Skip to main content
NCBI home page
PLOS One logoLink to PLOS One
. 2026 Jan 14;21(1):e0321347. doi: 10.1371/journal.pone.0321347

Associations between sleep duration and depression, mental health, physical health, and general health in U.S. adults: A population-based study

Mojisola Fasokun 1, Oluwasegun Akinyemi 2, Fadeke Ogunyankin 3, Phiwinhlanhla Ndebele-Ngwenya 2,*, Kaelyn Gordon 2, Seun Ikugbayigbe 4, Uzoamaka Nwosu 5, Mariam Michael 6, Kakra Hughes 7, Temitope Ogundare 8
Editor: Valentina Alfonsi9
PMCID: PMC12803467  PMID: 41533644

Abstract

Introduction

Adequate sleep is vital for maintaining mental and physical health. In the United States, a substantial proportion of adults report sleep durations that fall outside the recommended range. Prior research has associated insufficient or excessive sleep with adverse health outcomes; however, few studies have systematically quantified these associations across multiple health indicators using nationally representative data.

Objective

This study aims to evaluate the impact of short sleep duration on four key health outcomes: depression diagnosis, number of self-reported poor mental health days, number of physically unhealthy days, and self-rated general health status, using nationally representative U.S. data.

Methodology: Methods

We analyzed nationally representative data from the Behavioral Risk Factor Surveillance System (BRFSS) collected between 2016 and 2023. Sleep duration was self-reported and categorized into three groups: short sleep (≤5 hours), recommended sleep (6–8 hours), and long sleep (≥9 hours), with short sleep serving as the reference category. The primary health outcomes included: (1) self-reported diagnosis of depression, (2) number of poor mental health days, (3) number of poor physical health days, and (4) self-rated general health, measured on a 5-point Likert scale from excellent to poor. To estimate the effect of sleep duration on these outcomes, we applied Inverse Probability Weighting (IPW) to derive the Average Treatment Effect (ATE), adjusting for key demographic and socioeconomic covariates. All analyses incorporated BRFSS complex survey weights to ensure national representativeness.

Results

The study included 318,000 adults (63.3% female; 74.5% White) with a mean age of 51.3 ± 18.4 years. Among individuals with recommended sleep duration (6–8 hours), the baseline prevalence of depression was 39.5% (95% CI: 39.4%–39.7%). Compared to this group, short sleep duration (≤5 hours) was associated with a 14.1 percentage point increase in depression incidence (95% CI: 13.8%–14.4%), while long sleep duration (≥9 hours) was linked to a 12.9 percentage point increase (95% CI: 12.5%–13.3%). Those with short sleep reported an average of 5.3 poor mental health days (95% CI: 5.3–5.4), 4.4 poor physical health days (95% CI: 4.3–4.4), and a higher prevalence of poor general health, 10.0% (0.1, 95% CI: 9.7%–10.2%), compared to individuals with recommended sleep. Similarly, individuals with long sleep duration (≥9 hours) also reported more poor mental (4.6 days, 95% CI: 4.5–4.7) and physical health days (3.2 days, 95% CI: 3.1–3.3), along with a higher prevalence of poor general health, 20.3% (20.3%%, 95% CI 19.4%–21.3%) compared to those with recommended sleep.

Conclusion

Both short (≤5 hours) and long (≥9 hours) sleep durations are significantly associated with increased risk of depression, more days of poor mental and physical health, and worse self-rated general health compared to recommended sleep (6–8 hours). Promoting optimal sleep duration through targeted public health interventions, education, and screening may improve population well-being and reduce sleep-related health disparities.

Introduction

Sleep is a fundamental biological necessity essential for overall health, cognitive function, and emotional well-being [1,2]. Despite its critical role, sleep deprivation remains a growing public health concern in the United States [3,4]. The Centers for Disease Control and Prevention (CDC) recognizes insufficient sleep as a public health epidemic, with a significant proportion of adults failing to meet the recommended seven to nine hours of sleep per night [5]. National surveys indicate that approximately one in three U.S. adults experience short sleep duration, with even higher prevalence among racial minorities, low-income populations, and those with demanding work schedules [6–8]. The consequences of insufficient or disrupted sleep extend far beyond personal health, impacting productivity, increasing accident risk, and imposing substantial economic burdens estimated in the billions of dollars annually [9–12]. Furthermore, sleep deprivation is linked to a lower quality of life, heightened risk of chronic illness, and impaired daily functioning, thereby affecting overall well-being [13,14].

