Table 5.
Summary of Findings Table.
| Certainty assessment | No of patients | Effect | Certainty | Importance | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| No of studies | Study design | Risk of bias | Inconsistency | Indirectness | Imprecision | Other considerations | Telemedicine | Usual practice | Relative (95% CI) | Absolute (95% CI) | ||
| HbA1c% | ||||||||||||
| 243 | Randomized trials | Serious a | Serious b | Not serious | Not serious | None | 29 671 | 27 465 | − | MD 0.415 lower (0.482 lower to 0.348 lower) | ⨁⨁◯◯ LOW | IMPORTANT |
Abbreviations: CI, confidence interval; HbA1c%, glycated hemoglobin; MD, mean difference.
The size of the review implies that no single study contributes dominant weights in the meta-analysis. Indeed, study weights range from 0.26 to 0.77. The correlation between effect size, that is, the MD, and risk of bias across studies was low and insignificant (Spearman ρ = −0.06, P = .3961). However, only 17% of the studies were evaluated to have a low risk of bias. Consequently, the risk of bias was downgraded to one level and assessed as serious.
Effect size point estimates vary widely between studies, with significant effects favoring both alternatives. I2 was high, both with and without controlling for statistically significant study covariates simultaneously in the meta-regression (I2 = 87.8% and I2 = 93%).