Prevalence of sleep disturbance among chinese college students: an updated meta-analysis

Abstract

Background

Sleep disturbance is common among college students. However, the prevalence and associated factors among Chinese college students require an updated synthesis. This study aimed to estimate the pooled prevalence of sleep disturbance in this population and to examine potential sources of between-study variation.

Methods

A systematic search was conducted in four Chinese databases (CNKI, Wanfang, VIP, and Sinomed) and five international databases (PubMed, EMBASE, Web of Science, Cochrane Library, and PsycINFO) from inception to December 22, 2025. The study protocol was pre-registered in PROSPERO (CRD420251270413). Cross-sectional studies reporting sleep disturbance among Chinese college students assessed using the Pittsburgh Sleep Quality Index were included, with no restrictions on cut-off thresholds. Recall timeframes included the past month, 1–2 weeks, or several months. Random-effects meta-analysis was used to pool prevalence estimates with 95% confidence intervals, and a 95% prediction interval was additionally reported for the overall pooled estimate to reflect between-study heterogeneity. Statistical heterogeneity was assessed using the I² statistic. Subgroup analyses were performed to explore methodological and population-level factors, and between-group differences were tested using chi-square statistics.

Results

A total of 232 studies involving 495,641 undergraduate students were included. The pooled prevalence of sleep disturbance was 26.4% (95% confidence interval: 24.6% to 28.3%, 95% prediction interval: 7.4% to 59.1%), with substantial heterogeneity (I² = 99.57%). Methodological factors were significantly associated with prevalence estimates, particularly the Pittsburgh Sleep Quality Index cut-off score and the recall timeframe (both P < 0.001). Prevalence was highest during the COVID-19 pandemic (33.5%), compared with the pre-pandemic period (24.1%) and the post-pandemic period (21.0%) (P = 0.001). Higher pooled estimates were observed in more recent publication periods, increasing from 22.9% in 2014 and earlier to 28.7% in 2020 and later (P = 0.018). No statistically significant differences were identified across gender, academic year, major, region, or only-child status. Higher prevalence was observed among smokers, drinkers, and students reporting poorer family economic status, although these differences did not reach statistical significance.

Conclusions

Sleep disturbance, as defined by the Pittsburgh Sleep Quality Index, is prevalent among Chinese college students. Estimates should be interpreted cautiously because of extreme heterogeneity and reliance on self-reported, predominantly cross-sectional data. The findings underscore the public health relevance of sleep health on campuses and support continued monitoring and health promotion, as well as more standardized measurement in future studies.

Clinical trial number

Not applicable.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12888-026-07997-z.

Introduction

Sleep disorders are when people have trouble falling asleep or maintaining asleep, their sleep is disturbed, and they can’t function well during the day even when they have enough time and a good place to sleep [1]. There are many types of sleep disorders, including insomnia, sleep-related breathing disorders, central disorders of hypersomnolence, circadian rhythm sleep-wake disorders, parasomnias, sleep-related movement disorders, and other sleep disorders [2]. Sleep constitutes around one-third of human existence and is essential for the restoration of physiological processes and the maintenance of mental and physical homeostasis. Sleep disorders have emerged as a prevalent health concern in modern society, particularly among college students [3, 4]. The rising prevalence of sleep problems in this demographic has been attributed to a complex array of stressors, including academic pressures, work-related anxiety, interpersonal conflicts, and emotional difficulties [5]. Symptoms of sleep disorders in students include tiredness in lectures, nocturnal insomnia, and even neurasthenia, which collectively are associated with their physical and mental well-being, as well as academic performance [6–8]. Reports indicate that college students worldwide experience sleep problems to varying degrees [9, 10]. Given the unique contextual circumstances in China, the epidemiology of sleep disorders among Chinese college students warrants an updated assessment. Rapid societal changes in China over the past decade, including digitalization, intensified academic and employment competition, and the COVID-19 pandemic, may plausibly be associated with higher levels of sleep disturbance among college students through several interconnected mechanisms. For instance, increased evening screen time and constant connectivity have been associated with delayed sleep onset and disrupted circadian rhythms, which are in turn linked to sleep disturbance [11–13]. Concurrently, elevated academic demands and uncertainty may exacerbate physiological stress and cognitive-emotional arousal, which are often linked to insomnia and non-restorative sleep [14]. Additionally, sleep duration may be reduced when study obligations, social media use, or nighttime activities displace time otherwise dedicated to rest, leading to irregular sleep patterns. These factors are likely compounded by campus environments and broader societal pressures [15], and their effects may be amplified during periods of major disruption, such as the COVID-19 pandemic [16, 17]. Evidence from Chinese college student populations similarly indicates that poorer sleep quality is associated with worse academic functioning and mental health. For example, sleep disturbance, as measured by the Pittsburgh Sleep Quality Index (PSQI), has been associated with lower academic engagement among Chinese college students [18]. Moreover, recent large-scale studies in China have shown that sleep disturbance and poor sleep quality are positively associated with depressive and anxiety symptoms, underscoring the close interplay between sleep and psychological well-being in Chinese college settings [19, 20]. These associations are likely bidirectional: psychological distress may contribute to poorer sleep, whereas sleep disturbance may, in turn, aggravate depressive and anxiety symptoms by disrupting emotion regulation and daytime functioning [21–24]. Collectively, these findings suggest that sleep disturbance is not only a prevalent health concern but also a salient correlate of academic and mental health outcomes among Chinese college students. Several prior meta-analyses have summarized sleep problems among Chinese college students. However, most were conducted before more recent evidence became available and provided limited evaluation of methodological sources of heterogeneity, such as PSQI cut-off values and recall timeframes. The present study extends prior work by updating the evidence to December 22, 2025, incorporating substantially more studies, reporting prediction intervals alongside pooled prevalence under extreme heterogeneity, and conducting subgroup analyses to evaluate temporal patterns and methodological moderators. This updated synthesis provides a comprehensive descriptive summary of PSQI-defined sleep disturbance among Chinese college students, offering a useful reference for understanding the magnitude and distribution of this health concern.

Methods

This systematic review and meta-analysis adhered to the standards outlined in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA 2020) [25] and the Meta-Analysis of Observational Studies in Epidemiology (MOOSE) [26]. The study protocol was pre-registered in the PROSPERO database (CRD420251270413).

Literature search strategy

A systematic search was conducted across Chinese databases (CNKI, Wanfang, VIP and Sinomed) and foreign databases (PubMed, EMBASE, Web of Science, Cochrane Library and PsycINFO). The time range encompassed all eligible studies from the inception of each database up to December 22, 2025. The search strategy involved entering the following terms: P((“China” OR “Chinese”) AND (“university student” OR “college student” OR “undergraduate student” OR “adolescents” OR “young adults”)), O(“dyssomnia” OR “sleep disorders” OR “sleep disturbance” OR “sleep problem” OR “sleep dysfunction” OR “insomnia” OR “sleep quality”), and S(“prevalence” OR “epidemiology survey” OR “cross-sectional study”). In PubMed, both MeSH terms and free-text keywords were used in combination. In other databases, these terms were entered in free-text format. To minimize omissions, the reference lists of retrieved articles were also screened.

Inclusion and exclusion criteria

Studies are included researches meeting the following criteria: (1) Participants were full-time undergraduate students enrolled in Chinese higher education institutions (including mainland of China, Hong Kong, Macau, and Taiwan regions); (2) Sample size ≥ 500 individuals; (3) The PSQI was used as the assessment tool; (4) The outcome measure was the prevalence of sleep disturbance based on PSQI scores; (5) The study design was cross-sectional. The exclusion criteria were as follows: (1) Studies that included master’s students, doctoral students, vocational college students, transfer students, part-time students, or international students were excluded; studies including mixed student populations were also excluded unless data specifically for college students could be extracted separately; (2) Reviews, conference proceedings, dissertations, or studies with non-cross-sectional designs; (3) Studies with missing data, duplicate publications, or low methodological quality, defined as a score of 3 or less on the Agency for Healthcare Research and Quality (AHRQ) cross-sectional study quality assessment scale, were excluded; (4) Studies for which the full text was unavailable; (5) Studies published in languages other than English or Chinese; (6) Studies using substantially modified versions of the PSQI were excluded unless the authors clearly stated equivalence to, or validation against, the original PSQI or a validated Chinese version. Two researchers (ZH and GZW) independently screened the literature, extracted data, and cross-checked findings. Disagreements were resolved through discussion or by consultation with a third researcher (PQH). Studies were excluded based on the title when clearly irrelevant. Abstracts and full texts were then reviewed to determine eligibility for inclusion.

Data extraction

Two reviewers (ZH and GZW) independently extracted the following data: (1) Basic study information: research title, first author, publication year, study period, study region, etc.; (2) General characteristics of study subjects: age, grade level, ethnicity, lifestyle habits, family background, only child status, male proportion, medical student status, valid sample size, etc.; (3) Key elements for assessing risk of bias: response rate, sampling method, AHRQ quality assessment, etc.; (4) Outcome definitions and measurement parameters: criteria for sleep disturbance related to PSQI, timeframe for sleep disturbance, etc.; (5) Outcome measures: prevalence of sleep disturbance across study populations. Any discrepancies between the two reviewers were resolved through discussion or consultation with a third reviewer (PQH).

Quality appraisal

Two reviewers (ZH and GZW) assessed the risk of bias in included studies according to items recommended by the U.S. Agency for Healthcare Research and Quality [27, 28]. Responses were recorded as “Yes,” “No,” or “Unclear,” with “Yes” scored as 1 point, “No” or “Unclear” scored 0 points. A total score of 0–3 indicated low quality, 4–7 indicated moderate quality, and 8–11 indicated high-quality literature. Disagreements between reviewers were resolved through discussion with a third reviewer (PQH).

Statistical analysis

Stata 15.0 was used for all analyses. Between-study heterogeneity was assessed using Cochran’s Q test, with P < 0.05 indicating statistically significant heterogeneity, and was quantified using the I² statistic. Pooled prevalence estimates and 95% confidence intervals (CIs) were synthesized using a random-effects model for proportions implemented with Stata’s metaprop command. The Freeman–Tukey double arcsine variance-stabilizing transformation was applied, and pooled results were back-transformed to the prevalence scale. Prespecified subgroup analyses were conducted to explore potential sources of heterogeneity, including publication period, sex, PSQI cut-off values, recall timeframe, and other study-level characteristics when available. Subgroup differences were assessed using Cochran’s Q test for between-subgroup heterogeneity. Meta-regression was not performed because several key study-level moderators were highly imbalanced, sparsely represented, or available for only a limited number of studies. This would have reduced the analyzable dataset and produced estimates that were unstable and difficult to interpret. Given the extreme heterogeneity (I² > 99%), the pooled prevalence was interpreted as a descriptive summary rather than a single generalizable parameter. Accordingly, a 95% prediction interval (PI) was additionally reported to reflect the expected range of true prevalence in future comparable studies. Prediction intervals were constructed under a random-effects model on the logit scale using the pooled logit estimate, its standard error, and the between-study variance τ², and were then back-transformed to the prevalence scale using the inverse-logit function. Sensitivity analyses were performed using a leave-one-out approach. Publication bias was evaluated using Begg’s and Egger’s tests. When small-study effects were suggested, the trim-and-fill method was applied as a sensitivity analysis. Our analytic approach and reporting of heterogeneity were informed by recent systematic reviews and meta-analyses of sleep-related outcomes, as illustrated by Gupta et al. [29].