Multiple factors contribute to sleep duration variability in the U.S., including occupational stress, socioeconomic conditions, behavioral patterns, and comorbid health conditions [15–17]. Modern lifestyle disruptions such as prolonged screen exposure, inconsistent sleep routines, and psychosocial stress further intensify this trend [18]. Additionally, sleep disturbances commonly co-occur with mental and physical health disorders such as anxiety, depression, and chronic pain, fostering a bidirectional relationship that compounds health risks [19,20]. Reduced sleep is associated with more frequent poor mental health days, greater emotional distress, and impaired physical functioning [21–23]. Consequently, understanding the nuanced relationship between varying levels of sleep and health outcomes is essential to guide effective public health strategies.

While existing literature has documented associations between inadequate sleep and adverse health outcomes [24,25], most studies have treated sleep duration as a binary exposure (sufficient vs. insufficient) and have not comprehensively examined a broader range of health impacts across differentiated sleep categories [26–28]. Moreover, few studies have concurrently assessed the effects of both short (<6 hours) and long (≥9 hours) sleep on a wide spectrum of outcomes, including depression, poor mental and physical health days, and self-rated general health, an important indicator of holistic well-being [29–32]. Recent evidence suggests that both insufficient and excessive sleep durations may be linked to adverse health profiles [33,34], but population-level analyses using representative U.S. data are limited.

To address these gaps, this study investigates the association between sleep duration categorized into three distinct groups: short sleep (≤5 hours), recommended sleep (6–8 hours), and long sleep (≥9 hours) and four self-reported health outcomes: incidence of depression, number of poor mental health days, number of poor physical health days, and general health status. By utilizing a nationally representative dataset and advanced causal modeling techniques, we aim to provide robust, population-level estimates that inform public health policies and clinical practices related to sleep hygiene and health promotion.

Methodology

Study design and data source

This study utilizes data from the Behavioral Risk Factor Surveillance System (BRFSS) [35,36], a nationally representative, cross-sectional survey conducted annually by the Centers for Disease Control and Prevention (CDC). The BRFSS collects health-related data from U.S. adults through telephone-based interviews, providing extensive information on health behaviors, chronic conditions, and healthcare access [37]. Data from 2016 to 2023 were used for this study, providing a broad temporal scope to examine trends and associations between sleep duration, mental health outcomes, and physical activity levels. To assess changes in sleep patterns over time, we conducted a trend analysis using BRFSS data from 2016, 2018, 2020, and 2022—the years in which the full sleep module was administered nationally. The analytic sample was derived following the flow diagram presented in Fig 1.

Fig 1. Study sample derivation flowchart – BRFSS 2016–2023.

Fig 1

Open in a new tab

This flow diagram illustrates the selection process for the final analytic sample used in the study examining associations between sleep duration and health outcomes. Starting with 3,911,129 adult respondents from the 2016–2023 Behavioral Risk Factor Surveillance System (BRFSS), 1,712,145 participants were excluded due to missing data on sleep duration. An additional 2,788,668 respondents were excluded for missing data on one or more of the four primary outcome variables: depression (n = 1,299,214), poor mental health days (n = 1,190,226), poor physical health days (n = 289,872), and general health status (n = 9,356). The resulting final analytic sample included 318,654 respondents with complete information. These were categorized into: short sleep (≤5 hours, n = 65,579; 19.7%), recommended sleep (6–8 hours, n = 224,164; 70.5%), and long sleep (≥9 hours, n = 31,257; 9.8%).

Primary explanatory variable: Sleep duration

Sleep duration was measured using the BRFSS survey item: “On average, how many hours of sleep do you get in a 24-hour period?” Respondents provided a numeric value from 1 to 24 hours. For analysis, sleep duration was categorized into three groups: Short sleep (≤5 hours) [38,39], Recommended sleep (6–8 hours) [38,40], and Long sleep (≥9 hours) [38–40]. This categorization was informed by recommendations from the American Academy of Sleep Medicine (AASM) [41,42] and the Centers for Disease Control and Prevention (CDC) [43,44], which endorse 7–9 hours of sleep as optimal for adult health. In all models, short sleep (≤5 hours) served as the reference group, allowing comparisons to both recommended and long sleep durations.

