Using administrative health data to inform health service planning for specialist cancer care in Nova Scotia, Canada

Abstract

Introduction
Research has demonstrated that primary care providers can safely and effectively provide follow-up care after a person has received treatment for cancer. Yet, discharge to primary care after cancer treatment is variable, despite the fact that cancer systems are challenged to provide follow-up care given constrained cancer specialist resources.

Objectives and Approach
To inform cancer system planning, we examined (1) cancer centre routine follow-up (CC-FUP) care for prevalent cancer types and (2) changes in CC-FUP over time. From the Nova Scotia Cancer Registry, we identified all persons diagnosed in Nova Scotia, Canada, with an invasive breast, colorectal, gynecological, or prostate cancer between 01/01/2006 and 31/12/2013. We linked this dataset to cancer centre/clinic data and identified a non-metastatic cancer survivor cohort (n=12,267). Descriptive statistics were computed to describe patterns of care. Negative binomial regression was used to examine changes over time for both CC-FUP and all cancer centre visits, adjusting for other covariates.

Results
Nearly half of survivors (48.4\%) had at least one CC-FUP visit, which varied by disease site (range: 30.2-62.4\%). Variation existed across providers, with six oncologists providing 34.7\% of the CC-FUP visits to the study population. Year of diagnosis was associated with receipt of CC-FUP care, with each successive calendar year associated with an 8\% increase in CC-FUP visits (IRR=1.08, 95\%CI=1.07-1.10). Similarly, each successive calendar year was associated with a 14\% increase in all cancer centre visits (IRR=1.14, 95\%CI=1.13-1.15). Results were shared with cancer system decision-makers at regular intervals to inform ongoing analyses.

Conclusion/Implications
Both the number of CC-FUP visits and all visits increased over time, with the latter at a greater rate. The increases were much higher than assumed by cancer system decision makers (2\% increase per year) for resource planning, demonstrating the value of population-based administrative data to informing health service planning.

Introduction

Research has demonstrated that primary care providers can safely and effectively provide follow-up care after a person has received treatment for cancer. Yet, discharge to primary care after cancer treatment is variable, despite the fact that cancer systems are challenged to provide follow-up care given constrained cancer specialist resources.

Objectives and Approach

To inform cancer system planning, we examined (1) cancer centre routine follow-up (CC-FUP) care for prevalent cancer types and (2) changes in CC-FUP over time. From the Nova Scotia Cancer Registry, we identified all persons diagnosed in Nova Scotia, Canada, with an invasive breast, colorectal, gynecological, or prostate cancer between 01/01/2006 and 31/12/2013. We linked this dataset to cancer centre/clinic data and identified a non-metastatic cancer survivor cohort (n=12,267). Descriptive statistics were computed to describe patterns of care. Negative binomial regression was used to examine changes over time for both CC-FUP and all cancer centre visits, adjusting for other covariates.

Results

Nearly half of survivors (48.4%) had at least one CC-FUP visit, which varied by disease site (range: 30.2-62.4%). Variation existed across providers, with six oncologists providing 34.7% of the CC-FUP visits to the study population. Year of diagnosis was associated with receipt of CC-FUP care, with each successive calendar year associated with an 8% increase in CC-FUP visits (IRR=1.08, 95%CI=1.07-1.10). Similarly, each successive calendar year was associated with a 14% increase in all cancer centre visits (IRR=1.14, 95%CI=1.13-1.15). Results were shared with cancer system decision-makers at regular intervals to inform ongoing analyses.

Conclusion/Implications

Both the number of CC-FUP visits and all visits increased over time, with the latter at a greater rate. The increases were much higher than assumed by cancer system decision makers (2% increase per year) for resource planning, demonstrating the value of population-based administrative data to informing health service planning.