193 6 Cochrane review on integrated disease management for COPD to allow measurement of the true intervention e ect. Speci cally, consistent reporting on exacerbation outcomes and on severity of exacerbations may overcome the di culties we encountered in this review, for which we found a myriad of exacerbation de nitions. Researchers are encouraged to use recent Global initiative for Chronic Obstructive Lung Disease (GOLD) guidelines to provide unambiguous de nitions of disease severity and to evaluate e ects of IDM programmes on mild, moderate, and severe exacerbations (GOLD 2020). Subgroup analyses undertaken as part of this update stimulate new questions in relation to IDM and its contextual embedding. Di erences in subgroups based on the dominant intervention component call for further research to identify which intervention component, or which combination of components, is most e ective in IDM programmes, and for which patient groups. Similarly, the context-speci c e ects we observed in the subgroup analysis suggest that the country in which the IDM programme is embedded and the level of usual care it is compared to greatly impact the magnitude of e ect. This still means that the individual components of IDM programmes are important and will improve patient outcomes, as shown in this review. However, the contrast of a new IDM programme versus usual care becomes smaller when usual care itself already routinely contains several of the components. Other factors that remain uncertain are the optimal duration and intensity of the intervention and the combination of healthcare providers involved. These questions can be examined in a meta-regression analysis, which could shed light on the contribution of each individual factor or combination of factors to observed treatment e ects. Although the observed e ect of - 3.89 on the SGRQ did not reach the proposed MCID of -4 points for medium-term follow-up, there could be a proportion of patients in the intervention group that does exceed the 4 points of improvement. These so called ‘responders’ would clinically bene t more from IDM than from usual care. In our review, only Bourbeau 2003 reported the proportion of people who improved by 4 points or more on the SGRQ. Hence, we echo Cates 2015 and urge trialists to also report, besides the mean di erence, the spread of individual responses to the intervention or treatment. This information can be used for more complete assessment of clinical importance and helps to reveal the population bene t. Last, process-related outcomes raised issues that require consideration beyond this current review. For example, special attention should be given to evaluating the actual implementation of IDM programmes in existing healthcare structures, which should include outcomes related to patient satisfaction, feasibility, programme compliance, and assessment of personal and contextual determinants of implementation and treatment e ects. Pragmatic, real-life RCTs including both clinical and process-related outcomes and qualitative assessment with long-term follow-up are needed to evaluate IDM programmes as comprehensive packages in routine primary and secondary care practice. As part of this, cost-e ectiveness remains an important outcome, to allow for reimbursement and to inform health policy development and clinical guidance.
RkJQdWJsaXNoZXIy MTk4NDMw