsible primary care physician as those who accessed TeamCare. To track comparator patients from a comparable point in time, index dates were set for the comparators following the temporal distribution among TeamCare participants. Subjects were assigned a most responsible physician based on the plurality of contacts in the previous 12 months and were excluded if they visited a CHC physician between 1 April 2015 and 31 March 2017. The second comparison group included a 1% random sample of the Ontario population. An index date of 31 March 2017 was assigned to all subjects in the comparison groups.

Baseline characteristics were measured at the index date. Subjects were excluded if they were less than 18 years of age or greater than 105 years of age, were not an Ontario resident, were not eligible for provincial health insurance or were missing data on key variables (age and sex). Subjects recruited for this study were not directly involved in this research.

Physicians

We also compared the most responsible physicians of TeamCare participants to all other practising primary care physicians in Ontario.

Patient and public involvement

No patient involved.

Variable definitions

Patient-level demographic variables included age, sex and rurality. Rurality was defined using the Rurality Index of Ontario (RIO) score. The RIO score is 0-to-100-point index of census subdivision population density and distance to nearest referral centres, where higher scores indicate higher rurality. Patients were grouped into major urban (RIO=0–9), non-major urban (RIO=10–39) and rural (RIO=40+).46

Patient social complexity was measured with an indicator for whether a patient was a recent migrant to Ontario (ie, within the last 10 years), neighbourhood income quintile and marginalisation. Marginalisation was measured using the Ontario Marginalisation Index—an area-based index of measures of dependency, material deprivation, ethic concentration and residential instability.47 Patient medical complexity was measured using the ACG® System Aggregated Diagnosis Groups (ADGs) and Resource Utilisation Bands (RUBs). ADGs are based on International Classiciation of Diseases (ICD)-10 codes and group diagnoses based on severity and likelihood of persistence.48 There are 32 ADGs, which can be further condensed into 12 Collapsed ADGs based on likelihood of persistence, severity and types of healthcare services required.48 RUBs further group the ADGs into six categories based on expected resource use: 0—no use or invalid diagnosis; 1—healthy use; 2- low; 3—moderate; 4—high and 5—very high use.48 ICES-derived disease cohorts were also used for specific chronic conditions. These cohorts are derived using validated algorithms for asthma,49 Congestive Heart Failure (CHF),50 Chronic Obstructive Pulmonary Disease (COPD),51 hypertension52 and diabetes.53 Cohorts were also generated for dementia and chronic psychotic illness using validated algorithms.54 55

To measure healthcare utilisation, patients’ mean non-urgent ED visits (Canadian Emergency Department Triage and Acuity Scale score=4 or 5), all-cause ED visits, primary care physician visits and specialist physician visits in the 12 months prior to index date were assessed.56

Physician characteristics included age, sex, rurality of practice based on RIO score, Canadian medical graduate (yes/no), number of years since graduation, participation in a Family Heath Team, the number patient visits in the previous 12 months and patient roster size.

Statistical analyses

First, crude frequencies of TeamCare patient characteristics were measured. Second, to determine if TeamCare reached a target population of individuals with complex health and social needs, we compared TeamCare participants with the two comparison groups on crude baseline characteristics and healthcare utilisation in the year prior to the date of exposure. The following comparisons were made: (1) TeamCare exposure group versus non-TeamCare patients of the most responsible physicians and (2) TeamCare exposure group versus non-TeamCare 1% random sample of the general population. For comparison across categorical variables, χ2 tests and Cramer’s V were used to assess statistical significance and effect size, respectively.57 58 For continuous variables, t-tests and Hedge’s g59 statistics were used. A p<0.05 was used as a threshold to determine statistical significance. See online supplemental table S2 for the interpretation of Cramer’s V and Hedge’s g effect sizes, noting that while these measures are most suitable for our research questions, the thresholds are context-dependent and should be used cautiously for interpretation. All analyses were conducted using Stata V.13.1.

Results

Characterising TeamCare patients

One thousand one hundred and forty-eight patients flagged as TeamCare patients had a date of first encounter at a CHC between 1 April 2015 and 31 March 2017 and were included in the TeamCare group (see table 1). Across all variables, less than 1% of patients had missing data, and the missing values were evenly distributed across patient groups. Most patients in the TeamCare exposure group were female (63.7%), above the age of 60 (79.6%) and lived in major urban centres (55.4%). Only a small proportion of the group (5.7%) were migrants to Ontario within the last 10 years, based on the first year of OHIP eligibility.

Table 1.

