Table 2.
Subgroup analyses.
| Subgroups | Categories (Number of studies) | Proportion (%) | 95% CI(%) | Events | Sample size | I2 (%) | Q (P) |
|---|---|---|---|---|---|---|---|
| Community | Yes (11) | 45.1 | 37.7–52.8 | 4,119 | 10,696 | 97.8 | 6.32 (0.012) |
| No (13) | 58.2 | 51.4–64.6 | 1,788 | 3,224 | 92.2 | ||
| Geographical region | North (6) | 53.3 | 40.0–66.2 | 2,567 | 6,242 | 98.0 | 0.09 (0.76) |
| South (17) | 51.0 | 43.6–58.3 | 3,243 | 7,539 | 97.1 | ||
| Publication language | Chinese (22) | 53.6 | 46.7–60.4 | 3,899 | 8,491 | 97.0 | 5.80 (0.016) |
| English (2) | 40.0 | 32.0–53.8 | 2,009 | 5,429 | 94.2 | ||
| Age groupa | ≥63.7 (11) | 57.1 | 46.5–67.0 | 3,117 | 7,510 | 97.8 | 1.34 (0.24) |
| <63.7 (11) | 48.9 | 39.9–57.9 | 2,373 | 5,513 | 97.3 | ||
| Survey yeara | 2013–2017 (10) | 46.9 | 36.6–57.4 | 1,952 | 4,961 | 97.7 | 2.28 (0.13) |
| 2007–2012 (10) | 57.9 | 48.2–67.0 | 3,488 | 7,992 | 97.8 | ||
| Sample sizea | ≥232 (12) | 42.9 | 36.0–50.2 | 5,002 | 12,396 | 98.1 | 11.8 (0.001) |
| <232 (12) | 62.3 | 54.1–69.9 | 903 | 1,524 | 89.7 | ||
| Sampling method | Probability (8) | 41.5 | 33.9–49.5 | 3,950 | 10,150 | 98.0 | 8.03 (0.005) |
| Non-probability (16) | 58.2 | 49.8–66.1 | 1,955 | 3,770 | 95.6 | ||
| Cut-off of CPSQI | >5 (4) | 51.7 | 39.2–63.8 | 2,319 | 5,942 | 50.4 | 0.52 (0.91) |
| >6 (3) | 60.5 | 33.8–82.2 | 461 | 810 | 97.5 | ||
| >7 (15) | 51.1 | 42.6–59.5 | 2,796 | 6,487 | 97.7 | ||
| >10 (2) | 50.1 | 42.1–58.1 | 329 | 681 | 97.3 |
Bolded values: P < 0.05; Q: Cochran’s Q;
a: Continuous variables, such as age, survey year and sample size, were dichotomized using median splitting methods in the subgroup analyses. CPSQI: Chinese version of the Pittsburgh Sleep Quality Index; Probability sampling method: cluster sampling; multistage sampling; random sampling; stratified sampling; Non-probability sampling method; convenience and consecutive sampling.