We used a knowledge translation framework to implement asthma guidelines in a region and in a practice setting that did not have the requisite administrative, clinical, and human resources for effective guideline implementation. We engineered health system change by creating a regional administrative infrastructure, by creating asthma specific knowledge tools and by implementing an interdisciplinary care model. We demonstrated the functionality of the program by implementing six evidence-based best practices in a majority of patients. Finally, we demonstrated improvements in asthma-related health outcomes including asthma symptoms, urgent healthcare utilization and absenteeism that were sustained over time. This project is a case study demonstrating how quality improvement/knowledge translation can be usefully guided by a conceptual framework and result in positive health and health system outcomes.
Despite the identified knowledge-to-practice gap internationally, there are relatively few studies evaluating the implementation of evidence-based asthma care in a primary care setting [
22–
26]. The pediatric asthma management program, Easy Breathing, has been implemented by primary care physicians in large health maintenance organizations and urban Medicaid populations in the USA [
22,
24,
25]. A proactive asthma program was implemented in a pediatric private practice population in Australia and a provincial program for children and adults in a community healthcentre population in Canada [
23,
26]. Similar to our study, these studies demonstrate improvements in urgent health service use, absenteeism and symptom control. A novel aspect of our study is implementation in private practices that operate outside of a large healthcare organization, where there is no infrastructure to support the development, implementation and evaluation of quality improvement initiatives. We created electronic knowledge tools to support program portability, scalability, resource sharing and program evaluation.
Implementing a program in a community setting without access to a large administrative data set required that we use patient self-report to measure health services utilization. Although self-reported health service utilization is common in the chronic disease management literature, and although self-report has been validated against administrative databases in other settings, utilization in this study was not validated against an administrative data set [
27–
29]. We did however utilize several methods to increase the accuracy of the self-reported data including: measuring acute events that are easily understood by the subject (exacerbations requiring urgent care), conducting structured interviews with qualified staff, utilizing the shortest meaningful reporting interval and repeating the same inquiry consistently on all assessments [
30].
When tailoring this project to meet regional needs, we selected a design that would maximize physician participation. This design did not include a randomly assigned control group. While we are not aware of any systematic bias affecting our outcomes, without a control group we cannot exclude the possibility that bias influenced our results. Several factors support a cause-and-effect conclusion between our intervention and improved health outcomes. Our cohort was recruited from low acuity settings and we demonstrated asthma control levels that aligned with published surveys. This suggests our cohort was a valid representation of primary care and mitigates the risk that regression to the mean bias enriched our results [
8,
10,
11]. There was strong internal consistency in our outcomes over time; early improvements were sustained for almost 2 years and across all health outcome measures. Finally, there was a notable external consistency between the effect sizes in our study and those of comparable controlled efficacy trials [
1–
4,
22–
26].
Despite measures to facilitate follow-up appointments, the number of patients lost to follow-up was high, reflecting the real challenges of implementation in a community setting. We evaluated our data in two ways to assess for the impact of incomplete follow-up. First, we compared the baseline characteristics of the patients returning for follow-up with those who did not. The groups were not different on a majority of parameters; however, patients who returned for follow-up were older, or had a lower FEV1, or were more likely to be on a combination product than patients who did not return. While the retention of patients with parameters indicative of more severe asthma is clinically desirable, it may have increased the magnitude of the reported improvements. Another concern is that lost patients may over-represent patients who did not improve or who deteriorated thereby enriching the reported outcomes. We conducted a sensitivity analysis to address this concern directly. In the most likely scenario of the analysis, where we assumed that every patient lost to follow-up remained unchanged from baseline, all reported health outcome improvements were confirmed.
In Canada, the mean cost of caring for patients in an ambulatory asthma clinic population has been estimated at $2550/person/year [
31]. The development, implementation and evaluation of our project cost $501/patient which is comparable to the $664/patient cost of a diabetes self-management program in our jurisdiction [
32]. The ongoing primary care asthma program that followed our demonstration project is maintained at a cost of $290/patient. Estimated program cost savings can be calculated based on the post-intervention reduction in the number of urgent health visits and days absent multiplied by the average cost. Applying this simple model over the 2-year study interval, we estimate cost savings of $166 880.00 ($321/patent) on urgent health services and $145 656.00 ($281/patient) on absenteeism [
33–
34]. Program savings offset costs in a ratio of 2.1:1 ($602/$290), suggesting that our program may be cost effective.
Closing the gap between research knowledge and evidence-based care requires a systematic approach that is adaptable to a range of healthcare settings. A knowledge-translation framework can guide multi-level organizational change, facilitate asthma guideline implementation and improve health outcomes, with modest program expenditures in community primary care practices.