Patient characteristics of the TeamCare exposure group versus comparison groups

TeamCare patients
(reference) N=1148		 Non-TeamCare patients of
most responsible physicians
N=546 989		 Non-TeamCare Ontario Population
1% random sample
N=117 753
Characteristic n (%) n (%) Effect size* P value n (%) Effect sizeƚ P value
 Female 737 (63.7) 307 315 (56.2) 0.007 <0.001 60 143 (51.1) 0.026 <0.001
 Age 0.043 <0.001 0.124 <0.001
 <30 72 (6.2) 79 570 (14.5) 22 808 (19.4)
 30–39 74 (6.4) 71 651 (13.1) 19 576 (16.6)
 40–49 82 (7.1) 79 767 (14.6) 20 266 (17.2)
 50–59 103 (8.9) 106 053 (19.4) 21 913 (18.6)
 60–69 206 (17.8) 97 939 (17.9) 16 675 (14.2)
 70–79 259 (22.4) 67 646 (12.4) 10 132 (8.6)
 80–89 278 (24.0) 35 413 (6.5) 5030 (4.3)
 ≥90 74 (6.4) 8950 (1.6) 1353 (1.1)
 Rurality 0.019 <0.001 0.0868 <0.001
 Major urban 636 (55.4) 359 226 (65.7) 86 241 (73.2)
 Non-major urban 165 (14.4) 102 139 (18.7) 22 241 (18.9)
 Rural 347 (30.2) 84 221 (15.4) 8343 (7.1)
 Migrant to Ontario within last 10 years 66 (5.7) 48 618 (8.9) 0.005 <0.001 13 251 (11.3) 0.0171 <0.001
Neighbourhood income quintile 0.022 <0.001 0.040 <0.001
 Quintile 1 (lowest) 305 (26.6) 88 662 (16.2) 21 830 (18.5)
 Quintile 2 341 (29.7) 102 674 (18.8) 22 755 (19.3)
 Quintile 3 222 (19.3) 108 594 (19.9) 23 289 (19.8)
 Quintile 4 147 (12.8) 120 362 (22.0) 25 301 (21.5)
 Quintile 5 (highest) 131 (11.4) 124 956 (22.8) 24 043 (20.4)
Dependency 0.0218 <0.001 0.066 <0.001
 1 (lowest) 128 (11.1) 134 081 (24.5) 32 249 (27.4)
 2 162 (14.1) 92 793 (17.0) 22 672 (19.3)
 3 132 (11.5) 90 364 (16.5) 20 949 (17.8)
 4 236 (20.6) 91 924 (16.8) 19 627 (16.7)
 5 (highest) 489 (42.6) 135 995 (24.9) 21 585 (18.3)
Material deprivation 0.024 <0.001 0.035 <0.001
 1 (lowest) 95 (8.3) 122 951 (22.5) 20 468 (17.4)
 2 159 (13.9) 123 569 (22.6) 22 932 (19.5)
 3 287 (25.0) 113 453 (20.7) 23 194 (19.7)
 4 231 (20.1) 88 978 (16.3) 24 436 (20.8)
 5 (highest) 375 (32.7) 96 206 (17.6) 26 052 (22.1)
Ethnic concentration 0.018 <0.001 0.066 <0.001
 1 (lowest) 441 (38.4) 126 784 (23.2) 18 482 (15.7)
 2 132 (11.5) 107 252 (19.6) 18 982 (16.1)
 3 233 (20.3) 109 498 (20.0) 21 050 (17.9)
 4 208 (18.1) 119 426 (21.8) 24 218 (20.6)
 5 (highest) 133 (11.6) 82 197 (15.0) 34 350 (29.2)
Residential instability 0.019 <0.001 0.053 <0.001
 1 (lowest) 69 (6.0) 80 187 (14.7) 25 089 (21.3)
 2 141 (12.3) 97 070 (17.7) 22 114 (18.8)
 3 219 (19.1) 114 367 (20.9) 21 383 (18.2)
 4 238 (20.7) 111 087 (20.3) 21 570 (18.3)
 5 (highest) 480 (41.8) 142 446 (26.0) 26 926 (22.9)
 No of ADGs, mean±SD 8±4 6±3 0.55 <0.001 4±3 −0.99 <0.001
Prevalence of chronic conditions
 Asthma 46 (4.0) 19 078 (3.5) 0.0013 0.338 3020 (2.6) 0.009 0.002
 CHF 142 (12.4) 15 444 (2.8) 0.0263 <0.001 2055 (1.7) 0.077 <0.001
 COPD 114 (9.9) 13 440 (2.5) 0.0220 <0.001 1761 (1.5) 0.066 <0.001
 Hypertension 197 (17.2) 73 253 (13.4) 0.0051 <0.001 12 389 (10.5) 0.021 <0.001
 Diabetes mellitus† 145 (12.6) 46 234 (8.5) 0.0069 <0.001 7984 (6.8) 0.023 <0.001
 Chronic psychotic illness‡ 57 (5.0) 7218 (1.3) 0.0146 <0.001 1258 (1.1) 0.036 <0.001
 Dementia 185 (16.1) 14 432 (2.6) 0.0382 <0.001 1761 (1.5) 0.113 <0.001
Resource Utilisation Bands 0.0307 <0.001 0.109 <0.001
 0–1 (no—lowest expected use) 33 (2.9) 15 261 (2.8) 22 278 (18.9)
 2 52 (4.5) 64 172 (11.7) 18 486 (15.7)
 3 456 (39.7) 306 470 (56.0) 54 471 (46.3)
 4 282 (24.6) 107 345 (19.6) 16 184 (13.7)
 5 (highest expected use) 325 (28.3) 53 741 (9.8) 6334 (5.4)
Open in a new tab

