The influence of variations in implementation of a COPD disease management program across primary care teams on costs and effects

Background: The implementation of disease management (DM) is expected to differ between practices. Variations in the implementation can be caused by differences in translation from the intervention to the specific practice caused by different starting levels of DM and different barriers and facilitators influencing the implementation. The implementation variations will probably have an impact on costs and effects of DM. Failing to recognize these differences in implementation may lead to inappropriate conclusions about the cost-effectiveness of DM. Therefore this study focused on the relationship between implementation of the intervention and costs and effects. Methods: This study combined qualitative and quantitative research methods to enable a comprehensive evaluation of the implementation of the RECODE intervention. In the qualitative part the implementation variations between practices were explored by conducting interviews with practice nurses (or GPs). This part also paid attention to barriers and facilitators of implementing DM. The outcomes of the qualitative research were used to estimate scores on 18 indicators that represent the implementation of aspects of the RECODE intervention. In addition to the separate indicators, subtotal scores for the implementation of CCM elements and total implementation scores were calculated. The quantitative part of this study used these implementation variables to investigate the influence of implementation variations on costs and effects of DM. Results: The results of this study confirm the expectation that the RECODE intervention gets tailored to the specific practice and therefore the implementation varies between practices. The variations emerged because of different starting levels of DM and barriers and facilitators encountered during the implementation. The implementation variations (represented by the implementation variables) could not explain most of the differences between practices in changes in costs and effects between baseline and 18 months after the start of the RECODE study. The few significant influences that were found had an unexpected direction; implementing (more aspects of) DM resulted in poorer disease-specific quality of life (higher CCQ scores). Some findings were consistent with the expectations. First, the use of funding for physiotherapy improved dyspnea (lower MRC scores) of patients. Second, the implementation of individual treatment plans increased the generic quality of life (higher EQ-5D scores) in the subgroup of patients with more COPD symptoms (MRC>2). Conclusion: The implementation variations this study identified did not explain most of the differences between the practices in cost and effects. Therefore it is desirable to elaborate on the findings in this study and further explore the heterogeneity between practices and its influence on costs and effects of DM. This information could support future cost-effectiveness studies to give more accurate conclusions about the cost-effectiveness of DM.