188 6 Chapter 6 large heterogeneity between studies. For outcomes of functional and maximum exercise capacity, we downgraded the certainty of evidence owing to large heterogeneity that may be caused by an intervention-speci c e ect (i.e. IDM programmes with exercise as the dominant component showed more positive results for exercise capacity). We deemed the quality of evidence for respiratory-related hospital admissions as high. We downgraded one level for all-cause hospital admissions because of considerable heterogeneity and inconsistency in direction of e ect. We also downgraded the certainty of evidence for outcomes of hospital days per patient and ED visits due to inconsistency in e ects. Potential biases in the review process Several methodological strengths minimised the risk of bias in this review. As de nitions of IDM are still under debate, we strictly determined the inclusion criteria for an IDM programme a priori and published this in our review protocol (Kruis 2011). Our de nition was derived from de nitions published in the literature (Peytremann-Bridevaux 2009; Schrijvers 2009). Overall, researchers reported on “multiple interventions, designed to manage chronic conditions, with a focus on a multidisciplinary approach”. Furthermore, these de nitions suggest that IDM interventions should “focus on maximum clinical outcome, regardless of treatment setting(s) or typical reimbursement patterns”. As a result, we chose to include all interventions, independent of treatment setting, and to keep our de nition as simple as possible, to be easily understandable for readers and easy to use when readers check on all relevant literature. Therefore, we restricted included trials to multi-component, multi-disciplinary programmes of at least 12 weeks’ duration. Furthermore, we performed comprehensive searches to identify possible studies, leading to identi cation of more than 10,000 potentially relevant abstracts. Subsequently, three di erent assessors assessed the abstracts. We reached consensus on all included studies. Final decisions of course are open to interpretation or criticism. However, we have applied a systematic approach to including and excluding studies in this review, have followed the criteria pre-speci ed in the protocol, and have used robust methods for data collection and ‘Risk of bias’ assessment. We were able to retrieve additional data from 17 study authors but did not receive a response from eight authors despite multiple reminders. This may have introduced bias. Another limitation of this review is inconsistent reporting in the included studies, in terms of adjusting for baseline di erences. We decided on a conservative approach, using unadjusted mean di erences for all randomised controlled trials (RCTs) and adjusted only values corrected for clustering e ects, to overcome inconsistency between study authors’ corrections. Inconsistency in reporting also resulted in the need for computing standard deviations of the mean change using appropriate analysis methods. Last, there may have been large heterogeneity in control groups, resulting from country-speci c healthcare systems and COPD regulations for COPD treatment (i.e. reimbursements). Because the level of detail in reporting usual care varied greatly between studies (possibly also due to journal guidelines), we decided it was more informative to further investigate di erences between regions instead of di erences between types of usual care, as was performed in the previous version of
RkJQdWJsaXNoZXIy MTk4NDMw