*Effect size measure is Cramer’s V for binary/categorical variables and Hedge’s g for continuous variables.

†

‡

ADGs, Johns Hopkins Aggregated Diagnosis Groups; CHF, congestive heart failure; COPD, chronic obstructive pulmonary disease.

The TeamCare group was heavily skewed to the lower income quintiles, with over half (56.3%) of TeamCare patients in the first and second quintiles. On the Ontario Marginalisation Index, the distribution of scores in the TeamCare group were skewed to higher (ie, worse) scores on three of the four factors: dependency, material deprivation and residential instability. Only 11.6% of TeamCare patients lived in areas with high ethnic concentration (score=5), while 38.4% lived in areas with the lowest ethnic concentration (score=1).

TeamCare patients tended to have high expected resource use in the 12 months prior to participation in the programme based on the Johns Hopkins RUBs: 52.9% of patients had high or very high expected resource use (RUB=4–5) and 39.7% had moderate expected resource use (RUB=3). The mean number of ADGs in the sample was 8 (SD=4). In terms of chronic conditions, 17.2% of TeamCare patients had hypertension, 16.1% had dementia, 12.6% had diabetes, 12.4% had CHF, 9.9% had COPD, 5.0% had chronic psychotic illness and 4.0% had asthma.

Comparison of TeamCare patients to comparison groups

On 31 March 2017, a 1% random draw of the Ontario population generated a sample of 117 753 subjects, and 546 989 subjects were identified as other patients of most responsible primary care physicians of TeamCare patients. Descriptive characteristics of the TeamCare patient group and the two comparison groups are presented in table 1, with effect sizes and p values provided for each comparison group in reference to the TeamCare group.

The TeamCare exposure group had a higher proportion of rural patients compared with the non-TeamCare patient group (30.2% TeamCare vs 15.4% non-TeamCare; p<0.001). TeamCare Patients also had a higher mean number of ADGs and higher prevalence of each of the chronic conditions examined; all differences were statistically significant except for asthma. The prevalence of CHF (12.4% vs 2.8%), COPD (9.9% vs 2.5%), diabetes mellitus (12.6% vs 8.5%), chronic psychotic illness (5% vs 1.3%) and dementia (16.1% vs 2.6%) was much higher in the TeamCare exposure group. Patients in the exposure group also had two more (8 vs 6) ADGs on average (p<0.001). Though the difference in overall distribution of patients across RUBs between the two groups was statistically significant, the difference between the proportion of individuals in the two lowest RUBs—representing no or low expected use—was not (2.9% TeamCare vs 2.8% Other Primary Care patients; p=0.862). TeamCare patients had higher mean utilisation across all four utilisation measures in the year prior to their date of first encounter when compared with the patient populations of their most responsible physicians.

Compared with the general population, TeamCare patients were more likely to be female (63.7% vs 51.1%; Cramer’s V=0.0257; p<0.001). The age distributions of the two groups also differed significantly, with TeamCare patients heavily skewed to the older age groups (60 and above). The TeamCare exposure group had a higher proportion of patients living in rural areas (30.2% vs 7.1%; p<0.001) and a lower proportion of recent migrants to the province (5.7% vs 11.3%; Cramer’s V=0.0171; p<0.001).

Overall, the distributions of TeamCare patients and the random sample of the general population across income quintiles differed significantly (Cramer’s V=0.0398; p<0.001). The random sample of the general population was relatively evenly distributed across the five income quintiles, while over half (56.3%) of TeamCare patients lived in areas in the lowest two income quintiles (vs 37.8% of the general population sample). Distributions across each of the Ontario Marginalisation Index dimensions differed significantly between the two groups, with TeamCare patients tending to score higher on dependency (Cramer’s V=0.0663; p<0.001), material deprivation (Cramer’s V=0.0351; p<0.001), and residential instability (Cramer’s V=0.0529; p<0.001), and lower on ethnic concentration (Cramer’s V=0.0657; p<0.001).