Results

Literature screening results

An initial search of databases and other sources identified 7,578 records. After removing duplicates, the titles and abstracts of 5,641 records were screened. Subsequently, 819 full-text articles were assessed for eligibility. Of these, 587 were excluded after full-text review, and the reasons for exclusion are presented in Fig. 1. Ultimately, 232 studies were included in the analysis [30–261]:60 in English and 172 in Chinese, encompassing 495,641 undergraduate students. The basic characteristics and quality appraisal of the 232 studies are summarized in Table 1.

Fig. 1.

PRISMA flow chart for the study selection process

Table 1.

Basic information and quality scores of the literature

No.	Author, year	Area	Region	Study Period	Pandemic Context	Sampling method	Medical student	Grade	Age	Proportion of males (%)	Effective sample	Response rate (%)	PSQI Cut-off	Prevalence (%)	Time frame	AHRQ Score
1	Bai et al.(2017)	Liaoning	N	2015.11	PRE	C; R; S	NR	1–4	NR	21.13	2496	96.93	≥ 8	18.11	NR	5
2	Cao et al.(2009)	Xinjiang	N	NR	NR	R	NR	1–4	21.32 ± 2.56	50.07	1362	85.12	≥ 8	24.60	NR	4
3	Chang et al.(2016)	Taiwan	S	2014.3-2014.4	PRE	C	NR	NR	19.23 ± 1.10	40.89	1230	94.98	≥ 6	58.62	NR	8
4	Chang et al.(2018)	Tianjin	N	2018.3	PRE	R	Both	1–4	20.12 ± 1.14	58.28	827	92.09	≥ 8	21.64	NR	5
5	Chen et al.(2017a)	Anhui	S	2016.9	PRE	R	Yes	NR	19.72 ± 1.43	49.30	1441	92.61	≥ 6	35.53	NR	8
6	Chen et al.(2017b)	Hebei	N	NR	NR	C; R; S	NR	1–4	20.03 ± 1.89	48.35	817	96.12	NR	32.68	NR	4
7	Chen et al.(2017c)	NR	NR	NR	NR	C; S	Yes	1–4	NR	40.21	873	89.08	≥ 8	33.45	NR	6
8	Chen et al.(2020)	Hunan	S	NR	NR	C; R	Yes	1–3	NR	NR	869	96.56	≥ 8	42.58	LM	5
9	Chen et al.(2023)	Anhui	S	NR	NR	CS	NR	NR	NR	42.75	1165	86.30	≥ 8	29.79	NR	6
10	Chen et al.(2024a)	Guizhou/Shanxi/Shandong/Hubei/Anhui/Zhejiang	Both	2023.3-2023.4	DUR	CS	NR	1–4	18.40 ± 1.50	39.21	2099	79.06	≥ 5	62.27	LM	8
11	Chen et al.(2024b)	Chongqing	S	2023.3-2023.5	DUR	C	NR	NR	NR	NR	770	96.00	≥ 8	33.77	NR	10
12	Chen et al.(2025)	Hebei	N	NR	NR	CS	NR	NR	23.40 ± 3.60	50.89	617	95.07	≥ 5	51.05	LM	5
13	Cheng et al.(2018)	Anhui	S	NR	NR	C; R	Yes	4–5	NR	35.77	548	91.33	≥ 8	25.73	LM	6
14	Chong et al.(2025)	Jilin/Shandong/Anhui/Gansu	Both	2024.11	POST	CS	NR	1–4	18.99 ± 1.31	40.67	1618	93.33	NR	12.68	NR	8
15	Chu et al.(2018)	Anhui	S	NR	NR	C; R	Both	NR	NR	56.58	866	98.41	≥ 8	19.63	LM	6
16	Cui et al.(2020)	Anhui	S	2018.2-2018.5	PRE	C; R; S	Yes	NR	NR	47.01	1019	82.44	≥ 8	27.10	LM	8
17	Deng et al.(2018)	Guangxi	S	2015.4-2015.6	PRE	R	NR	1–3	20.16 ± 1.35	52.36	4614	98.21	≥ 8	9.82	NR	6
18	Ding et al.(2022)	Hebei	N	2020.12-2021.1	DUR	R	Yes	1–5	NR	35.53	5140	85.77	≥ 6	31.87	NR	8
19	Ding et al.(2024)	Heilongjiang	N	2023.10	POST	CS	Both	1–5	NR	47.93	820	92.66	≥ 8	52.93	LM	8
20	Du et al.(2022)	Shandong	N	NR	NR	C; R	Yes	1–3	NR	38.32	775	91.18	≥ 8	23.70	LM	6
21	Fan et al.(2014)	Shanghai	S	NR	NR	Cen.	No	1–3	NR	52.33	772	59.71	≥ 8	13.60	NR	4
22	Fan et al.(2016)	Guangdong	S	2014.9-2015.9	PRE	C; R	Both	NR	21.06 ± 4.95	51.10	5202	94.58	≥ 8	44.39	NR	7
23	Fan et al.(2017)	Fujian	S	NR	NR	C; R; S	Both	1–3	NR	37.14	1960	98.00	≥ 8	37.14	LM	6
24	Fan et al.(2020)	Zhejiang	S	2020.2	DUR	CS	NR	1–4	NR	46.78	932	93.20	≥ 8	56.87	NR	9
25	Fang et al.(2019)	Guangdong	S	2017.6-2017.9	PRE	CS	Both	NR	21.0 ± 0.90	29.46	2064	97.27	≥ 9	9.35	NR	8
26	Feng et al.(2012)	Xinjiang	N	2012.3	PRE	C	Both	1–4	NR	33.96	1219	98.31	≥ 8	48.40	NR	4
27	Feng et al.(2014)	Wuhan	S	2011.11-2011.12	PRE		Both	1	18.9 ± 0.90	57.41	1106	92.20	≥ 6	17.70	LM	10
28	Gan et al.(2018)	Jiangsu	S	2016.3-2016.6	PRE	C; S	Yes	1–3	20.52 ± 1.38	16.32	582	97.00	≥ 8	32.47	LM	5
29	Gao et al.(2021)	Xizang	S	2019.10-2019.11	PRE	C; S	Both	2–3	20.22 ± 1.12	45.11	1288	88.04	≥ 8	20.50	NR	6
30	Geng et al.(2014)	Jiangsu	N	2012.10	PRE	M; R; S	Both	1–4	NR	48.24	1194	97.87	≥ 8	23.79	NR	7
31	Guo et al.(2016a)	Henan	N	NR	NR	R	No	1–4	20.86 ± 1.33	23.45	631	90.14	≥ 9	14.30	NR	7
32	Guo et al.(2016b)	Ningxia Hui Autonomous Region	N	NR	NR	C; M; S	Both	1–4	NR	28.19	777	97.12	≥ 8	36.55	NR	5
33	He et al.(2020)	Guangdong	S	NR	NR	CS	Both	1–4	20.85 ± 1.23	26.67	525	99.06	≥ 8	22.10	LM	4
34	He et al.(2025)	Hubei	S	2023.4	DUR	R; S	Both	1–4	19.37 ± 1.50	38.53	2668	66.92	NR	31.37	NR	7
35	Hou et al.(2020)	Guangdong	S	NR	NR	R; S	Both	1–4	NR	47.49	838	83.80	≥ 8	25.66	LM	7
36	Hu et al.(2018)	Chongqing	S	NR	NR	S	NR	1–3	19.70 ± 1.30	28.96	891	95.29	≥ 8	33.67	LM	5
37	Hu et al.(2019)	Heilongjiang	N	NR	NR	R; S	Yes	1–3	18.97 ± 0.81	24.74	780	94.43	≥ 8	41.03	NR	4
38	Hu et al.(2023a)	Hangzhou	S	NR	NR	R	Yes	NR	NR	NR	617	100.00	≥ 6	0.39	NR	5
39	Hu et al.(2023b)	Anhui/Jiangxi/Shanghai/Xinjiang	Both	2021.9-2021.12	DUR	C; S	NR	1–4	21.12 ± 1.17	43.10	6363	97.64	≥ 6	72.52	LM	8
40	Hu et al.(2023c)	Jiangsu	S	2021.10	DUR	C; R; S	Both	1–3	18.97 ± 1.10	0.00	1349	97.75	≥ 6.5	27.06	NR	8
41	Hu et al.(2024a)	Jiangsu	S	2022.10-2022.11	DUR	C; CS	Both	1–4	NR	30.03	8458	98.38	≥ 6	39.29	LM	9
42	Hu et al.(2024b)	Jiangsu	S	2022.10-2022.11	DUR	CS	Both	1–5	18.83 ± 1.05	30.03	8457	98.49	≥ 8	26.64	LM	10
43	Huang et al.(2007)	Zhejiang	S	NR	NR	C; R	Both	1–3	NR	39.75	1024	94.81	≥ 8	33.98	NR	5
44	Huang et al.(2013)	Shanghai	S	2011.5-2011.6	PRE	C; R; S	Both	1–5	20.60 ± 1.40	47.76	1051	95.55	≥ 8	33.78	NR	8
45	Huang et al.(2017)	Henan	N	NR	NR	R	Both	1–5	NR	31.61	987	96.76	≥ 7	28.98	NR	5
46	Huang et al.(2018)	Henan	N	NR	NR	R	Both	1–5	NR	31.17	972	98.58	≥ 7	15.85	NR	4
47	Huang et al.(2022)	Neimenggu/Guizhou/Beijing	Both	2020.9-2020.10	DUR	CS	Both	NR	NR	NR	711	84.04	≥ 8	14.91	LM	8
48	Huang et al.(2024a)	Xinjiang	N	2022.10-2023.5	DUR	CS	NR	NR	NR	38.46	1469	97.93	≥ 8	55.41	NR	5
49	Huang et al.(2024b)	Sichuan	S	NR	NR	C: CS	NR	NR	NR	25.38	9408	93.62	NR	19.33	NR	5
50	Huo et al.(2010)	Inner Mongolia Autonomous Region	N	NR	NR	C; R	NR	1–4	22.17 ± 1.48	38.16	773	95.67	≥ 8	11.13	NR	4
51	Ji et al.(2020)	Hubei/ Gansu	Both	2020.4	DUR	CS	Yes	1–4	NR	24.47	515	97.54	≥ 8	17.28	NR	6
52	Ji et al.(2022)	Hainan	S	NR	NR	C	Yes	NR	20.17 ± 1.54	43.28	989	98.90	≥ 8	22.75	LM	4
