175 6 Cochrane review on integrated disease management for COPD Therefore, results for these groups should be interpreted carefully (Analysis 1.14). 2.2.2 Subgroup analysis based on dominant component of intervention Four studies (102 participants) reporting on the 6MWD had some kind of exercise training as their dominant component (Engstrom 1999; Fernandez 2009; Gottlieb 2011; Güell 2000). In six studies, structural follow-up was considered the dominant component (Kalter-Leibovici 2018; Kessler 2018; Ko 2016; Littlejohns 1991; Vasilopoulou 2017; Zhang 2020). One study provided individualised education as the dominant component (Wakabayashi 2011), and another study included self-management as the dominant component (Jimenez-Reguera 2020). Therefore, these could not be pooled. A test for subgroup di erences showed a statistically signi cant di erence (Chi² = 10.56, df = 4, P = 0.03; Analysis 1.14). Subgroup analysis for exercise training as the dominant component showed that the 6MWD improved by 68.21 metres (95% CI 44.75 to 91.68; I² = 3%). This e ect was almost twice the MCID of 35 metres. Also, studies with telemonitoring as the dominant component showed a large improvement of 59.94 metres (95% CI 42.59 to 77.29; I² = 32%). Studies with structural follow-up as the dominant component showed statistically signi cant di erences in favour of IDM (MD 35.14, 95% CI 2.83 to 67.45). However, heterogeneity remained substantial. 2.2.3 Subgroup analysis based on region of study Three studies reporting on 6MWD with medium-term follow-up were performed in Northwestern Europe (Engstrom 1999; Gottlieb 2011; Littlejohns 1991), ve in Southern Europe (Fernandez 2009; Güell 2000; Jimenez-Reguera 2020; Kessler 2018; Vasilopoulou 2017), four in East Asia (Ko 2016; Wakabayashi 2011; Wang 2017; Zwar 2016), and one in Western Asia (Kalter-Leibovici 2018). A test for subgroup di erences indicated statistically signi cant di erences in e ect between subgroups (Chi² = 19.09, df = 3, P = 0.00003). Pooling of studies performed in Northwestern Europe showed no statistically signi cant di erence between IDM and control (MD 18.18, 95% CI -7.87 to 44.24; I² = 4%). A statistically signi cant di erence was found for the Southern Europe subgroup (MD of 61.73) and the East Asia subgroup (MD of 42.67). Pooling indicated considerable heterogeneity in the subgroup of studies from Southern Europe (I² = 68%) and East Asia (I² = 90%); results for these subgroups should therefore be interpreted carefully (Analysis 1.16). 2.3 Functional exercise capacity - long-term Six studies on 7288 participants published long-term results on the 6MWD (Gottlieb 2011; Güell 2000; Kalter- Leibovici 2018; Lou 2015; van Wetering 2010; Zhang 2020). The MD was 48.83 metres in favour of IDM and was of statistically and clinically signi cant relevance (95% CI 16.37 to 80.49; I² = 90%) (Analysis 1.13). Sensitivity analysis could not explain heterogeneity and showed a smaller non-statistically signi cant mean di erence (MD 36.4; I² = 94%; Analysis 1.13) noted by a wide CI (95% CI -6.43.97 to 79.24).
RkJQdWJsaXNoZXIy MTk4NDMw