The mean of 8 ADGs (SD=4) in the exposure group was double that of the general population (mean=4, SD=3; p<0.001). The TeamCare group also had a significantly higher prevalence of each of the chronic conditions measured. For instance, the prevalence of CHF and Dementia were over 7 and 10 times higher in the exposure group, respectively. The distribution of patients across RUBs differed significantly between the two groups, with TeamCare patients tending to have higher expected resource use than the general population (Cramer’s V=0.1093; p<0.001).

TeamCare patients had higher mean utilisation across all four utilisation measures in the year prior to their date of first encounter when compared with the general population (table 2).

Table 2.

Healthcare utilisation in the year prior to index date

TeamCare
N=1148		 Non-TeamCare patients of most responsible physicians
N=546 989		 Non-TeamCare Ontario population
1% random sample
N=117 753
Characteristics Mean±SD Mean±SD Effect size (Hedge’s g) P value Mean±SD Effect size (Hedge’s g) P value
Non-urgent ED visits 0.50±1.44 0.23±0.81 0.342 <0.001 0.13±0.57 0.640 <0.001
All-cause ED visits 2.01±3.75 0.70±1.72 0.580 <0.001 0.40±1.16 0.956 <0.001
Primary care physician visits 7.77±8.77 5.55±6.68 0.333 <0.001 3.85±5.95 0.655 <0.001
Specialist visits 5.45±6.82 3.15±5.19 0.443 <0.001 2.01±4.27 0.801 <0.001
Open in a new tab

Effect sizes and p values are reported for each comparison group in reference to the TeamCare exposure group.

Characterising physicians of TeamCare patients

Three hundred and fifty-seven physicians were identified as the most responsible primary care providers of TeamCare patients and included in the physician group. The Non-TeamCare primary care physicians group comprised 11 103 general practitioners or family physicians who did not have rostered patients in the TeamCare patient group. See online supplemental file 1 for more details on physician characteristics.

TeamCare physicians were not significantly different from non-TeamCare physicians except on a few characteristics. TeamCare physicians were more likely than non-TeamCare physicians to practice in rural areas (11.5% TeamCare vs 7.2% Other Physicians; p=0.002) and varied in terms of roster size: TeamCare physicians had a median roster size of 1180 (IQR 852–1601), while other physicians had a median roster size of 818 (IQR 0–1417); p<0.001. Surprisingly, the difference in physicians practising in an FHT model was not significant between the two groups: 14.3% TeamCare vs 13.6% other Physicians; Cramer’s V=0.0047; p=0.882).

Interpretation

The comparison of TeamCare patients to non-TeamCare patients of their most responsible primary care physicians suggests that TeamCare patients had greater clinical complexity and were more likely to live in neighbourhoods with lower income and a higher degree of marginalisation than non-TeamCare patients. TeamCare patients had a higher mean number of ADGs, higher expected utilisation (RUB scores) and a higher prevalence of all chronic conditions measured except asthma, including nearly five times the rate of chronic psychotic illness.

Similarly, compared with the general population, TeamCare patients were more likely to live in low-income areas and tended to score higher on most dimensions of the Ontario Marginalisation Index, indicating that TeamCare patients experienced a higher degree of marginalisation than the population on average.

TeamCare patients had more frequent encounters with the healthcare system in the year prior to the intervention relative to both comparison groups. TeamCare patients had a significantly higher mean number of non-urgent ED visits, all-cause ED visits, physician visits and specialist visits.

Patient populations facing complex medical and socioeconomic challenges with high unmet needs are known to experience poor health outcomes and interact frequently with the health system, particularly with the ED.5 8 The findings of this study align with the literature on patients with complex needs; TeamCare patients had significantly higher utilisation of the ED for non-urgent issues as well as for any reason, primary care physician visits and specialist visits in the 12 months prior to entering the programme.

The results of this study suggest that there were few significant differences between the most responsible primary care physicians of TeamCare patients and other physicians in the province who did not participate in TeamCare, except that TeamCare physicians were more likely than non-TeamCare physicians to practice in rural areas and had larger roster sizes and number of visits over the past year. Rurality and physician roster size are dimensions known to be related to access to healthcare,60 and a larger roster has been shown to be associated with lower access to primary care for individuals and gaps in the delivery of prevention, health promotion and chronic disease management.61 In this regard, TeamCare appears to be more common in areas of greater perceived need.

A surprising finding was that the proportion of physicians that used TeamCare practising in an FHT was not significantly lower than that in the general population. Given that TeamCare is targeted to patients who do not have access to an interprofessional primary care team, it is surprising that just under 15% of most responsible physicians of TeamCare patients were practising in an existing interprofessional FHT model. One possible explanation for this finding is that the CHCs participating in TeamCare offered more extensive services (social services in particular) than some FHTs, and physicians felt it appropriate to refer patients to these services.

