Concurrent session I: Access, quality, and cost

Friday, March 10, 11:00 am – 12:00 pm
Salon F
Chair: Rick Glazier

Demonstration of Starfield’s observations about relationship between primary care quality and health system cost at the level of primary care teams in Ontario

Presenting author: Carol Mulder, Association of Family Health Teams of Ontario

Co-author: Rick Glazier

Objective: Starfield observed that high quality primary care was related to lower per-capita healthcare costs. The regional/national nature of these observations make them less actionable for providers. This study examines the quality-cost relationship at primary care team level in Ontario to facilitate healthcare sustainability via high quality primary care.

Methods: All 184 members of the Association of Family Health Teams of Ontario (AFHTO) were invited to contribute quality and cost data to Data to Decisions (D2D), a performance measurement report, now in its 4th iteration. Consistent data extraction and compilation were enabled by team-level Quality Improvement Decision Support (QIDS) specialists. A composite measure of quality was calculated from patient experience, administrative and EMR data. Each component of the composite was weighted by its importance to patients in their relationship with their provider. It was compared to per-capita healthcare costs for primary care, hospitalization, diagnostic and consulting services and institutional care.

Results: Data were available for 120 teams (65%). The average quality score was 52.9 (S.D. =11.57, 25.2 – 85.3, compared to an approximated Ontario average of 30.6 based on publicly-reported aggregate data. The average per-capita cost (without institutional costs) was $2456.16 (S.D. = 504.65, 1234.68-4030.03). Linear regression analyses (controlling for patient panel size, Standardized Adjusted Clinical Group Morbidity Index (SAMI), rurality, teaching status and a surrogate measure of EMR maturity) showed that quality was negatively related to cost (R2 = 0.427). The relationship was stronger for urban than rural teams, suggesting other factors might be contributing to costs in different settings.

Conclusion: Starfield’s observation that high primary care quality is related to lower healthcare system costs applies to this setting. Factors affecting costs in rural/urban settings need further examination. Active participation of AFHTO members (and QIDS specialists in particular) in consistent, ongoing measurement via D2D was a crucial enabler of this study.

Is increasing physician supply the key to solving access/provision challenges in primary health care in British Columbia?

Presenting author: Lindsay Hedden, Centre for Clinical Epidemiology and Evaluation

Co-authors: Morris L. Barer, Kimberlyn M. McGrail, Michael R. Law, Ivy L. Bourgeault

Background/Introduction/Objective: Reports of access issues in primary health care are ubiquitous in British Columbia (BC), and despite significant financial and policy attention, 300,000 British Columbians are currently without, and are activity seeking, a primary care physician (PCP). At the same time, per capita physician supply has been increasing since 1986. We used a population-based, retrospective cohort study to examine the extent to which workforce feminization, aging, and secular period effects may explain the dual observations of increasing supply alongside frequently heard claims about difficulties in accessing primary care in BC.

Methods: We used fee-for-service (FFS) and alternative payments (non-FFS) billings for all PCPs in BC for the years 2005/6-2011/12. We modeled the trend in per-physician remuneration, patient contact counts, and practice size over the study period using longitudinal multivariate mixed effects linear models. Models included gender and age, and adjusted for training and practice locations, and the proportion of payments from non-FFS sources. We simulated differences in population-level spending on physician services, patient contacts, and per-physician practice size attributable to shifts in workforce demographics and secular period effects, accounting for the overall growth in the physician population.

Results: We found limited change in per-physician remuneration over the study period, after taking out the effects of changes in fee levels; however, total patient contact counts and per-PCP practice sizes declined by 10% (111,577 total contacts/year) and 13% (38 patients/physician/year) respectively. Although workforce feminization, and to a lesser extent aging, contributed to these declines, the primary driver is a broad secular trend toward reduced clinical activity over time. Removing the effect of this trend while allowing for demographic shifts would have resulted in an additional 2.03 million patient contacts by 2011/12, the equivalent of one additional physician visit for half of BC’s population.