Outcome variables

The primary outcome variables in this study, drawn from the Behavioral Risk Factor Surveillance System (BRFSS), include self-reported depression diagnosis, poor mental health days, physically unhealthy days, and general health status.

Depression diagnosis

Depression status was assessed using the BRFSS chronic condition item: “(Ever told) (you had) a depressive disorder (including depression, major depression, dysthymia, or minor depression)?” Responses were coded as 1 = Yes and 2 = No. Individuals responding “Yes” were classified as having a self-reported history of depressive disorder. Non-substantive responses—including “Don’t know/Not sure” (code = 7) and “Refused” (code = 9)—were excluded from the analysis. This binary variable was analyzed as a categorical outcome to estimate the association between sleep duration and the likelihood of depression diagnosis.

Mentally unhealthy days

Mentally unhealthy days were assessed using the BRFSS core question: “Now thinking about your mental health, which includes stress, depression, and problems with emotions, for how many days during the past 30 days was your mental health not good?” Responses ranged from 0 to 30 days, reflecting the respondent’s self-reported number of days with poor mental health in the past month. Participants selecting “None” were coded as 0 days, while responses such as “Don’t know/Not sure” (77) and “Refused” (99) were treated as missing and excluded from analysis. This variable was modeled as a continuous outcome to quantify the monthly burden of mental distress among respondents.

Physically unhealthy days

Physically unhealthy days were assessed using the BRFSS core question: “Now thinking about your physical health, which includes physical illness and injury, for how many days during the past 30 days was your physical health not good?” Responses ranged from 0 to 30 days, reflecting the number of days respondents experienced poor physical health in the past month. Responses indicating “None” were coded as 0 days, while non-substantive responses (e.g., “Don’t know/Not sure” = 77, “Refused” = 99) were treated as missing and excluded from analysis. This variable was modeled as a continuous outcome to capture variations in physical health burden among participants.

General health status

Self-rated general health was measured using the standard BRFSS item: “Would you say that in general your health is excellent, very good, good, fair, or poor?” Responses were captured on a five-point Likert-type scale ranging from 1 (excellent) to 5 (poor). For analytical purposes, the variable was treated as a continuous measure, with higher scores indicating poorer self-perceived health. This subjective rating has been widely validated as a reliable indicator of overall health status and a strong predictor of morbidity and mortality in population-based research.

Covariates and confounder adjustment

To minimize confounding and improve the validity of estimated associations between sleep duration and health outcomes, the analysis adjusted for a comprehensive set of demographic and socioeconomic covariates sourced from the Behavioral Risk Factor Surveillance System (BRFSS). These covariates included:

  • Age group (18–34, 35–54, and ≥55 years)

  • Sex (male, female)

  • Body Mass Index (BMI) category (underweight, normal weight, overweight, obese)

  • Educational attainment (high school or less, some college, college graduate)

  • Household income (<$25,000, $25,000–$49,999, $50,000–$99,999, ≥ $100,000)

  • Marital status (married, previously married, never married)

  • Covid (pre-covid (2016–2019), post-covid (2020–2023))

These covariates were selected based on prior literature linking them to both sleep behaviors and health outcomes, and to control for potential confounding due to variations in socioeconomic status, lifestyle factors, and demographic characteristics. All models were also adjusted for state of residence and survey year to account for geographic and temporal variation in reporting patterns and policy contexts.

By including these covariates in the analytic models, the study aimed to more accurately isolate the independent association between sleep duration and the primary health outcomes of interest.

Propensity score adjustment (IPW)

To adjust for potential confounding in estimating the causal effect of sleep duration on various health outcomes, we employed inverse probability weighting (IPW) based on propensity scores. Although the Results section references PSM conceptually, our approach used IPW — a method within the propensity score framework that achieves covariate balance across treatment groups without the need for matched pairs. The rationale for using IPW was to retain the full sample while reducing selection bias due to non-random assignment of sleep duration categories, allowing for estimation of the average treatment effect (ATE).