Overall, the results of this study suggest that the TeamCare initiative reached a target population of patients with complex health and social needs. TeamCare patients were more complex (on average) than the non-TeamCare patient population of their most responsible physicians, suggesting that physicians generally referred patients with higher needs than those not referred TeamCare patients also had more social and medical complexity than the general adult population. The most responsible primary care physicians of TeamCare patients did not differ significantly from other physicians in the province except on geography and roster size, which may have contributed to poorer access for their patients.

An important limitation of this work is that data constraints limited the ability to identify patients who were referred to the programme but did not receive services. As a result, it was not possible to determine whether there were any systematic differences between patients who participated in the programme and those who were referred but did not participate. It is also important to note that the data used in this study were assembled for administrative reporting and payment purposes and do not capture all characteristics that would make a physician likely to refer patients to TeamCare. However, the characteristics described demonstrate that a voluntary referral programme can target patients perceived to be at highest risk for adverse outcomes. There were also limitations in the data that may influence interpretation of the results. One of the programmes, SPiN, contributed very few patients to the overall TeamCare patient sample, while PCO contributed just over 50%. Results were generated by the TeamCare programme, but small sample sizes limited reporting on this level.

The use of administrative health data was a strength of this research; it captured all patients who participated in TeamCare and enabled comparison to other patients cared for by referring physicians and to the general population. In particular, the comparison to a random sample of the general population increased the generalisability of the main findings. The use of administrative data also allowed for the comparison of nearly complete and validated information on patient diagnostic characteristics and healthcare utilisation between participants and representative comparators.

Conclusion

TeamCare fills an important gap in the Ontario primary care landscape, serving a population of patients with complex needs that did not previously have access to interprofessional team-based care. The initiative has grown considerably since it was first implemented and continues to expand to other regions and evolve its programme model to include additional primary care organisations and model types and to serve more patients. The results from this study have the potential to inform further efforts to expand the TeamCare programme model across the province of Ontario, as well as the implementation of other voluntary referral-based interprofessional primary care programmes in other jurisdictions. Future work must further evaluate TeamCare and analyse TeamCare’s impact on the health and healthcare utilisation of its patients.

Supplementary Material

Reviewer comments
bmjopen-2022-065362.reviewer_comments.pdf (148.2KB, pdf)
Author's manuscript
bmjopen-2022-065362.draft_revisions.pdf (1.7MB, pdf)

Footnotes

Twitter: @wwodchis, @RaynerJen

Contributors: SJ, WPW, JR and DR authors made substantial contributions to the conception, design, and interpretation of data. WPW and JR supported the acquisition of the administrative health data. SJ led the statistical analysis. SJ and DR drafted the final manuscript and WPW and JR revised it critically for important intellectual content. SJ, WPW, JR and DR agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. DR is the guarantor for this study.

Funding: 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). This study was also supported by the Health System Performance Research Network (HSPRN grant #06034) and the Ontario SPOR Support Unit. The analyses, conclusions, opinions and statements expressed herein are solely those of the authors and do not reflect those of the funding or data sources; no endorsement is intended or should be inferred

Competing interests: None declared.

Patient and public involvement: Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

Provenance and peer review: Not commissioned; externally peer reviewed.

Supplemental material: This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.

Data availability statement

The data that support the findings of this study are approved for use by data stewards and accessed through a process managed by ICES (www.ices.on.ca/DAS (email: das@ices.on.ca). We are not permitted to share the data used in this analysis with other researchers. The full dataset creation plan and underlying analytic code are available from the authors on request, understanding that the computer programs may rely on coding templates or macros that are unique to this project and to individual data centres and are therefore either inaccessible or may require modification.

Ethics statements

Patient consent for publication

Not applicable.

Ethics approval

The use of data in this project was authorised under section 45 of Ontario’s Personal Health Information Protection Act. The study also received ethics approval from the University of Toronto Research Ethics Board (Protocol #36927).