Conclusion: Despite overall growth in PCP supply in BC, shrinking per-physician levels of service delivery and practice size have meant that the increased supply has not kept pace with demand. Workforce feminization and aging both contributed to the reduction in activity over time, but their effects were dwarfed by the secular decline in activity levels observed among all PCPs.

Access to primary health care and the hospitalization rates for ambulatory care sensitive conditions in First Nations communities in Manitoba

Presenting author: Naser Ibrahim, University of Manitoba

Co-authors: Josée Lavoie, Wanda Philips, Stephanie Sinclair, Grace Kyoon-Achan, Kathi Avey Kinew, Alan Katz

The objective of this study was to assess the performance of models of PHC delivered in FN communities in Manitoba, using hospitalization rates for Ambulatory Care Sensitive Conditions (ACSC) as a key indicator. The iPHIT study is a partnership between Nanaadawewigamig, the First Nations Health and Social Secretariat of Manitoba (FNHSSM), researchers from the University of Manitoba, Manitoba Health, and 63 FN Communities from Manitoba.

We used administrative claims data (1986-2013) housed at the Manitoba Centre for Health Policy which includes all Manitoba residents insured under the Manitoba Health Services Insurance Plan. Data was analyzed using generalized estimating equation logistic regression, controlled for age, sex, and socioeconomic status to find the relationship between hospitalization rates for ACSC and models of PHC in FN communities.

The study shows that the hospitalizations rates for acute and chronic conditions have decreased over time in FN communities, but the rates are increasing for mental health related admissions. FN communities served by Nursing Stations had lower admission rates for all ACSC compared to the other models of care. Nevertheless, Manitoba FN people living in FN communities have significantly higher hospitalization rates compared to other Manitoba residents.

In conclusion, local access to a broader complement of responsive PHC is associated with lower admission rates for ACSC. Improving access to PHC in FN communities, to an extent comparable to a Nursing Station model could lead to a reduction in avoidable hospitalization rates in these communities.

Development of a provincial evaluation framework for A GP for Me: Measuring collective impact

Presenting author: Sarah Fielden, Doctors of BC

Co-authors: Andi Cuddington, Kathi Hendry, Garey Mazowita, John Hamilton, and the GPSC Evaluation Working Group

Context: Evidence suggests that patients who have an ongoing relationship and are “attached” to a primary care provider have better health outcomes. A GP for Me is a province-wide initiative funded by the Government of BC and Doctors of BC with the goals of helping BC residents find a regular family physician and strengthening the physician/patient relationship.

Problem/issue: A provincial evaluation framework was required to help guide and assess the initiative. This presentation describes the process of its development. It utilized a two-pronged approach that addressed the need for province-wide as well as community-based measurement in partnership with the Divisions of Family Practice (DofFP) across BC. The evaluation framework was based on the five core conditions of the Collective Impact evaluation approach* and included learnings of three prototype Divisions’ projects (2010-2013).

Intervention: A GP for Me supports DofFP-led projects emerging from community-based assessment of need. Interventions are diverse throughout the province but include themes of: physician recruitment and retention; ‘medical-home’ development; and optimizing physician practice office efficiency. Creation of the evaluation framework development was informed by the US Center for Disease Control model and included stakeholder engagement and an overarching logic model.

Measurement: Based on reviewing Canadian primary care evaluation models, domains were selected to categorize outcomes: health system performance; quality of care; governance and policy; and partnership. Key outcomes were informed by goals of the initiative as well as the US Institute of Health Improvement’s Triple Aim. Quality indicators were selected based on three main criteria: scientific soundness; potential feasibility; and capturing an important performance aspect.

Challenges & lessons learned: Challenges of the process included: timeliness, data access and reporting, and the multiplicity of strategies proposed to address community-level needs. The primary lessons learned relate to beginning the evaluation process and engaging with relevant partners as early as possible and aligning measurement with existing evidence.