Table 2.
Summary of subgroup analysis for the effect of inpatient hypoglycaemia on length of stay and mortality
| Category | Studies included (N) | Resulta | I² (%) |
|---|---|---|---|
| Subgroup analysis of the effect of hypoglycaemia on length of stay | |||
| All studies | 9 | 4.08 (2.36 to 5.79) | 99 |
| Removal of papers rated as poor quality | 4 | 3.59 (0.80 to 7.62) | 99 |
| Removal of non‐cohort studies | 7 | 4.15 (2.11 to 6.19) | 99 |
| Removal of studies with imputed standard deviation | 6 | 3.62 (2.09 to 5.14) | 98 |
| General ward location only | 6 | 3.24 (1.01 to 5.47) | 97 |
| Hospital location not specified only | 4 | 5.08 (2.14 to 8.02) | 100 |
| Inclusion of non‐serious hypoglycaemia definitions only | 5 | 4.37 (2.13 to 6.61) | 98 |
| Inclusion of serious hypoglycaemia definitions only | 2 | 2.87 (‐0.36 to 6.10) | 98 |
| General ward areas only and removal of outliers | 8 | 2.14 (1.30 to 2.99) | 70 |
| General ward location and removal of papers rated as poor quality | 2 | 1.63 (0.87 to 2.40) | 67 |
| Hospital location not specified, and removal of papers rated as poor quality | 2 | 5.58 (3.55 to 7.62) | 97 |
| Subgroup analysis of the effect of hypoglycaemia on mortality | |||
| All studies | 11 | 1.69 (1.40 to 2.03) | 92 |
| Removal of papers rated as poor quality | 5 | 1.81 (1.36 to 2.42) | 90 |
| Removal of non‐cohort studies | 4 | 1.62 (1.16 to 2.26) | 95 |
| Removal of studies with imputed data | 9 | 1.57 (1.29 to 1.91) | 87 |
| In‐hospital mortality only | 7 | 2.09 (1.64 to 2.67) | 94 |
| In‐hospital mortality and non‐serious hypoglycaemia definition | 5 | 2.15 (1.98 to 2.33) | 0 |
| Intensive care unit mortality only (all used serious hypo definitions) | 2 | 0.75 (0.49 to 1.16) | 0 |
| 90‐day and post discharge mortality only | 2 | 1.26 (1.08 to 1.47) | 0 |
a
Mean difference in length of stay in days or risk ratio (95% confidence interval).