References

  • 1.Webster F, Rice K, Bhattacharyya O, et al. The mismeasurement of complexity: provider narratives of patients with complex needs in primary care settings. Int J Equity Health 2019;18:107. 10.1186/s12939-019-1010-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Safford MM. The complexity of complex patients. J Gen Intern Med 2015;30:1724–5. 10.1007/s11606-015-3472-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Safford MM, Allison JJ, Kiefe CI. Patient complexity: more than comorbidity. The vector model of complexity. J Gen Intern Med 2007;22 Suppl 3:382–90. 10.1007/s11606-007-0307-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Schaink AK, Kuluski K, Lyons RF, et al. A scoping review and thematic classification of patient complexity: offering a unifying framework. J Comorb 2012;2:1–9. 10.15256/joc.2012.2.15 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Valderas JM, Starfield B, Sibbald B, et al. Defining comorbidity: implications for understanding health and health services. Ann Fam Med 2009;7:357–63. 10.1370/afm.983 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Hong CS, Atlas SJ, Ashburner JM, et al. Evaluating a model to predict primary care Physician-Defined complexity in a large academic primary care practice-based research network. J Gen Intern Med 2015;30:1741–7. 10.1007/s11606-015-3357-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Hochman M, Asch SM. Disruptive models in primary care: caring for High-Needs, high-cost populations. J Gen Intern Med 2017;32:392–7. 10.1007/s11606-016-3945-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Ryan J, Abrams MK, Doty MM, et al. How High-Need patients experience health care in the United States. findings from the 2016 Commonwealth fund survey of High-Need patients. Issue Brief 2016;43:1–20. [PubMed] [Google Scholar]
  • 9.Shah T, Lewis C, Tsega M. Quick reference guide to promising care models for patients with complex needs. Washington, D.C: The Commonwealth Fund, 2019. [Google Scholar]
  • 10.D'Amour D, Oandasan I. Interprofessionality as the field of interprofessional practice and interprofessional education: an emerging concept. J Interprof Care 2005;19 Suppl 1:8–20. 10.1080/13561820500081604 [DOI] [PubMed] [Google Scholar]
  • 11.Health Professions Networks Nursing & Midwifery, Human Resources for Health. . Framework for Action on Interprofessional Education & Collaborative Practice. Geneva: Switzerland: World Health Organization, 2010. [Google Scholar]
  • 12.editors.Mathias P, Thompson T, Øvretveit J. Interprofessional working for health and social care. Houndmills: Macmillan, 1997: 234. [Google Scholar]
  • 13.Zygmunt A, Asada Y, Burge F. Is team-based primary care associated with less access problems and self-reported unmet need in Canada? Int J Health Serv 2017;47:725–51. 10.1177/0020731415595547 [DOI] [PubMed] [Google Scholar]
  • 14.Khan S, McIntosh C, Sanmartin C. Primary health care teams and their impact on processes and outcomes of care. Ottawa, ON: Statistics Canada, 2008. [Google Scholar]
  • 15.Adam P, Brandenburg DL, Bremer KL, et al. Effects of team care of frequent attenders on patients and physicians. Fam Syst Health 2010;28:247–57. 10.1037/a0020944 [DOI] [PubMed] [Google Scholar]
  • 16.Ritchie C, Andersen R, Eng J, et al. Implementation of an interdisciplinary, team-based complex care support health care model at an academic medical center: impact on health care utilization and quality of life. PLoS One 2016;11:e0148096. 10.1371/journal.pone.0148096 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Bodenheimer T, Wagner EH, Grumbach K. Improving primary care for patients with chronic illness. JAMA 2002;288. [DOI] [PubMed] [Google Scholar]
  • 18.Edwards ST, Peterson K, Chan B. Effectiveness of intensive primary care interventions: a systematic review. J Gen Intern Med 2017. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Saint-Pierre C, Herskovic V, Sepulveda M. Multidisciplinary collaboration in primary care: a systematic review. Fam Pract 2017. [DOI] [PubMed] [Google Scholar]
  • 20.Green LA, Chang H-C, Markovitz AR, et al. The reduction in ED and hospital admissions in medical home practices is specific to primary Care-Sensitive chronic conditions. Health Serv Res 2018;53:1163-1179. 10.1111/1475-6773.12674 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Adaji A, Melin GJ, Campbell RL, et al. Patient-Centered medical home membership is associated with decreased hospital admissions for emergency department behavioral health patients. Popul Health Manag 2018;21:172-179. 10.1089/pop.2016.0189 [DOI] [PubMed] [Google Scholar]
  • 22.Chu L, Sood N, Tu M, et al. Reduction of emergency department use in people with disabilities. Am J Manag Care 2017;23:e409–15. [PubMed] [Google Scholar]