53	Ji et al.(2023)	Fujian	S	2022.4	DUR	C; R	NR	NR	19.0 ± 1.40	31.86	5379	76.92	≥ 8	29.63	LM	7
54	Jiang et al.(2019)	Hubei	S	2018.4	PRE	CS	NR	1–4	20.30 ± 1.80	34.52	733	93.73	≥ 8	21.28	NR	6
55	Jiang et al.(2021)	Anhui	S	NR	NR	Cen.	NR	NR	NR	NR	2688	96.34	NR	12.10	NR	7
56	Jiao et al.(2021)	Beijing/Tianjin/Hubei	Both	2020.3	DUR	R; S	NR	NR	20.30 ± 1.30	50.43	2808	93.51	≥ 8	12.39	NR	10
57	Jin et al.(2014)	Zhejiang	S	2012.11	PRE	C; CS	Both	1–3	19.73 ± 1.03	21.02	1632	54.64	NR	22.73	LM	8
58	Jin et al.(2018)	Anhui	S	NR	NR	C; R	Both	1–4	20.74 ± 1.31	50.94	1333	88.87	≥ 7	25.73	FM	8
59	Kang et al.(2016)	Guangxi	S	NR	NR	C; S	Yes	1–3	20.11 ± 1.30	0.00	1267	98.22	≥ 8	34.49	NR	4
60	Ke et al.(2018)	Hubei	S	NR	NR	M; R	Both	1–3	NR	34.56	570	NR	≥ 8	17.72	NR	7
61	Lei et al.(2024)	Shanghai/Jiangxi/Hubei/Shanxi	Both	2023.10-2023.12	POST	CS	NR	NR	NR	40.39	14,379	84.58	≥ 6	16.32	LM	9
62	Li et al.(2009)	Heilongjiang	N	NR	NR	C	Yes	1–5	21.0 ± 1.0	39.41	4808	90.50	≥ 9	25.96	NR	4
63	Li et al.(2015a)	Ningxia Hui Autonomous Region	N	NR	NR	C; R; S	Yes	1–4	NR	NR	869	98.75	≥ 8	42.58	LM	6
64	Li et al.(2015b)	Sichuan	S	2013.5	PRE	C; S	Both	1–3	NR	27.98	729	91.13	≥ 8	14.40	LM	6
65	Li et al.(2016a)	Liaoning	N	2014.11	PRE	C; S	NR	1–4	20.40 ± 1.10	49.10	1053	96.52	≥ 8	14.72	NR	5
66	Li et al.(2016b)	Liaoning	N	2014.11	PRE	C; R	Yes	NR	NR	30.39	770	97.10	≥ 8	15.84	NR	5
67	Li et al.(2018)	East China/ Central China/Southwest China/Northwest China	Both	2017.3-2017.4	PRE	C; R; S	Yes	1–3	NR	27.21	6534	91.49	≥ 8	10.33	LM	9
68	Li et al.(2019a)	Anhui	S	2017.10-2017.11	PRE	C; R; S	Yes	1–4	19.80 ± 1.38	31.18	1437	89.81	≥ 8	14.06	NR	7
69	Li et al.(2019b)	Guangdong	S	NR	NR	C; R; S	Yes	1–3	NR	35.44	951	79.25	≥ 8	19.35	NR	6
70	Li et al.(2020a)	Anhui/ Jiangxi	S	2019.4-2019.5	PRE	C; S; CS	Both	NR	18.80 ± 1.20	38.06	1135	96.27	≥ 8	13.30	NR	7
71	Li et al.(2020b)	Jiangxi/ Liaoning	Both	2019.10-2019.11	PRE	C; CS	NR	NR	20.10 ± 1.60	43.64	1164	92.53	≥ 8	15.98	LM	8
72	Li et al.(2020c)	Jilin	N	2016	PRE	C; S	Both	1–5	19.76 ± 1.45	52.67	6284	83.79	≥ 6	31.05	LM	10
73	Li et al.(2020d)	Shanghai	S	NR	NR	R	Yes	1–4	NR	32.99	982	98.20	NR	21.18	LM	4
74	Li et al.(2020e)	NR	NR	2018.4	PRE	C	NR	1–2	19.48 ± 0.93	15.89	598	94.92	≥ 8	49.50	LM	7
75	Li et al.(2021a)	Hubei	S	2019.10-2019.12	PRE	C; R	Both	1–4	NR	43.26	712	98.89	≥ 8	19.38	LM	8
76	Li et al.(2021b)	Jiangsu	S	2019.10-2019.11	PRE	C	Both	1–5	NR	40.26	3366	90.97	≥ 8	15.72	NR	7
77	Li et al.(2022)	Hunan	S	2020.10	DUR	CS	NR	1–4	20.0 ± 1.19	32.94	2347	NR	≥ 6	48.57	NR	5
78	Li et al.(2023)	Heilongjiang/Jiangxi/Liaoning/Shanxi	Both	2021.9-2021.12	DUR	CS	NR	NR	19.0 ± 1.70	49.68	1872	80.62	≥ 8	14.00	NR	8
79	Li et al.(2024)	Shanghai/Jiangxi/Hubei	S	2023.8-2023.10	POST	CS	NR	1–4	19.62 ± 1.37	35.08	13,920	85.93	≥ 6	16.23	NR	8
80	Li et al.(2025a)	Anhui/Jiangxi	S	2019.4-2019.5	PRE	C; S	Both	NR	18.70 ± 1.20	31.56	903	76.59	≥ 8	11.63	LM	8
81	Li et al.(2025b)	Anhui	S	2023.3-2023.5	DUR	R; S	Yes	1–5	NR	44.91	550	NR	≥ 11	37.30	NR	7
82	Li et al.(2025c)	Guizhou	S	2024.1-2024.6	POST	CS	Both	NR	NR	28.13	686	NR	≥ 16	24.78	NR	5
83	Liao et al.(2007)	Chongqing	S	NR	NR	C; M; S	Both	NR	21.50 ± 2.52	48.29	878	80.85	≥ 8	13.78	NR	5
84	Liao et al.(2016)	Jilin	N	2014.8-2014.9	PRE	C	Both	1–5	20.72 ± 1.59	46.86	956	83.79	≥ 8	13.81	NR	7
85	Lin et al.(2015)	Shanxi	N	NR	NR	R	NR	NR	20.76 ± 1.18	50.79	1012	97.68	≥ 8	8.10	NR	8
86	Lin et al.(2016)	Zhejiang	S	2014.2-2014.5	PRE	S	No	NR	NR	0.00	554	98.23	≥ 9	27.44	LM	6
87	Lin et al.(2019)	Shanxi	N	2018.4-2018.6	PRE	R; S	NR	1–4	NR	50.76	1984	99.20	≥ 8	8.06	NR	10
88	Lin et al.(2025)	Jiangxi/Hubei/Shanghai/Shanxi	Both	2023.10-2023.11	POST	CS	NR	1–4	NR	39.64	14,767	89.50	≥ 6	15.26	LM	8
89	Liu et al.(1994)	Shandong	N	NR	NR	R	Yes	NR	20.85 ± 1.81	57.50	560	93.33	≥ 8	13.93	NR	5
90	Liu et al.(2009)	Jilin	N	2008	PRE	CS	NR	NR	NR	49.55	783	97.88	NR	13.03	LM	6
91	Liu et al.(2011)	Guizhou	S	2010.3-2010.9	PRE	C	Both	1–4	20.52 ± 1.52	40.40	2126	94.53	≥ 8	27.38	NR	7
92	Liu et al.(2016)	Anhui	S	NR	NR	R	Yes	1–5	21.14 ± 1.53	40.25	646	80.75	≥ 8	24.30	LM	4
93	Liu et al.(2019)	Anhui	S	NR	NR	C	Yes	1–3	19.65 ± 1.45	38.20	1034	94.00	≥ 8	9.18	LM	6
94	Liu et al.(2021)	Guangdong	S	2019.12-2020.1	PRE	CS	Both	1–5	20.54 ± 1.08	38.74	573	99.65	≥ 8	35.43	LM	6
95	Liu et al.(2023a)	Jiangxi	S	2021.10-2021.12	DUR	R	NR	NR	19.0 ± 1.03	50.08	1258	92.84	≥ 8	16.53	LM	5
96	Liu et al.(2023b)	Anhui	S	NR	NR	CS	Both	1–5	20.15 ± 1.43	47.83	690	94.52	≥ 8	28.70	NR	4
97	Liu et al.(2024)	Hunan/Hubei/Guizhou	S	2023.1-2023.3	DUR	C; R; S	NR	1–4	NR	49.02	1524	95.07	≥ 8	39.20	NR	6
98	Liu et al.(2025)	Zhejiang	S	2023.11-2023.12	POST	CS	No	1	18.50 ± 0.79	37.71	1408	NR	≥ 8	26.85	LM	9
99	Lu et al.(2018)	NR	NR	2016	PRE	Cen.	Both	NR	NR	70.92	3081	77.03	≥ 8	48.10	NR	6
100	Luo et al.(2025)	Fujian	S	NR	NR	M; R; S	No	NR	20.22 ± 1.52	31.36	2057	NR	≥ 8	59.16	NR	7
101	Luo et al.(2026)	Guangdong	S	2023.9-2023.10	POST	CS	NR	NR	NR	41.93	1307	72.61	≥ 6	37.80	NR	9
102	Lv et al.(2025)	Hubei	S	2021.9-2022.1	DUR	CS	Yes	1–5	NR	NR	2893	81.40	≥ 6	70.40	LM	10
103	Ma et al.(2020)	Northeast China	N	2018.4-2019.12	PRE	R; S	NR	1–3	19.30 ± 0.98	29.61	1550	97.48	≥ 6	39.42	LM	10
104	Mei et al.(2022)	Jilin	N	NR	NR	R; S	Both	1–5	20.70 ± 1.60	47.36	946	82.91	≥ 5	33.19	LM	8
105	Meng et al.(2021)	Liaoning	N	2016.12-2017.1	PRE	C	Both	NR	20.66 ± 1.66	28.11	4234	94.09	≥ 8	12.94	NR	8
106	Niu et al.(2017)	Ningxia Hui Autonomous Region	N	2016.3-2016.4	PRE	C; R; S	Both	1–4	20.98 ± 1.47	37.83	2263	94.29	≥ 8	30.62	NR	7
107	Pan et al.(2021)	Guangdong	S	2018.2-2018.4	PRE	C; R; S	Both	1–2	20.50 ± 1.733	35.30	730	96.05	≥ 7	23.56	NR	7
108	Pan et al.(2022)	Shanxi	N	2022.1	DUR	CS	NR	NR	20.11 ± 1.71	51.20	1412	77.97	≥ 8	33.29	NR	6
109	Peng et al.(2014)	Guangdong	S	2011.10	PRE	C; R; S	NR	NR	20.19 ± 1.69	47.53	1296	92.57	≥ 8	28.40	NR	5
110	Qian et al.(2010)	Heilongjiang	N	NR	NR	C; R; S	Both	NR	NR	44.08	583	97.17	≥ 8	33.28	NR	5
111	Qin et al.(2022)	Sichuan	S	2022.11-2022.12	DUR	CS	Both	NR	20.13 ± 1.53	32.20	1379	98.64	≥ 8	20.30	NR	8
112	Qiu et al.(2024)	Qinghai/Tibet Autonomous Region	S	NR	NR	R	NR	1–4	NR	49.27	3026	91.95	≥ 6	74.29	LM	8
