Theme: 
Understanding our client populations and their health journeys
Source
Format: 
Presentation
Year: 
2022
Details: 

#Context

Information about the individuals’ Social Determinants of Health could help primary care (PC) providers prioritize and tailor care, but such data are not routinely collected in electronic medical records (EMRs). Neighbourhood level Measures of Vulnerability (MOVs) derived from census data may provide good proxy information. We studied three commonly used multi-dimensional Canadian neighbourhood MOVs to identify which have the strongest ability to discriminate health measures between most deprived and least deprived individuals.

#Methods

We linked the 2019 PC EMRs of 1.2 million Canadians housed at the Canadian Primary Care Sentinel Surveillance Network to the 2016 Canadian census data using postal code, to study the association between MOVs and multiple health measures capturing screening, prevalence, incidence, and management of chronic disease risk factors. Each MOV included multiple dimensions: Pampalon [2], Canadian Marginalisation Index (Can-Marg) [4], Canadian Index of multiple deprivation (CIMD) [4]. We derived age-sex standardised health measures for each MOV dimension and quintile (Q) and computed the relative health risk (RR) for Q5 (highest deprivation) in relation to Q1 (lowest deprivation) as a measure of their potential to discriminate. We report results on smoking in adults >18 years and diabetes incidence in last 3 years amongst those >40 years.

#Results

The overall prevalence of current smokers was 22%. Across all 10 MOV dimensions studied, the prevalence ranged from 14%-25% in Q1, and 21%-33% in Q5. The strongest discrimination was for Pampalon-material (RR: 2.1), CIMD-situational (RR: 2.4), and Can-Marg-material (RR:2.2). Overall diabetes incidence was 4.9% and ranged from 3.7%-4.8% in Q1 and 5%-7.1% in Q5. The strongest discrimination was for Pampalon-material (RR:2.1), CIMD-situational (RR:1.8), and Can-Marg-material (RR:1.9). Can-Marg-ethnic (immigration, minority), age related dependency, CIMD-economic, ethnic, and residential dimensions had RR<1.5. Overall, Pampalon was a better index of the three to discriminate health measures.

#Conclusions

Our preliminary results suggest that neighbourhood level MOVs are useful for identifying patients at higher risk of poor health measures. All 3 MOVs show good discrimination, with Pampalon being superior for some measures. We continue to analyse associations with other health measures and evaluate whether combining some dimensions within one MOV improves its ability to discriminate.

#Upon completion of this session, participants should be able to:

  1. To understand the strength of neighbourhood level measures of deprivation in predicting health measures
  2. To stimulate discussion on how these measures can be used to support clinical care