  • 23.Cuellar A, Helmchen LA, Gimm G, et al. The CareFirst patient-centered medical home program: cost and utilization effects in its first three years. J Gen Intern Med 2016;31:1382–8. 10.1007/s11606-016-3814-z [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Boult C, Green AF, Boult LB, et al. Successful models of comprehensive care for older adults with chronic conditions: evidence for the Institute of Medicine's "retooling for an aging America" report. J Am Geriatr Soc 2009;57:2328–37. 10.1111/j.1532-5415.2009.02571.x [DOI] [PubMed] [Google Scholar]
  • 25.McAlister FA, Bakal JA, Green L, et al. The effect of provider affiliation with a primary care network on emergency department visits and hospital admissions. CMAJ 2018;190:E276–84. 10.1503/cmaj.170385 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Wranik WD, Price S, Haydt SM, et al. Implications of interprofessional primary care team characteristics for health services and patient health outcomes: a systematic review with narrative synthesis. Health Policy 2019;123:550–63. 10.1016/j.healthpol.2019.03.015 [DOI] [PubMed] [Google Scholar]
  • 27.The Conference Board of Canada. . Final report: an external evaluation of the family health team (FHT) initiative. Ottawa: The Conference Board of Canada, 2014. [Google Scholar]
  • 28.Russell GM, Dahrouge S, Hogg W, et al. Managing chronic disease in Ontario primary care: the impact of organizational factors. Ann Fam Med 2009;7:309–18. 10.1370/afm.982 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Reiss-Brennan B, Brunisholz KD, Dredge C, et al. Association of integrated team-based care with health care quality, utilization, and cost. JAMA 2016;316:826–34. 10.1001/jama.2016.11232 [DOI] [PubMed] [Google Scholar]
  • 30.Tracy CS, Bell SH, Nickell LA. The impact clinic: innovative model of interprofessional primary care for elderly patients with complex health care needs. Can Fam Physician 2013;59:e148–55. [PMC free article] [PubMed] [Google Scholar]
  • 31.Harris MF, Advocat J, Crabtree BF, et al. Interprofessional teamwork innovations for primary health care practices and practitioners: evidence from a comparison of reform in three countries. J Multidiscip Healthc 2016;9:35–46. 10.2147/JMDH.S97371 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Hogg W, Lemelin J, Dahrouge S, et al. Randomized controlled trial of anticipatory and preventive multidisciplinary team care: for complex patients in a community-based primary care setting. Can Fam Physician 2009;55:e76–85. [PMC free article] [PubMed] [Google Scholar]
  • 33.Gater R, Goldberg D, Jackson G, et al. The care of patients with chronic schizophrenia: a comparison between two services. Psychol Med 1997;27:1325–36. 10.1017/S0033291797005631 [DOI] [PubMed] [Google Scholar]
  • 34.Gucciardi E, Espin S, Morganti A, et al. Implementing specialized diabetes teams in primary care in southern Ontario. Can J Diabetes 2015;39:467–77. 10.1016/j.jcjd.2015.07.006 [DOI] [PubMed] [Google Scholar]
  • 35.Lee L, Hillier LM, Gregg S. Partnerships for improving dementia care in primary care: extending access to primary care-based memory clinics in Ontario, Canada. Health Soc Care Community 2019;27:1574–85. 10.1111/hsc.12829 [DOI] [PubMed] [Google Scholar]
  • 36.Aggarwal M, Williams AP. Tinkering at the margins: evaluating the PACE and direction of primary care reform in Ontario, Canada. BMC Fam Pract 2019;20:128.. 10.1186/s12875-019-1014-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Marchildon GP, Hutchison B. Primary care in Ontario, Canada: new proposals after 15 years of reform. Health Policy 2016;120:732–8. 10.1016/j.healthpol.2016.04.010 [DOI] [PubMed] [Google Scholar]
  • 38.College of Family Physicians of Canada . A Vision for Canada: Family Practice The Patient’s Medical Home. The College of Family Physicians of Canada, 2011. [Google Scholar]
  • 39.Alliance for Healthier Communities . CHC fact sheet, 2017. Available: https://www.allianceon.org/chc-fact-sheet
  • 40.Sweetman A, Buckley G. Ontario’s Experiment with Primary Care Reform. The School of Public Policy SPP Research Papers 2014;7. [Google Scholar]
  • 41.Hutchison B, Glazier R. Ontario’s primary care reforms have transformed the local care landscape, but a plan is needed for ongoing improvement. Health Aff 2013;32:695–703. 10.1377/hlthaff.2012.1087 [DOI] [PubMed] [Google Scholar]
  • 42.Rudoler D, Laporte A, Barnsley J, et al. Paying for primary care: a cross-sectional analysis of cost and morbidity distributions across primary care payment models in Ontario Canada. Soc Sci Med 2015;124:18–28. 10.1016/j.socscimed.2014.11.001 [DOI] [PubMed] [Google Scholar]