Propensity scores were estimated via a multinomial logistic regression model, where sleep duration category (short sleep [≤5 hours], recommended sleep [6–8 hours], and long sleep [≥9 hours]) was modeled as a function of observed covariates. The following variables were included in the IPW model based on theoretical and empirical relevance to both sleep patterns and health outcomes: age group, sex (female), race/ethnicity, BMI category, educational attainment, income category, marital status, U.S. state of residence (stfip), and survey year (iyear). The syntax used for IPW estimation was: teffects ipw (Outcome) (sleep_cat i.age_group3 i.bmi_category i.RACE Femalei.Education i.income4cat i.marital_group3 COVID i.stfip i.iyear) [pweight = _llcpwt].

We applied BRFSS-provided probability weights (_llcpwt) in all models to ensure national representativeness. Balance diagnostics were conducted by comparing standardized mean differences (SMDs) for covariates before and after weighting. Covariate balance was considered acceptable if SMDs were reduced to below 0.1 following weighting. The use of IPW, rather than traditional matching (e.g., nearest-neighbor), was chosen to avoid sample loss and improve efficiency in a large, complex survey dataset. Robust standard errors were clustered at the state level to account for intra-state correlation.

Ethical considerations

This study used publicly available, de-identified data from the Behavioral Risk Factor Surveillance System (BRFSS), administered by the Centers for Disease Control and Prevention (CDC). As a secondary analysis of anonymized survey data, this research did not involve direct interaction with human participants and was exempt from Institutional Review Board (IRB) oversight. Informed consent was originally obtained by the CDC during data collection. Given the de-identified and publicly accessible nature of the dataset, no additional consent was required. The study complies with CDC data use policies and adheres to established ethical research standards.

Statistical analysis

To estimate the association between sleep duration and multiple health outcomes, we employed IPW using treatment-effects models within a counterfactual framework. This approach allows for the estimation of the ATE while accounting for observed confounders that influence both sleep duration and health outcomes.

Treatment and outcome models.

Sleep duration was operationalized categorically into three treatment groups:

  • Short sleep: ≤ 5 hours (reference group),

  • Recommended sleep: 6–8 hours,

  • Long sleep: ≥ 9 hours.

We implemented a multinomial logistic regression model to predict sleep category membership based on a rich set of covariates. These included:

  • Age group (18–34, 35–54, ≥55),

  • Sex (male, female),

  • Race/ethnicity,

  • Body Mass Index (BMI) category (underweight, normal, overweight, obese),

  • Education (high school or less, some college, college graduate),

  • Household income (<$25K, $25–49K, $50–99K, ≥ $100K),

  • Marital status (married, previously married, never married),

  • State of residence, and

  • Survey year.

Inverse probability weights were then generated from these estimated treatment probabilities and applied to model the expected values of each health outcome.

Estimation of effects

Each health outcome was modeled as the dependent variable in a weighted regression model:

  • Depression diagnosis (binary) was analyzed using a weighted linear probability model, allowing for straightforward interpretation of marginal effects as risk differences.

  • Poor mental health days, physically unhealthy days, and self-rated general health (continuous) were modeled using weighted mean regressions, assuming normally distributed residuals. These outcomes represent count or ordinal measures with relatively symmetric distributions, justifying use of linear estimators.

All analyses accounted for the complex survey design of BRFSS. We applied the provided BRFSS sampling weights (_llcpwt) to ensure population-level representativeness, and clustered standard errors at the state level to account for intra-state correlation. Fixed effects for state and year were included in all models to control for unobserved regional and temporal heterogeneity.

Statistical significance was assessed at the 0.05 alpha level, and robust standard errors were computed for all estimates. Analyses were conducted using Stata’s teffects ipw command, and convergence was confirmed through examination of estimation criteria.

Sensitivity analysis

To address reviewer concerns regarding the definition of sleep duration, we conducted a sensitivity analysis in which sleep was re-categorized according to American Academy of Sleep Medicine and Centers for Disease Control and Prevention guidance. In this re-specification, short sleep was defined as ≤6 hours, recommended sleep as 7–9 hours, and long sleep as ≥10 hours. We re-estimated all primary models using these revised sleep duration categories. The resulting predicted probabilities and regression estimates were summarized in supplementary tables (S1 File).

