The efficacy of using continuous glucose monitoring as a behaviour change tool in populations with and without diabetes: a systematic review and meta-analysis of randomised controlled trials

Table 3 Sensitivity analysis of HbA1c regression

Analysis	Studies	Mean difference			I²
Analysis	Studies	Mean	95% CI	p	Mean	95% CI
Main analysis	23	-0.28	-0.42, -0.15	< 0.001	88.5	84.0, 91.7
Influential cases removed¹	20	-0.29	-0.42, -0.17	< 0.001	78.0	66.5, 85.6
High risk of bias cases removed²	16	-0.29	-0.46, -0.12	< 0.001	91.2	87.3, 93.9
Conflict of interest cases removed³	13	-0.32	-0.47, -0.17	< 0.001	85.4	76.6, 90.8
Study duration ≥ 12 weeks⁴	20	-0.24	-0.36, -0.11	< 0.001	87.9	82.7, 91.5

¹ Three RCTs removed as outliers: Schembre et al., Zhang et al., Haak et al. ² Seven high risk of bias cases: Alfadhli et al., Aronson et al., Cosson et al., Guo et al., Meisenhelder-Smith, Price et al., Yeoh et al. ³Eleven RCTs declared a conflict of interest: Allen et al., Aronson et al., Chekima et al., Choe et al., Cosson et al., Furler et al., Haak et al., Jospe et al., Lee et al., Murphy et al., Price et al., Wada et al., Yoo et al. ⁴ Three RCTs were less than 12 weeks in duration: Allen et al., Chekima et al., Guo et al. HbA1c: glycated haemoglobin

ISSN: 1479-5868