113	Shen et al.(2015)	Jiangsu	S	2014.5	PRE	C	Both	1–3	20.50 ± 1.10	56.41	913	73.04	≥ 8	17.09	LM	6
114	Shi et al.(2005)	Guangdong	S	NR	NR	C; R	Both	NR	20.74 ± 1.27	48.62	1302	93.00	≥ 8	28.65	NR	5
115	Shi et al.(2013)	Jiangxi	S	NR	NR	C; R	Both	1–3	NR	43.76	1145	95.42	≥ 8	16.42	LM	7
116	Song et al.(2017)	Liaoning	N	2015.11	PRE	C; S	NR	1–3	19.70 ± 1.30	33.25	1224	93.29	≥ 8	14.22	NR	6
117	Song et al.(2023)	Guangxi	S	2021.11-2021.12	DUR	C	NR	1–4	20.30 ± 1.10	54.50	967	80.58	≥ 8	24.61	NR	6
118	Su et al.(2012)	Fujian	S	2010.5	PRE	CS	NR	NR	NR	37.19	562	93.67	≥ 7	22.80	NR	6
119	Su et al.(2021)	Jiangxi	S	2018.5	PRE	C; S; P	No	1–3	NR	48.54	2610	98.79	≥ 8	16.97	NR	7
120	Sun et al.(2009)	Xuzhou	N	NR	NR	R	Both	1–4	NR	51.61	589	98.17	≥ 8	66.89	NR	4
121	Sun et al.(2015)	Jiangsu	N	2015.5	PRE	C; R; S	Both	1–3	NR	57.89	900	100.00	≥ 8	14.56	LM	8
122	Sun et al.(2021)	Liaoning	N	2020.9-2020.10	DUR	C	Both	1–5	NR	51.10	1452	96.80	≥ 8	36.09	NR	9
123	Teng et al.(2017)	Guangdong	S	2015.11	PRE	C; CS	Yes	1	18.40 ± 0.90	35.14	2792	95.78	≥ 8	3.87	NR	7
124	Tian et al.(2010)	Hebei	N	NR	NR	C; R	Both	1–4	NR	22.53	688	98.29	≥ 8	33.72	NR	6
125	Tong et al.(2010)	Jiangsu	S	NR	NR	C; R	NR	1–4	20.45 ± 1.51	39.02	715	88.27	≥ 8	15.80	LM	5
126	Tong et al.(2023)	Anhui	S	2020.11-2020.12	DUR	C; M; S; CS	Both	1–3	19.14 ± 1.11	43.18	4768	94.17	≥ 8	18.23	NR	9
127	Wang et al.(2008)	Hainan	S	2006.10	PRE	C; R	Yes	1–5	21.07 ± 1.84	46.10	551	91.83	≥ 8	21.42	LM	5
128	Wang et al.(2012)	Hubei	S	NR	NR	R	NR	NR	20.99 ± 1.20	49.75	796	94.76	≥ 8	17.84	NR	4
129	Wang et al.(2014a)	Hainan	S	2013.5-2013.6	PRE	C; M; R	Both	1–5	20.60 ± 1.40	34.69	2341	91.80	≥ 8	23.32	LM	6
130	Wang et al.(2014b)	Jiangsu	S	NR	NR	C	Both	1–3	20.50 ± 1.20	43.85	1131	87.00	≥ 8	29.53	LM	5
131	Wang et al.(2016a)	Inner Mongolia Autonomous Region	N	2013	PRE	Cen.	Yes	1–5	NR	27.28	6085	NR	≥ 6	27.84	LM	6
132	Wang et al.(2016b)	NR	NR	2015.5-2015.6	PRE	C	No	1–4	20.10 ± 1.30	58.25	527	90.86	≥ 5	71.54	NR	5
133	Wang et al.(2016c)	Guangdong	S	NR	NR	C; R; S	Both	1–5	NR	41.92	1312	93.71	≥ 8	33.61	LM	4
134	Wang et al.(2018)	Sichuan	S	2017.7-2017.8	PRE	C; R; S	NR	NR	19.61 ± 1.19	40.92	501	98.24	≥ 8	18.56	LM	7
135	Wang et al.(2019a)	Anhui	S	2014.12-2015.2	PRE	C; M; R	NR	NR	19.50 ± 1.33	35.39	1328	94.86	≥ 6	38.18	LM	8
136	Wang et al.(2019b)	Jilin	N	2016	PRE	C; S	Both	1–5	NR	52.67	6284	83.79	≥ 6	33.53	NR	7
137	Wang et al.(2020)	Anhui	S	2018.9-2018.11	PRE	R; S	Yes	1–3	18.80 ± 1.18	41.52	3738	98.37	≥ 8	30.12	NR	6
138	Wang et al. (2022a)	NR	NR	2019.6-2019.8	PRE	SSnow	NR	1–4	20.32 ± 1.43	40.00	1040	97.11	≥ 6	46.83	LM	7
139	Wang et al.(2022b)	NR	NR	NR	NR	C; S	NR	1–4	NR	40.21	771	98.85	≥ 11	3.11	LM	8
140	Wang et al.(2022c)	Shanghai	S	2020.11	DUR	C	Yes	1–5	20.0 ± 1.0	48.27	663	93.91	≥ 8	24.13	LM	6
141	Wang et al.(2023a)	Xizang Autonomous Region	S	2021.6-2021.7	DUR	C; CS	NR	NR	19.90 ± 1.34	38.57	4325	88.54	≥ 6	45.69	LM	8
142	Wang et al.(2023b)	Qinghai	N	2021.3-2021.4	DUR	M	Both	1–4	19.95 ± 1.33	43.12	1438	97.10	≥ 6	27.26	NR	7
143	Wang et al.(2023c)	Jiangsu/Hubei/Chongqing/Shanxi	Both	2021.3-2021.4	DUR	C; R	NR	1–4	NR	48.11	1220	99.84	≥ 6	43.28	LM	7
144	Wang et al.(2024)	Shanxi	N	2022.10-2022.11	DUR	M; R;S	NR	1–4	NR	34.90	18,723	95.40	≥ 8	16.37	LM	9
145	Wang et al.(2025a)	Zhejiang	S	2025.6-2025.9	POST	C; CS	Yes	NR	18.98 ± 0.90	39.31	2712	92.69	≥ 11	4.28	LM	10
146	Wang et al.(2025b)	Chongqing Shandong	Both	NR	NR	CS	Both	1–5	21.32 ± 1.48	37.38	519	90.58	≥ 6	48.75	NR	7
147	Wen et al.(2019)	Guangdong	S	2016.12	PRE	C; R	Both	1–3	NR	42.31	624	97.20	NR	28.37	NR	4
148	Wu et al.(2008)	Wuhan	S	2006.6	PRE	C; S	Both	NR	22.50	49.46	645	92.14	≥ 8	24.50	NR	4
149	Wu et al.(2014)	Zhejiang	S	NR	NR	C; S	NR	1–4	NR	47.59	1372	96.62	≥ 8	23.62	LM	5
150	Wu et al.(2015)	Anhui	S	2013.10	PRE	R; S	Both	1–3	19.24 ± 1.41	41.56	4747	96.58	≥ 8	9.77	LM	8
151	Wu et al.(2021)	Guangdong	S	NR	NR	R	Yes	NR	NR	46.95	754	NR	NR	37.40	NR	6
152	Wumaier et al.(2022)	Xinjiang	N	2020.9-2020.11	DUR	C; S	Both	NR	NR	37.24	2323	92.92	≥ 6	34.52	LM	8
153	Xi et al.(2018)	Hunan	S	2017.11	PRE	R; S	Both	1–3	NR	34.50	2767	95.66	≥ 8	40.19	NR	7
154	Xia et al.(2015)	Jiangsu徐州	N	2012.9-2012.12	PRE	M; R	Both	NR	NR	48.16	1194	99.50	≥ 8	23.62	LM	6
155	Xian et al.(2023)	Chongqing	S	2022.9.19-2022.9.27	DUR	E; CS	Yes	2–5	NR	28.33	660	92.70	≥ 6	52.73	NR	7
156	Xiao et al.(2000)	Shanghai	S	NR	NR	S	No	1–4	NR	78.74	621	95.54	≥ 8	18.04	NR	4
157	Xiao et al.(2005)	NR	NR	NR	NR	C	Yes	1–5	21.0 ± 1.0	59.14	3204	90.46	≥ 8	25.97	NR	4
158	Xiao et al.(2016)	Jiangxi	S	2015.5-2015.6	PRE	C	NR	NR	19.68 ± 1.16	59.17	2422	96.88	≥ 8	18.70	NR	8
159	Xiao et al.(2017)	Hunan	S	NR	NR	C; R	Yes	1	18.50 ± 0.60	38.22	3006	94.11	≥ 7	13.57	LM	6
160	Xie et al.(2011)	Tianjin	N	2008.4-2008.6	PRE	C; R;S	Both	1–5	21.12 ± 1.57	50.11	3207	91.63	≥ 8	13.22	NR	6
161	Xie et al.(2019)	Anhui	S	NR	NR	CS	No	1–5	21.73 ± 2.33	12.74	777	92.50	≥ 8	42.60	LM	6
162	Xie et al.(2020a)	Hunan	S	2018.9-2018.12	PRE	CS	NR	1–3	18.21 ± 1.18	47.83	759	84.33	≥ 8	16.47	LM	9
163	Xie et al.(2020b)	Anhui/Jiangxi	S	2018.6-2018.7	PRE	C; R	Both	NR	NR	44.51	4624	96.59	≥ 8	15.57	LM	6
164	Xie et al.(2021)	Yunnan	S	2020.2	DUR	S	Yes	1–5	NR	36.40	1026	NR	≥ 8	33.20	1-2w	7
165	Xu et al.(2017)	Zhejiang	S	2014.10-2014.11	PRE	C	NR	NR	20.89 ± 1.30	28.06	588	94.84	≥ 8	12.93	NR	8
166	Xu et al.(2021)	Hunan	S	2019.9-2019.12	PRE	C; R; S	Both	1–5	NR	NR	874	97.11	≥ 8	26.43	NR	4
167	Xu et al.(2025)	Zhejiang	S	2024.10-2024.12	POST	C; R; S	NR	2	20.59 ± 0.62	42.04	2117	71.68	≥ 8	38.50	NR	8
168	Yan et al.(2017)	Gansu	N	2016.4-2016.5	PRE	C; R	Both	1–5	19.90 ± 1.20	50.66	685	97.86	≥ 8	15.47	NR	6
169	Yang et al.(2000)	NR	NR	NR	NR	C; S	Yes	1–4	20.71 ± 1.23	51.20	584	97.01	≥ 8	15.75	LM	5
170	Yang et al.(2011)	Hubei	S	NR	NR	R; S	Both	1–3	NR	50.51	887	98.56	≥ 7	30.10	LM	4
171	Yang et al.(2019)	Anhui	S	NR	NR	C; R	Yes	1–5	21.16 ± 1.44	47.67	1137	98.02	≥ 8	26.30	NR	5
172	Yang et al.(2020)	NR	NR	2019.5-2019.8	PRE	CS	NR	1–4	21.60 ± 3.10	54.99	1564	94.79	≥ 8	21.10	NR	7