Results

Table 1 presents the sociodemographic and health-related characteristics of the study population stratified by sleep duration categories: short sleep (≤ 5 hours), recommended sleep (6–8 hours), and long sleep (≥9 hours). The total sample included 318,000 individuals. Of these, 65,579 (19.7%) reported sleeping fewer than 6 hours, 224,164 (70.5%) reported sleeping 6–8 hours, and 31,257 (9.8%) reported sleeping more than 8 hours per night.

Table 1. Sociodemographic and health characteristics of the study population by sleep duration category (BRFSS 2016-2023).

Variables (N = 318,000) Total Population Sleep (<5hrs) Sleep (6–8hrs) Sleep (>8hrs) p-value
(N = 318,000) (n = 65,579) (19.7%) (n = 224,164) (70.5%) (n = 31,257) (9.8%)
Age <0.001
 18–34Yr. 66, 151 (20.7%) 11,005 (17.6%) 50,032 (22.3%) 5,114 (16.0%)
 35–54Yr. 101,844 (31.9%) 23,394 (37.3%) 71,341 (31.8%) 7,109 (22.3%)
  > 55Yr. 150,891 (47.3%) 28,245 (45.1%) 102,919 (45.9%) 19,727 (61.7%)
Sex <0.001
 Male 117,044 (36.7%) 23,842 (38.1%) 82,307 (36.7%) 10,895 (34.1%)
 Female 201,610 (63.3%) 38,737 (61.9%) 141,857 (63.3%) 21,016 (65.9%)
BMI <0.001
 Underweight 6,558 (2.2%) 1,534 (2.7%) 4,246 (2.1%) 778 (2.6%)
 Normal 82,030 (28.0%) 13,680 (23.8%) 60,318 (29.2%) 8,032 (27.3%)
 Overweight 89,225 (30.4%) 16,547 (28.8%) 64,273 (31.1%) 8,405 (28.6%)
 Obese 115,661 (39.4%) 25,622 (44.7%) 77,815 (37.7%) 12,224 (41.5%)
Race/Ethnicity <0.001
 White 225,911 (76.7%) 40,505 (69.9%) 161,969 (78.3%) 23,437 (79.5%)
 Black 25,503 (8.7%) 7,078 (2.2%) 15,923 (7.7%) 2,502 (8.5%)
 Hispanic 16,988 (5.8%) 3,855 (6.7%) 11,834 (5.7%) 1,299 (4.4%)
 Other 26,002 (8.8%) 6,536 (11.3%) 17,210 (8.3%) 2,256 (7.7%)
Education <0.001
 High School or less 113,071 (35.6%) 28,076 (45.0%) 72,092 (32.2%) 12,903 (40.5%)
 Some college 96,360 (30.3%) 20,432 (32.7%) 66,223 (29.6%) 9,705 (30.5%)
 College graduate 108,587 (34.1%) 13,929 (22.3%) 85,409 (38.2%) 9,249 (29.0%)
Household Income ($) <0.001
  < 25K 50,917 (19.0%) 15,255 (29.2%) 29,201 (15.4%) 6,461 (24.7%)
 25-50K 62,187 (23.2%) 14,361 (27.5%) 40,729 (21.5%) 7,097 (27.1%)
 50-100K 67,421 (25.2%) 11,750 (22.5%) 49,242 (26.0%) 6,429 (24.6%)
  > 100K 87,182 (32.6%) 10,818 (20.7%) 70,163 (37.1%) 6,201 (23.7%)
Marital Status <0.001
 Married 149,465 (47.2%) 24,900 (40.1%) 111,210 (49.9%) 13,355 (42.1%)
 Previously Married 97,385 (30.8%) 23,707 (38.2%) 61,937 (27.8%) 11,741 (37.0%)
 Never Married 69,751 (22.0%) 13,520 (21.8%) 49,595 (22.3%) 6,636 (20.9%)
Covid Era <0.001
 pre-Covid 153,094 (51.5%) 31,444 (53.6%) 106,157 (50.8%) 15,493 (52.0%)
 post-Covid 144,230 (48.5%) 27,278 (46.5%) 102,675 (49.2%) 14,277 (48.0%)
Outcomes
 Depression 150,491 (47.6%) 37,020 (59.7%) 95,249 (42.8%) 18,222 (57.5%) <0.001
 Poor Mental Health days 12.7 + 10.7 17.4 + 11.1 11.2 + 10.2 14.3 + 11.0 <0.001
 Poor Physical Health days 12.3 + 11.2 16.4 + 11.5 10.7 + 10.6 15.3 + 11.5 <0.001
 Poor General health days 3.3 + 1.1 3.7 + 1.1 3.1 + 1.1 3.6 + 1.1 <0.001
 Excellent 17,331 (5.5%) 1,873 (3.0%) 14,216 (6.4%) 1,242 (3.9%) <0.001
 Very Good 67,436 (21.2%) 6,747 (10.8%) 56,084 (25.1%) 4,605 (14.5%) <0.001
 Good 100,540 (31.6%) 16,143 (25.9%) 75,848 (33.9%) 8,549 (26.8%) <0.001
 Fair 86,257 (27.1%) 21,750 (34.8%) 54,393 (24.3%) 10,114 (11.7%) <0.001
 Poor 46,614 (14.7%) 15,935 (25.5%) 23,335 (10.4%) 7,344 (23.1%) <0.001
Open in a new tab

