Validation and Application of a Needs-Based Complexity Case-Mix System for Community-Based Primary Health Care Planning and Research

Abstract

Background: Canadian provinces are transforming their primary health care (PHC) systems by developing networks of resources at community and practice levels. This leaves provinces with the challenge of determining how to meaningfully measure the PHC needs of communities and practices. Currently, methods for systematically measuring PHC needs at community and practice levels are inadequate for PHC planning and are not routinely used. Objectives: The objectives of this research were to (1) operationalize and evaluate a complexity case-mix system to measure variation in PHC needs across practices and communities, and (2) apply the system to two areas of resource alignment relevant for PHC planning a) comparing the case-mix of PHC needs of patients unattached to a PHC provider with the attached population, and b) assessing if there are systematic differences in the case-mix of PHC needs between groups with different commuting patterns for PHC. Approach and Results: The setting of this research was Nova Scotia (NS), Canada. We adapted a previously developed case-mix approach for PHC, leveraging administrative data to classify Nova Scotians into a set of needs-based segments ranging from low to high complexity. We evaluated the complexity case-mix system against a set of criteria including consistency with research on what makes PHC needs complex, utility for PHC decision makers, and segment associations with healthcare use. Segments were consistent with research on what makes PHC needs complex, and decision makers found the NS complexity case-mix system to be useful, identifying multiple areas for application including evaluating PHC performance. Segments were further found to be predictive of healthcare costs and primary care costs. We then applied the case-mix system to the unattached patient centralized waitlist in NS and found that patients who are unattached to PHC have less complex needs than the attached population. We further rostered patients to providers and found variability in commuting for PHC by community, region, and need segment, with those who commute for PHC having less complex needs compared to those who do not commute. Impact: The results of this embedded research are enhancing the capacity to design and iteratively improve community PHC systems.