173	Yang et al.(2021)	Yunnan	S	NR	NR	CS	NR	NR	18.50 ± 0.96	28.40	845	93.89	≥ 8	19.53	NR	5
174	Yang et al.(2022a)	Gansu	N	2020.9-2020.10	DUR	C; R; S	Both	1–4	19.90 ± 1.40	33.28	1737	97.26	≥ 8	35.35	NR	7
175	Yang et al.(2022b)	Guizhou	S	NR	NR	CS	NR	NR	19.60 ± 1.08	28.47	1110	92.50	≥ 8	15.23	LM	8
176	Yang et al.(2022c)	NR	NR	2020.11-2021.1	DUR	C; R; S	No	1–4	NR	79.80	604	85.07	≥ 8	16.06	NR	8
177	Yang et al.(2022d)	Guangdong	S	2018.4-2021.5	PRE	CS	NR	NR	19.89 ± 1.41	24.89	936	NR	≥ 8	27.56	NR	5
178	Yang et al.(2023)	Liaoning/Anhui/Jiangxi	Both	NR	NR	C; M; S	NR	NR	19.12 ± 1.04	42.95	11,423	97.72	≥ 7	35.88	NR	8
179	Yang et al.(2003)	Taiwan	S	NR	NR	C; CS	NR	1	18.52 ± 0.93	72.11	1922	64.07	≥ 6	40.37	NR	5
180	Yao et al.(2022)	Shandong/Anhui/Henan/Shanxi/Gansu	Both	NR	NR	M	Yes	1–3	NR	33.59	3700	96.73	≥ 8	20.97	LM	7
181	Ye et al.(2013)	Guangdong	S	NR	NR	C	No	1–2	NR	100.00	571	99.30	≥ 8	10.16	NR	6
182	Ye et al.(2016)	Wuhan	S	2012.5-2012.6	PRE	C; M	NR	1–2	19.70 ± 1.20	59.20	2422	89.70	≥ 6	42.60	LM	8
183	Ye et al.(2019)	Shanghai	S	2018	PRE	C; R; S	NR	1–2	19.60 ± 0.90	62.33	4964	94.02	≥ 8	55.00	NR	6
184	Ye et al.(2022)	NR	NR	2020.2-2020.3	DUR	R	NR	1–4	19.58 ± 1.61	42.80	1106	94.77	NR	37.70	LM	8
185	Yin et al.(2025)	Shanxi	N	NR	NR	R	NR	1–5	NR	34.40	17,713	80.34	≥ 8	14.30	LM	7
186	You et al.(2020)	NR	NR	2017.12	PRE	R	NR	NR	20.20 ± 1.43	36.96	1104	80.76	≥ 9	25.45	LM	8
187	Yu et al.(2013)	Anhui	S	NR	NR	C; R	NR	1–2	19.23 ± 1.14	49.67	2744	92.05	≥ 8	17.38	NR	7
188	Yu et al.(2018a)	NR	NR	NR	NR	CS	NR	NR	20.28 ± 1.79	45.44	964	96.40	≥ 8	16.18	NR	6
189	Yu et al.(2018b)	Anhui	S	2017	PRE	C; S	Both	1–3	NR	0.00	1289	95.48	≥ 8	17.53	NR	5
190	Yu et al.(2022a)	Anhui	S	2021.10-2021.11	DUR	CS	Yes	NR	NR	9.87	628	NR	≥ 8	35.99	LM	6
191	Yu et al.(2022b)	Anhui	S	2019.10-2019.12	PRE	C; S	Both	1–4	NR	53.64	865	66.64	≥ 8	25.43	NR	8
192	Yuan et al.(2015a)	Shanxi	N	NR	NR	C	NR	1–3	20.51 ± 1.31	28.03	4548	88.81	≥ 8	19.83	LM	5
193	Yuan et al.(2022)	Hunan	S	NR	NR	C	No	NR	NR	48.29	585	97.33	≥ 8	31.85	LM	6
194	Zang et al.(2024)	Hubei	S	2020.9-2020.12	DUR	C; R	Yes	NR	19.27 ± 1.26	36.87	979	NR	≥ 8	17.47	LM	8
195	Zhai et al.(2022)	Anhui/Jiangxi	S	2019.4-2019.4	PRE	CS	Both	NR	18.68 ± 0.99	32.48	702	91.05	≥ 8	12.54	NR	8
196	Zhang et al.(2013)	NR	NR	2012	PRE	C; R	NR	NR	20.50 ± 0.90	69.89	837	98.47	≥ 8	16.96	NR	7
197	Zhang et al.(2014a)	Xinjiang	N	NR	NR	C; S	Yes	1–4	NR	34.60	552	92.77	≥ 8	42.93	NR	4
198	Zhang et al.(2014b)	Hebei	N	NR	NR	C; R	Both	1–4	NR	46.19	1325	98.73	≥ 8	15.77	NR	4
199	Zhang et al.(2015)	Xinjiang	N	NR	NR	C; S	Yes	NR	NR	NR	639	91.29	≥ 8	69.48	NR	5
200	Zhang et al.(2016a)	Henan	N	NR	NR	C	Both	NR	NR	42.74	503	91.45	≥ 7	22.47	NR	4
201	Zhang et al.(2016b)	Xinjiang	N	2014.11	PRE	C; S	Both	1–4	NR	28.53	771	85.67	≥ 8	18.94	NR	8
202	Zhang et al.(2017a)	Shandong	N	2013	PRE	C; R	NR	1–4	NR	69.76	840	NR	NR	14.52	NR	5
203	Zhang et al.(2017b)	Ningxia Hui Autonomous Region	N	NR	NR	C; R; S	Yes	1–4	20.98 ± 11.47	37.91	2263	94.29	≥ 8	21.30	NR	6
204	Zhang et al.(2018)	Shandong	N	2016.9-2016.10	PRE	R	Both	1	NR	38.21	547	91.17	≥ 8	26.51	NR	7
205	Zhang et al.(2021)	Anhui	S	2019.11	PRE	CS	Yes	1–3	19.32 ± 1.04	58.28	851	NR	≥ 8	16.22	NR	5
206	Zhang et al.(2022)	Sichuan/Fujian/Heilongjiang/Henan	Both	2021.9-2021.11	DUR	SSnow	NR	1–4	19.0	44.87	1928	96.26	≥ 6	45.07	NR	9
207	Zhang et al.(2023a)	Guangdong	S	2020.5-2020.7	DUR	C; R	NR	NR	20.72 ± 3.55	39.23	2526	98.98	NR	48.18	LM	7
208	Zhang et al.(2023b)	Hubei	S	2021.9-2022.1	DUR	CS	Yes	1–5	18.92 ± 1.32	34.01	3423	96.31	≥ 6	43.03	LM	8
209	Zhang et al.(2024a)	Guangdong/Anhui	S	2020.8	DUR	CS	NR	NR	20.70 ± 1.60	36.40	1793	69.31	NR	29.70	NR	7
210	Zhang et al.(2024b)	Heilongjiang/Sichuan/Fujian/Henan	Both	2021.9-2021.12	DUR	CS	NR	1–4	19.65 ± 1.71	44.76	1966	NR	≥ 11	5.29	1-2w	6
211	Zhang et al.(2024c)	Hunan/Jiangxi/Guangdong	S	2023.9-2023.10	POST	C; CS	No	1–3	20.87 ± 2.03	62.71	7205	95.51	≥ 8	24.61	NR	8
212	Zhang et al.(2025a)	Chongqing	S	2024.10-2024.11	POST	CS	NR	NR	18.42 ± 1.10	42.96	5803	91.11	≥ 6	15.54	NR	6
213	Zhang et al.(2025b)	Jiangxi/Hunan/Hubei	S	2023.11-2023.12	POST	CS	NR	NR	19.78 ± 1.36	44.06	6600	90.91	≥ 6	14.71	LM	7
214	Zhao et al.(2018)	Ningxia Hui Autonomous Region	N	2017.4-2017.6	PRE	C; R	Both	NR	20.80 ± 1.50	42.63	1818	90.95	≥ 8	48.90	NR	8
215	Zhao et al.(2020)	Heilongjiang	N	2018.11-2018.12	PRE	C; R; S	Both	NR	NR	NR	908	85.90	≥ 8	21.92	LM	8
216	Zhao et al.(2024)	Hunan	S	2023.2-2023.6	DUR	C; R	NR	1–2	18.80	44.61	3490	NR	≥ 8	30.40	NR	8
217	Zhao et al.(2025)	Shanxi	N	NR	NR	CS	Yes	1–3	19.92 ± 1.36	23.05	6524	96.92	≥ 8	27.21	LM	6
218	Zhen et al.(2021)	Shanxi	N	NR	NR	C; R	NR	2–3	NR	54.77	639	91.29	≥ 8	43.97	LM	4
219	Zheng et al.(2016a)	Beijing	N	2014.20-2014.12	PRE	C; R; S	Yes	2–5	20.20 ± 1.30	44.61	603	NR	NR	22.39	LM	8
220	Zheng et al.(2016b)	Beijing	N	2014.10	PRE	C; R; S	Both	2–5	20.10 ± 1.30	46.29	512	92.25	≥ 8	29.10	LM	9
221	Zheng et al.(2025)	Xinjiang/Guangdong/Fujian/Beijing/Henan/Shanxi	Both	2023.12-2024.6	POST	C; R; S	NR	NR	22.35 ± 0.42	52.29	960	98.87	≥ 11	25.00	NR	6
222	Zhou et al.(2013)	Jiangxi	S	NR	NR	C; R	Both	1–3	NR	43.76	1145	95.42	≥ 8	16.42	LM	5
223	Zhou et al.(2016)	Guangxi	S	NR	NR	R	NR	NR	NR	32.94	510	96.23	≥ 11	11.70	NR	5
224	Zhou et al.(2019)	NR	NR	NR	NR	CS	Both	1–5	NR	49.41	510	92.73	NR	25.69	NR	5
225	Zhou et al.(2022)	Guizhou	S	2018.12-2019.1	PRE	C; S	Both	1–4	19.80 ± 1.30	39.23	1063	88.58	≥ 6	53.70	LM	9
226	Zhu et al.(2015)	Hebei	N	NR	NR	CS	Both	1–5	22.0	37.25	553	92.16	≥ 8	22.60	LM	5
227	Zhu et al.(2016)	Heilongjiang	N	2014.9	PRE	C; R; S	NR	NR	20.68 ± 1.09	32.11	928	94.69	≥ 8	32.11	LM	7
228	Zhu et al.(2017)	Shanxi	N	NR	NR	C; R	Both	1–5	21.28 ± 1.81	89.04	712	97.53	≥ 8	32.16	NR	4
229	Zhu et al.(2024)	Jiangsu	S	2022.9	DUR	C; R; S	NR	1–4	NR	36.70	4670	78.09	≥ 5	52.61	NR	6
230	Zhu et al.(2025a)	NR	NR	NR	NR	R	NR	NR	19.02 ± 1.08	36.10	1209	90.77	≥ 8	23.60	LM	8
231	Zhu et al.(2025b)	Shanghai/Hubei/Jiangxi	S	2023.9-2023.11	POST	CS	NR	NR	19.70 ± 1.23	39.51	7954	88.87	≥ 11	8.40	LM	8
232	Zou et al.(2011)	Anhui	S	NR	NR	C; R	NR	1–4	18.66 ± 1.71	71.88	793	79.30	≥ 8	10.84	NR	5