Age distribution varied significantly across sleep categories. Participants aged 35–54 were most prevalent among short sleepers (37.3%), while long sleepers were predominantly older adults (>55 years, 61.7%). Among those reporting 6–8 hours of sleep, the majority were aged >55 years (45.9%), followed by those aged 35–54 years (31.8%) and 18–34 years (22.3%).

Sex differences were also significant. Males, who comprised 36.7% of the overall population, were slightly overrepresented in the short sleep group (38.1%) but underrepresented in the long sleep group (34.1%). In contrast, females were slightly overrepresented in the long sleep group (65.9%) compared to their overall proportion (63.3%). These differences, however, were marginal.

Body Mass Index (BMI) categories showed a significant relationship with sleep duration. The proportion of obese individuals was highest among short sleepers (44.7%) and lowest among those with 6–8 hours of sleep (37.7%). Normal weight was most common among those with recommended sleep (29.2%).

Race/ethnicity distribution significantly differed across sleep duration categories. Among individuals reporting short sleep (≤5 hours), 69.9% were White, 2.2% Black, 6.7% Hispanic, and 11.3% Other. In contrast, among those with recommended sleep (6–8 hours), 78.3% were White, 7.7% Black, 5.7% Hispanic, and 8.3% Other. Among long sleepers (>8 hours), 79.5% were White, 8.5% Black, 4.4% Hispanic, and 7.7% Other.

Educational attainment differed substantially by sleep duration. Individuals with a high school education or less—who made up 35.6% of the total population—were notably overrepresented among short sleepers (45.0%). In contrast, college graduates, whose baseline proportion was 34.1%, were slightly overrepresented among those with 6–8 hours of sleep (38.2%), suggesting healthier sleep patterns in this more educated group.

Income was positively associated with sleep duration. Nearly 29.2% of short sleepers reported incomes below $25,000 compared to 15.4% among those with 6–8 hours of sleep. High-income earners (> $100,000) were most concentrated among recommended sleepers (37.1%) and least common among short sleepers (20.7%).

Compared to their baseline proportions—married (47.2%), previously married (30.8%), and never married (22.0%)—previously married individuals were overrepresented among short sleepers (38.2%) but underrepresented among those with recommended sleep (27.8%). Married individuals were slightly overrepresented in the 6–8 hour sleep group (49.9%), suggesting more optimal sleep patterns in this group.

After the onset of the COVID-19 pandemic, subtle shifts in sleep patterns were observed. Compared to their overall representation (48.5%), post-COVID respondents were slightly overrepresented in the recommended sleep group (49.2%) but underrepresented among short sleepers (46.5%) and long sleepers (48.0%). These differences, however, were marginal at best.

Mental and physical health outcomes varied by sleep duration. The prevalence of depression was significantly higher among short sleepers (59.7%) and long sleepers (57.5%) compared to those with recommended sleep (42.8%) (p < 0.001). Additionally, poor mental health days were highest among short sleepers (mean 17.4 ± 11.1) and long sleepers (14.3 ± 11.0), compared to 11.2 ± 10.2 in the recommended group (p < 0.001). Similarly, physically unhealthy days followed the same pattern, with short sleepers reporting the highest mean (16.4 ± 11.5), followed by long sleepers (15.3 ± 11.5), and recommended sleepers (10.7 ± 10.6) (p < 0.001).