  • 43.Kiran T, Kopp A, Glazier RH. Those left behind from voluntary medical home reforms in Ontario, Canada. Ann Fam Med 2016;14:517–25. 10.1370/afm.2000 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Glazier RH, Redelmeier DA. Building the patient-centered medical home in Ontario. JAMA 2010;303:2186–7. 10.1001/jama.2010.753 [DOI] [PubMed] [Google Scholar]
  • 45.Glazier RH, Zagorski BM, Rayner J. Comparison of primary care models in Ontario by demographics, case mix and emergency department use, 2008/09 to 2009/10. Toronto, Canada: Institute for Clinical Evaluative Sciences, 2012. [Google Scholar]
  • 46.Kralj B. Measuring “rurality” for purposes of health-care planning: an empirical measure for Ontario.. Ont Med Rev. 2000 Oct. [Google Scholar]
  • 47.Matheson FI. Ontario marginalization index: technical document. Toronto, ON: St. Michael’s Hospital 2011;2017. [Google Scholar]
  • 48.Johns Hopkins University Bloomberg School of Public Health . The Johns Hopkins ACG System: Excerpt from Technical Reference Guide Version 9.0. Health Services Research & Development Center, 2009 Dec.. [Google Scholar]
  • 49.Gershon AS, Wang C, Guan J, et al. Identifying patients with physician-diagnosed asthma in health administrative databases. Canadian Respiratory Journal 2009;16:183–8. 10.1155/2009/963098 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Schultz SE, Rothwell DM, Chen Z, et al. Identifying cases of congestive heart failure from administrative data: a validation study using primary care patient records. Chronic Dis Inj Can 2013;33:160–6. 10.24095/hpcdp.33.3.06 [DOI] [PubMed] [Google Scholar]
  • 51.Gershon AS, Wang C, Guan J, et al. Identifying individuals with physcian diagnosed COPD in health administrative databases. COPD 2009;6:388–94. 10.1080/15412550903140865 [DOI] [PubMed] [Google Scholar]
  • 52.Tu K, Campbell NR, Chen Z-L, et al. Accuracy of administrative databases in identifying patients with hypertension. Open Med 2007;1:e18–26. [PMC free article] [PubMed] [Google Scholar]
  • 53.Lipscombe LL, Hwee J, Webster L, et al. Identifying diabetes cases from administrative data: a population-based validation study. BMC Health Serv Res 2018;18:316. 10.1186/s12913-018-3148-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 54.Jaakkimainen RL, Bronskill SE, Tierney MC, et al. Identification of Physician-Diagnosed Alzheimer’s Disease and Related Dementias in Population-Based Administrative Data: A Validation Study Using Family Physicians’ Electronic Medical Records. JAD 2016;54:337–49. 10.3233/JAD-160105 [DOI] [PubMed] [Google Scholar]
  • 55.Kurdyak P, Lin E, Green D, et al. Validation of a population-based algorithm to detect chronic psychotic illness. Can J Psychiatry 2015;60:362–8. 10.1177/070674371506000805 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 56.Bullard MJ, Musgrave E, Warren D, et al. Revisions to the Canadian emergency department triage and acuity scale (CTAS) guidelines 2016. CJEM 2017;19:S18–27. 10.1017/cem.2017.365 [DOI] [PubMed] [Google Scholar]
  • 57.Cohen J. Statistical power analysis for the behavioural sciences. L. Erbaum Associates 1988. [Google Scholar]
  • 58.Gravetter FJ, Wallnau LB. Statistics for the behavioural sciences. Boston: Cengage Learning, 2013. [Google Scholar]
  • 59.Maher JM, Markey JC, Ebert-May D. The other half of the story: effect size analysis in quantitative research. CBE Life Sci Educ 2013;12:345–51. 10.1187/cbe.13-04-0082 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 60.Andersen RM. Revisiting the behavioral model and access to medical care: does it matter? J Health Soc Behav 1995;36:1–10. 10.2307/2137284 [DOI] [PubMed] [Google Scholar]
  • 61.Muldoon L, Dahrouge S, Russell G, et al. How many patients should a family physician have? factors to consider in answering a deceptively simple question. Hcpol 2012;7:26–34. 10.12927/hcpol.2013.22885 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary data

bmjopen-2022-065362supp001.pdf (178.5KB, pdf)

Reviewer comments
bmjopen-2022-065362.reviewer_comments.pdf (148.2KB, pdf)
Author's manuscript
bmjopen-2022-065362.draft_revisions.pdf (1.7MB, pdf)

Data Availability Statement

The data that support the findings of this study are approved for use by data stewards and accessed through a process managed by ICES (www.ices.on.ca/DAS (email: das@ices.on.ca). We are not permitted to share the data used in this analysis with other researchers. The full dataset creation plan and underlying analytic code are available from the authors on request, understanding that the computer programs may rely on coding templates or macros that are unique to this project and to individual data centres and are therefore either inaccessible or may require modification.

Articles from BMJ Open are provided here courtesy of BMJ Publishing Group

RESOURCES