Abbreviations: (1)Region: ① N: North of China (north of the Qinling Mountains-Huai River line); ② S: South of China (south of the Qinling Mountains-Huai River line); ③ Both: Including both regions; ④ NR: Not reported.(2)Pandemic Context: Studies were categorized based on the timing of primary data collection: ①PRE: Pre-pandemic: concluded before January 30, 2020; ②DUR: During-pandemic: entirely or predominantly occurred between January 30, 2020 and May 5, 2023; ③POST: Post-pandemic: initiated after May 6, 2023; ④NR: period not clearly reported (“NR”) or explicitly spanning the above categories (e.g., from 2019 to 2020).(3) Sampling method: ①C: Cluster sampling; ②M: Multistage sampling; ③R: Random sampling; ④S: Stratified sampling; ⑤E: Episodic sampling; ⑥CS: Convenience sampling; ⑦ Cen.: Census; ⑧S: Snowball sampling; ⑨P: Purposive sampling. (4) Age was reported as presented in the original study (mean ± standard deviation or mean).(5) Timeframe: ①1–2 W: past 1–2 weeks; ②LM: last month; ③FM,: past few months; ④NR: not reported

Meta-analysis of sleep disturbance prevalence among chinese college students

This meta-analysis included 232 studies comprising 495,641 participants. Heterogeneity was extreme (I²=99.57%, P < 0.001; τ²=0.109). Using a random-effects model, the pooled prevalence of sleep disturbance among Chinese college students was 26.4% (95% CI: 24.6%–28.3%). The 95% PI ranged from 7.4% to 59.1%.

Subgroup analysis of sleep disturbance prevalence among chinese college students

To further explore the sources of heterogeneity and the distribution characteristics of sleep disturbance prevalence among Chinese college students, this study conducted systematic subgroup analyses based on demographic characteristics (gender, grade, place of origin, only child status, family circumstances), academic and major factors (professional category), time and contextual factors (publication period, pandemic context), geographical factors, research methodology factors (effective sample size, diagnostic criteria, assessment duration, literature quality), and health behavior factors (physical exercise, smoking, alcohol consumption). Related results are presented in Table 2.

Table 2.

Subgroup analysis

Subgroups		Number of included studies	Sample size	Events	I2(%)	Prevalence (%, 95% CI)	Between-subgroup comparison
							Q	P
Gender							0.53	0.47
	Male	90	78,833	21,418	99.0	25.7 (22.7–28.9)
	Female	94	117,704	34,640	99.3	27.4 (24.3–30.5)
Grade							0.89	0.83
	First	55	43,664	12,672	98.7	26.8 (23.1–30.7)
	Second	57	35,601	9969	98.4	27.8 (24.2–31.6)
	Third	54	25,375	6779	98.0	28.8 (24.8–33.0)
	Fourth or fifth	36	13,964	3535	97.1	29.3 (24.6–34.2)
Place of origin							0.95	0.62
	Urban	25	31,664	7435	99.0	24.2 (19.2–29.5)
	Township	9	3654	954	95.6	28.4 (21.3–36)
	Rural	24	38,891	8873	99.2	24.6 (19.7–29.8)
Only child							0.50	0.48
	Yes	14	15,102	4072	97.0	27.2 (22.9–31.7)
	No	14	25,791	7312	98.8	29.7 (24.5–35.2)
Family circumstances							4.46	0.11
	Good	10	3335	746	91.3	18.3 (13.7–23.4)
	Average	10	21,215	5580	99.3	22.9 (16.0- 30.6)
	Poor	11	6253	1789	98.0	28.3 (20.6–36.7)
Professional category							0.54	0.46
	Medical students	29	15,316	4505	97.3	28.1 (23.8–32.6)
	Non-medical students	28	31,658	8671	97.9	26 (22.6–29.6)
publication period							8.08	0.018
	2014 and earlier	40	49,474	11,537	98.3	22.9 (20.2–25.8)
	2015–2019	78	125,781	32,977	99.5	25.0 (21.8–28.3)
	2020 and later	114	320,386	85,351	99.7	28.7 (25.8–31.6)
Pandemic context							14.99	0.001
	Pre-pandemic	86	150,569	37,941	99.4	24.1 (21.3–27)
	During the pandemic	45	122,768	41,861	99.7	33.5 (28.4–38.8)
	Post-pandemic	15	82,256	14,025	99.5	21.0 (17.0-25.2)
Region							0.26	0.61
	North	66	141,150	34,296	99.3	27.0 (24.3–29.8)
	South	132	261,369	70,475	99.6	26.0 (23.5–28.6)
Effective sample							0.15	0.93
	501–900	87	59,820	15,983	98.4	26.1 (23.4–29)
	900–2000	80	101,125	27,698	99	26.2 (23.6–29.0)
	>2000	65	334,696	86,184	99.8	27.0 (23.3–30.9)
PSQI cut-off							69.50	<0.001
	≥ 16	1	686	170	0	24.8 (21.7–28.1)
	≥ 11	7	15,423	1416	99.1	11.6 (6.3–18.2)
	≥ 9	5	9161	1964	99.8	19.9 (12.3–28.8)
	≥ 8	155	277,197	65,690	99.3	24.4 (22.6–26.3)
	≥ 7	9	20,403	5970	99.0	24.0 (17.4–31.3)
	≥ 6.5	1	1349	365	0	27.1 (24.8–29.5)
	≥ 6	33	133,211	42,373	99.8	38.8 (32.3–45.5)
	≥ 5	5	8859	4770	99.7	54.2 (44.1–64.1)
Time frame							147.42	<0.001
	past month	95	243,227	62,188	99.7	26.9 (23.8–30.2)
	past 1–2 weeks	2	2992	445	0	12.6 (11.5–13.8)
	past several months	1	1333	343	0	25.7 (23.5–28.1)
Literature quality							0.006	0.94
	AHRQ scores 4–7	158	265,832	69,346	99.3	26.4 (24.4–28.4)
	AHRQ scores 8–11	74	229,809	60,519	99.8	26.5 (22.8–30.4)
Physical exercise							0.91	0.34
	Yes	11	10,491	2352	98.3	23.1 (16.6–30.2)
	No	11	23,682	5260	99.2	28.3 (20.6–36.7)
Smoking							3.40	0.065
	Yes	8	4455	1154	96.1	30.5 (21.4–40.4)
	No	8	42,208	9540	99.8	19.0 (11.7–27.6)
Drinking							2.69	0.10
	Yes	8	13,630	4005	98.0	28.0 (22.3–34.2)
	No	8	34,106	6909	99.7	19.5 (12.2–28.1)

Subgroup estimates based on fewer than 5 studies were considered exploratory and should be interpreted with caution

Gender

The prevalence rate among male college students was 25.7% (95% CI: 22.7%-28.9%, I² = 99.0%, P < 0.001), while that among female college students was 27.4% (95% CI: 24.3%-30.5%, I² = 99.3%, P < 0.001). Subgroup analysis revealed no statistically significant difference in prevalence between genders (Q = 0.53, P = 0.47). All analyses exhibited high heterogeneity (I² > 99%, P < 0.001).

Grade

The prevalence rate among first-year college students was 26.8% (95% CI: 23.1%-30.7%, I² = 98.7%, P < 0.001), the prevalence among second-year college students was 27.8% (95% CI: 24.2%-31.6%, I² = 98.4%, P < 0.001), the prevalence among third-year college students was 28.8% (95% CI: 24.8%-33.0%, I² = 98.0%, P < 0.001), and fourth-year and above college students had a prevalence of 29.3% (95% CI: 24.6%-34.2%, I² = 97.1%, P < 0.001). Although prevalence values showed a slight upward trend with grade level, the difference between groups was not statistically significant (Q = 0.89, P = 0.83).

Place of origin

Subgroup analysis revealed that the prevalence of sleep disturbance among college students from urban, township, and rural areas was 24.2% (95% CI: 19.2%-29.5%, I² = 99.0%, P < 0.001), 28.4% (95% CI: 21.3%-36.0%, I² = 95.6%, P < 0.001), and 24.6% (95% CI: 19.7%-29.8%, I² = 99.2%, P < 0.001), respectively. Although the point estimates for the prevalence among students from townships were slightly higher, the differences between subgroups were not statistically significant (Q = 0.95, P = 0.62).

Only child

There was no statistically significant difference in the prevalence of sleep disturbance between only child and non-only child among college students. Specifically, the combined prevalence rate was 27.2% (95% CI: 22.9%–31.7%, I² = 97.0%, P < 0.001) among only child and 29.7% (95% CI: 24.5%–35.2%, I² = 98.8%, P < 0.001). The test for group differences showed no statistical significance (Q = 0.50, P = 0.48).

Family circumstances

The combined prevalence of sleep disturbance among college students from families with “good,” “average,” and “poor” economic status was 18.3% (95% CI: 13.7%–23.4%, I² = 91.3%), 22.9% (95% CI: 16.0%–30.6%, I² = 99.3%), and 28.3% (95% CI: 20.6%–36.7%, I² = 98.0%), respectively. The test for between-group differences did not reach conventional statistical significance (Q = 4.46, P = 0.11).

Professional category

Medical students exhibited a prevalence rate of 28.1% (95% CI: 23.8%-32.6%, I² = 97.3%, P < 0.001), while non-medical undergraduate students showed a prevalence rate of 26.0% (95% CI: 22.6%-29.6%, I² = 97.9%, P < 0.001). However, the difference between groups was not statistically significant (Q = 0.54, P = 0.46).

Publication period

Subgroup analysis by publication period showed pooled prevalence estimates of 22.9% (95% CI: 20.2%–25.8%, I² = 98.3%, P < 0.001) for studies published in 2014 or earlier, 25.0% (95% CI: 21.8%–28.3%, I² = 99.5%, P < 0.001) for 2015–2019, and 28.7%(95% CI: 25.8%–31.6%, I² = 99.7%, P < 0.001) for 2020 or later. The difference across periods was statistically significant (Q = 8.08, P = 0.018).

Pandemic context

The period during which the pandemic was officially declared a “Public Health Emergency of International Concern” by the World Health Organization (January 30, 2020, to May 5, 2023) is considered the “pandemic period [262, 263]” (which can be seen in Table 1 footnote for specific groupings). The prevalence rate before the COVID-19 pandemic was 24.1% (95% CI: 21.3%–27.0%, I² = 99.4%, P < 0.001). The prevalence during the pandemic period was 33.5% (95% CI: 28.4%–38.8%, I² = 99.7%, P < 0.001), and 21% for post-pandemic(95% CI: 17.0%–25.2%, I² = 99.5%, P < 0.001). Subgroup analysis revealed statistically significant differences in sleep disturbance prevalence across the three periods (Q = 14.99, P = 0.001).

Region

The prevalence rate in the north of China was 27.0% (95% CI: 24.3%-29.8%, I² = 99.3%, P < 0.001), while the south of China had a prevalence of 26.0% (95% CI: 23.5%-28.6%, I² = 99.6%, P < 0.001). The difference between groups was not statistically significant (Q = 0.26, P = 0.61).

Effective sample

The prevalence rate for an effective sample size of 501–900 was 26.1% (95% CI: 23.4%–29.0%, I² = 98.4%, P < 0.001). The prevalence rate for effective sample sizes of 900–2000 was 26.2% (95% CI: 23.6%–29.0%, I² = 99.0%, P < 0.001), and the prevalence rate for an effective sample size > 2000 was 27.0% (95% CI: 23.3%-30.9%, I² = 99.8%, P < 0.001). Subgroup analysis revealed no statistically significant differences among the three groups (Q = 0.15, P = 0.93).

PSQI cut-off

To improve clarity, we standardized PSQI cut-off categories in descending order: ≥16, ≥ 11, ≥9, ≥ 8, ≥7, ≥ 6.5, ≥ 6, and ≥ 5. Subgroup differences were significant (Q = 69.50, P < 0.001), with a general pattern of higher pooled prevalence at more lenient cut-offs. Specifically, the pooled prevalence was 24.8% for PSQI ≥ 16 (1 study), 11.6% for ≥ 11 (7 studies), 19.9% for ≥ 9 (5 studies), 24.4% for ≥ 8 (155 studies), 24.0% for ≥ 7 (9 studies), 27.1% for ≥ 6.5 (1 study), 38.8% for ≥ 6 (33 studies), and 54.2% for ≥ 5 (5 studies).

Time frame

The prevalence rate for the past month was 26.9% (95% CI: 23.8%-30.2%, I² = 99.7%, P < 0.001), the prevalence rate for the past 1–2 weeks was 12.6% (95% CI: 11.5%–13.8%, I² = 0), and the prevalence rate for the past several months was 25.7% (95% CI: 23.5%–28.1%, I² = 0). Subgroup analysis revealed extremely significant differences across time periods (Q = 147.42, P < 0.001).

Literature quality

A subgroup analysis was conducted based on the quality scores of included studies. Results indicated that study quality did not significantly influence estimates of sleep disturbance prevalence among Chinese college students. The pooled prevalence rate for the group of 158 moderate-quality studies (AHRQ scores 4–7) was 26.4% (95% CI: 24.4%–28.4%), while the pooled prevalence in the 74 high-quality studies (AHRQ scores 8–11) was 26.5% (95% CI: 22.8%-30.4%). The prevalence difference between the two groups was not statistically significant (Q = 0.006, P = 0.94). This indicates that the primary findings of this study are robust to variations in study quality.

Physical exercise

Subgroup analysis of physical exercise behavior revealed no statistically significant difference in the prevalence of sleep disturbance between college students who exercised and those who did not. The combined prevalence among exercisers was 23.1% (95% CI: 16.6%–30.2%, I² = 98.3%, P < 0.001), while non-exercisers had a prevalence of 28.3% (95% CI: 20.6%–36.7%, I² = 99.2%, P < 0.001). The difference between groups was not statistically significant (Q = 0.91, P = 0.34).

Smoking

The combined prevalence of sleep disturbance among smoking college students was 30.5% (95% CI: 21.4%–40.4%, I² = 96.1%, P < 0.001), significantly higher than the 19.0% among non-smoking college students (95% CI: 11.7%–27.6%, I² = 99.8%, P < 0.001). The between-group difference test showed no statistical significance (Q = 3.40, P = 0.065).

Drinking

The combined prevalence among drinking college students was 28.0% (95% CI: 22.3%–34.2%, I² = 98.0%, P < 0.001). The pooled prevalence among non-drinking college students was 19.5% (95% CI: 12.2%–28.1%, I² = 99.7%, P < 0.001). The between-group difference test showed no statistical significance (Q = 2.69, P = 0.10).

Publication bias assessment

Publication bias in this meta-analysis was evaluated using both Begg’s rank correlation test and Egger’s regression test. Begg’s test revealed no significant publication bias (z = 0.88, P = 0.379). The Egger regression test produced an intercept of 1.867 (95% CI: -0.028 to 3.763), with P = 0.053, slightly above the 0.05 significance threshold. Together, the results of both tests provided no statistically significant evidence of publication bias, suggesting that the existing research findings are well-represented. However, the P-value of the Egger test was close to the critical threshold, highlighting the need for caution in future research regarding the potential non-publication of small-sample or negative findings. The funnel plot, shown in Fig. 2, demonstrates that the scatter points are generally symmetrically distributed, with no clear evidence of publication bias.

Fig. 2.

Publication bias test funnel plot

Sensitivity analysis

Sensitivity analyses were conducted by sequentially excluding each study to assess the impact of individual studies on the overall pooled effect size. Results indicated that the pooled estimate of sleep disturbance prevalence among Chinese college students remained stable after excluding any single study, suggesting that the overall conclusions of this meta-analysis exhibit high robustness and are not unduly influenced by any individual study.

Discussion

This updated meta-analysis included 232 studies with a total of 495,641 Chinese college students. The results revealed an overall prevalence of sleep problems among Chinese college students of 26.4% (95% CI: 24.6%–28.3%). Given the extreme between-study heterogeneity, the 95% PI was wide, ranging from 7.4% to 59.1%, indicating that the true prevalence may vary markedly across study settings. Potential contributors to this variability include differences in PSQI cut-off values, recall timeframes, pandemic-related study contexts, and regional or population characteristics. The pooled estimate should therefore be interpreted as a descriptive summary rather than a single value applicable across all settings. Compared with the 2018 meta-analysis estimate of 25.7% [264], the pooled prevalence in the present study was higher by 0.7% points. This modest difference should be interpreted cautiously and should not be taken as evidence of a true temporal increase, as it may reflect differences in the size and recency of the evidence base, as well as differences in study characteristics and measurement choices. Nonetheless, by synthesizing a larger body of evidence over a longer observation period, this meta-analysis provides a more up-to-date and comprehensive estimate of PSQI-defined sleep disturbance among Chinese college students. Collectively, these findings highlight the public health importance of sleep disturbance in this population and support the need for ongoing surveillance and targeted, campus-based health promotion efforts.

The study indicates extremely high heterogeneity (I² = 99.57%), indicating substantial variability across studies that cannot be fully accounted for by the examined moderating factors alone. Under these conditions, the pooled prevalence estimate of 26.4% should be interpreted as a descriptive summary reflecting the overall magnitude and general scope of sleep problems among Chinese college students, rather than as a precise or universally generalizable epidemiological parameter. Although the random-effects model provides an estimate of the central tendency across studies, the true prevalence is likely to vary considerably across different contexts, time periods, and measurement approaches. To further investigate potential sources of heterogeneity, subgroup analyses were conducted. While prevalence estimates varied across strata, overall heterogeneity remained high, suggesting that additional unmeasured factors may continue to contribute to between-study variability.

To avoid reiterating numerical results reported in the Results section, we focus the Discussion on interpreting the observed heterogeneity and its methodological and contextual implications. Specific findings are summarized as follows: (1) Methodological factors. Research methodology appears to be a primary contributor to differences in reported prevalence. Although the PSQI is widely used as a screening instrument for sleep disturbance, cutoff values vary across studies, and prevalence estimates differed significantly according to the threshold applied (Q = 69.50, P < 0.001). This is primarily due to differences in study populations, language versions, and validation strategies. To preserve comparability with the original reports, we categorized studies according to the specific cutoff values used in each study rather than imposing a uniform threshold. Some nonstandard cutoffs (e.g., ≥ 6.5) were retained because they reflected study-specific validation strategies or operational definitions. However, these categories were supported by only a small number of studies and should therefore be interpreted cautiously. The effect of the time frame was equally important. The prevalence rates for the previous 1–2 weeks (12.6%) were significantly lower than those for the preceding month (26.9%) or several months (25.7%) (Q = 147.42, P < 0.001), suggesting that prevalence estimates may differ by recall timeframe, although the evidence is limited and alternative explanations cannot be excluded. It is important to note that only one study covered the previous several months, and two studies covered the previous 1–2 weeks. Given the small number of studies in these subgroups, the corresponding point estimates (25.7% and 12.6%) may be unstable and influenced by study design, population characteristics, or measurement approaches. Consequently, the existing data are inadequate to accurately ascertain the genuine disparity in prevalence estimates between short-term (1–2 weeks) and long-term (several months) recollection periods. Beyond measurement-related differences, we also examined methodological characteristics at the study level, including effective sample size, to evaluate potential small-study effects. Subgroup analyses by sample size were not statistically significant (P = 0.93), suggesting limited evidence of a systematic influence of study size on pooled prevalence within the included scope. (2) Major societal events. The COVID-19 pandemic was associated with higher pooled prevalence estimates of sleep disturbance among Chinese college students. The pooled prevalence was 33.5% during the pandemic, compared with 24.1% before the pandemic and 21.0% after the pandemic. This pattern is consistent with previous research [265]. The higher prevalence observed during the pandemic may reflect disruption of daily routines, increased stress, and reduced social and academic stability [266]. However, these explanations remain speculative and should not be interpreted causally, given the predominantly cross-sectional, self-reported nature of the included evidence and the potential for residual confounding. In addition, the pandemic-period classification was based on the WHO Public Health Emergency of International Concern timeframe, which provided a common temporal anchor across studies but may not fully capture regional variation in lockdown measures, campus restrictions, and reopening policies across China [267]. (3) Publication period. When stratified by publication period, the pooled prevalence was 22.9% for studies published in 2014 or earlier, 25.0% for those published between 2015 and 2019, and 28.7% for those published in 2020 or later (Q = 8.08, P = 0.018). These findings suggest that pooled prevalence estimates differed across publication periods. Although publication year is not equivalent to the timing of data collection, this subgroup analysis remains informative for understanding how reported prevalence estimates vary across the evolving evidence base. The observed differences may reflect changes in study composition, measurement characteristics, and the greater representation of more recently published studies. Therefore, these between-period differences should be interpreted cautiously and should not be taken as direct evidence of a true temporal increase. (4) Group comparisons. Subgroup analyses revealed no statistically significant differences in prevalence across most demographic, behavioral, or geographic characteristics, including gender, academic year, place of origin, only child status, family economic status, major category, physical exercise, smoking, alcohol consumption and region. This pattern suggests that sleep problems are broadly distributed across Chinese college students. Nonetheless, several comparisons showed notable numerical trends. Point estimates of prevalence were higher among students who smoked (30.5%) than among non-smokers (19.0%; P = 0.065), and among students who self-reported “poor” family economic status (28.3%) compared to those from “good” family backgrounds (18.3%; between-group P = 0.11). These differences did not reach conventional levels of statistical significance and should therefore be interpreted as descriptive trends rather than confirmed associations. Given the cross-sectional design of the included studies and P-values exceeding 0.05, causal inferences cannot be drawn. However, these patterns may generate hypotheses for future longitudinal or interventional research examining potential pathways linking smoking behavior or socioeconomic status with sleep health in this population. More broadly, recent evidence suggests that sleep-related difficulties in university students may co-occur with wider affective, interpersonal, and behavioral vulnerabilities within multidimensional psychosocial profiles. This broader perspective helps situate student sleep health within overall well-being, although such evidence should be interpreted as contextual background rather than as direct evidence regarding sleep-disturbance prevalence [268].(5) Study quality. Prevalence estimates derived from studies of moderate and high methodological quality were comparable, suggesting that the principal findings of this meta-analysis are robust and not materially influenced by variations in study quality.

Limitations of this study: (1) All results were based on the PSQI, a self-report instrument primarily designed to screen for sleep quality problems rather than to establish clinical diagnoses. Therefore, the pooled estimates more accurately reflect the epidemiology of PSQI-defined sleep disturbance rather than the prevalence of clinically diagnosed sleep disorders in the strict sense. In addition, substantial variation in PSQI cut-off values across studies contributed to marked differences in reported prevalence, highlighting the need for standardized thresholds in future research. Moreover, none of the included studies incorporated objective sleep assessments, such as actigraphy or polysomnography. Self-reported sleep quality may be influenced by current mood states and reporting biases, which could affect the classification of PSQI-defined sleep disturbance and contribute to variability in prevalence estimates. (2) Although extensive subgroup analyses were performed, residual heterogeneity remained extremely high, indicating that additional unmeasured study-level factors may continue to contribute to between-study variability. These may include academic stress, depressive symptoms and other psychiatric comorbidities, chronotype, caffeine intake, socioeconomic status, and patterns of smartphone/screen use, which were often inadequately measured or inconsistently reported across the included studies and therefore could not be examined comprehensively. The wide 95% PI (7.4%–59.1%) further reflects substantial variability across settings and populations. (3) Every study that was included was cross-sectional, which precludes causal inference and prevents robust evaluation of temporal or directional relationships between sleep problems and associated factors. In addition, considerable residual heterogeneity may have reduced the precision of the pooled prevalence estimates. (4) The inclusion criterion requiring a minimum sample size of 500, although intended to improve estimate stability, may have excluded smaller studies that capture unique subpopulations or local contexts, thereby potentially reducing the comprehensiveness of the evidence base. (5) Certain subgroup analyses, especially those based on specific PSQI cutoff values and behavioral characteristics, were informed by a limited evidence base, potentially affecting the precision and reliability of the associated pooled estimates. These findings should therefore be considered exploratory and interpreted with caution. (6) Although Begg’s test suggested no significant publication bias, the Egger’s test P-value was close to the significance threshold (P = 0.053), indicating a potential risk of small-study effects. While the trim-and-fill method suggested minimal impact on the pooled estimate, the possibility of unpublished null or negative findings cannot be entirely excluded.

Conclusion

In summary, this updated meta-analysis, based predominantly on cross-sectional observational studies, indicates that PSQI-defined sleep disturbance remains common among Chinese college students, with a pooled prevalence of 26.4%(95% CI: 24.6% to 28.3%, 95% PI: 7.4%–59.1%). This substantial burden, together with the wide prediction interval indicating marked contextual variability, underscores that sleep health is an important public health concern in this population and warrants continued monitoring. Sleep outcomes in college students may reflect an interplay of individual vulnerability, health-related behaviors, academic and campus environments, and broader societal stressors, with these influences potentially amplified during major social disruptions.

Given the observational nature of the included studies and the extremely high heterogeneity, this pooled estimate of PSQI-defined sleep disturbance should be interpreted as a descriptive synthesis of reported prevalence across diverse settings rather than as a single precise epidemiological parameter. Methodological variability, particularly differences in PSQI cut-off thresholds and assessment time frames, likely contributes to between-study variation. These findings underscore the need for improved standardization in measurement and reporting in future research.

Although most subgroup comparisons did not reach statistical significance and the cross-sectional evidence does not allow causal inference, several implications for campus health policy warrant consideration. First, the consistently high prevalence suggests that universities may consider integrating sleep health education and routine screening into student health services, with clear pathways for timely referral and support for students experiencing clinically significant sleep difficulties. Second, although not statistically significant, directional patterns observed for health behaviors and socioeconomic indicators may help generate hypotheses and guide the prioritization of support for potentially vulnerable groups. These patterns, however, require confirmation in longitudinal studies with adequate adjustment for confounding. Third, the higher prevalence observed during the pandemic period highlights the importance of scalable and responsive campus support systems, allowing mental health and sleep-related services to be strengthened rapidly during major societal disruptions.

Although most subgroup comparisons did not reach statistical significance and the cross-sectional evidence does not allow causal inference, several public health implications may warrant consideration. The consistently high prevalence highlights the importance of continued attention to sleep health in university settings. In addition, although not statistically significant, directional patterns observed for health behaviors and socioeconomic indicators may help generate hypotheses and inform future efforts to identify potentially vulnerable groups. These observations, however, require confirmation in longitudinal studies with adequate adjustment for confounding. The higher prevalence observed during the pandemic period also highlights the importance of maintaining flexible and responsive campus health systems during major societal disruptions.

Improving sleep health among college students will likely require a coordinated public health approach involving universities both and public health stakeholders. Future research should prioritize longitudinal designs, standardized PSQI operationalization, including consistent cut-offs and recall timeframes, and rigorous evaluation of modifiable risk and protective factors to better inform future research, prevention, and health promotion efforts.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

PSQI

Pittsburgh Sleep Quality Index

CI

confidence interval

PI

prediction interval

PRISMA 2020

Preferred Reporting Items for Systematic Reviews and Meta-Analyses

MOOSE

Meta-Analysis of Observational Studies in Epidemiology

AHRQ

Agency for Healthcare Research and Quality

e.g.

exempli gratia

Author contributions

Qinghe Peng: Conceptualization, Methodology, Investigation, Supervision, Validation, Project administration, Writing-review & editing. Hao Zhang and Ziwei Guo: Data Curation, Validation, Writing-original draft. All authors read and approved the final manuscript.

Funding

Funding for this research was provided by the Anhui Provincial Department of Education 2024 Comprehensive Education and Ideological-Political Capacity Enhancement Project: “Ying Shan Hong” Counselor Master Teacher Studio (Project Number: sztsjh-2024-8-11); Anhui Provincial Department of Education 2025 Key Humanities and Social Sciences Project for Higher Education Institutions: Coupling Mechanisms and Optimization Pathways for Cultivating a Sense of Community for the Chinese Nation through Cultural and Museum Research and Study under the Perspective of Cultural Embedding: An Empirical Study Based on Anhui-Xinjiang Practices (Project Number.: 2025AHGXSK30457).

Data availability

The data used to support the findings of this study are included in the article.

Declarations

Ethics approval and consent to participate

This review article is based on previously published studies, and no new human or animal experiments were conducted by the authors. All participants provided informed consent.

Consent for publication

All authors have agreed to publish this manuscript.

Declaration of generative AI and AI-assisted technologies in the manuscript preparation process

During the preparation of this work the authors used ChatGTP-4.0, Open AI in order to improve readability and language. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the published article.

Competing interests

The authors declare no competing interests.