General health status differed by sleep duration (Table 1). Adults reporting recommended sleep had the most favorable profiles, with higher predicted probabilities of excellent (0.077) and very good health (0.258) than those with short (0.049 and 0.197, respectively) or long sleep (0.033 and 0.151, respectively). In contrast, the predicted probabilities of fair and poor health were lowest among those with recommended sleep (0.231 and 0.100) and highest among long sleepers (0.313 and 0.203). These patterns indicate that both short and long sleep durations are associated with worse self-reported health compared with recommended sleep (Table 1).

Sensitivity analyses

Re-analysis using revised sleep-duration categories (≤6 hours, 7–9 hours, ≥ 10 hours) yielded results consistent with the primary findings (S1 File). The U-shaped association between sleep duration and general health status persisted; both short and long sleep remained associated with higher predicted probabilities of fair and poor health relative to recommended sleep. Point estimates differed only slightly in magnitude, and overall inference was unchanged.

Temporal trends in sleep duration (2016–2022)

To assess temporal variation in self-reported sleep duration (Table 2), we conducted a trend analysis using BRFSS data from 2016, 2018, 2020, and 2022—the four years in which the complete sleep module was nationally administered. As shown in Table 2, the prevalence of short sleep duration (≤5 hours per night) modestly declined from 21.5% in 2016 (95% CI: 20.9–22.1%) to 19.8% in 2020 (95% CI: 19.0–20.5%), before increasing slightly to 20.5% in 2022 (95% CI: 19.9–21.1%). In contrast, the proportion of adults reporting recommended sleep duration (6–8 hours) gradually increased across survey waves, rising from 69.5% in 2016 to 71.0% in 2022. Meanwhile, long sleep duration (≥9 hours) remained relatively stable, ranging between 8.5% and 10.0% across the same period. These patterns suggest a modest but consistent shift toward healthier sleep durations over time, providing further context for interpreting the health associations reported in this study (S1 Fig 1 in S1 File).

Table 2. Weighted prevalence of self-reported sleep duration categories across survey years in the United States, BRFSS 2016–2022.

Sleep Duration Year Proportion Std. Err. 95% CI
Short Sleep 2016 0.215 0.003 0.209–0.221
2018 0.213 0.003 0.207–0.220
2020 0.198 0.004 0.190–0.205
2022 0.205 0.003 0.199–0.211
Recommended 2016 0.695 0.003 0.689–0.702
2018 0.698 0.003 0.691–0.705
2020 0.702 0.004 0.694–0.711
2022 0.710 0.003 0.704–0.717
Long Sleep 2016 0.090 0.002 0.086–0.094
2018 0.088 0.002 0.085–0.092
2020 0.100 0.003 0.094–0.106
2022 0.085 0.002 0.081–0.089
Open in a new tab

Proportions reflect weighted estimates based on national sampling weights from the Behavioral Risk Factor Surveillance System (BRFSS). Sleep duration was classified as: Short Sleep (≤5 hours), Recommended Sleep (6–8 hours), and Long Sleep (≥9 hours). Estimates are restricted to survey years in which the complete sleep module was administered nationally (even years only).

Association between sleep duration and incidence of depression

Table 3 presents the estimated Average Treatment Effects (ATE) of sleep duration on the incidence of self-reported depression, using inverse probability weighting (IPW) models. Recommended sleep (6–8 hours) served as the reference group for comparison.

Table 3. Average treatment effects of sleep duration on incidence of depression.

Outcome Sleep Categories Coef. Std. Err. z-value p-value 95% CI
Depression Short Sleep 0.141 0.002 90.20 <0.001 0.138 0.144
Long Sleep 0.129 0.002 65.99 0.521 0.125 0.133
Recommended 0.395 0.0001 581.66 <0.001 0.394 0.397
Open in a new tab

This table presents the estimated Average Treatment Effects (ATE) of sleep duration categories on the incidence of self-reported depression, using inverse probability weighting (IPW) models. The reference group is individuals reporting recommended sleep (6–8 hours per night). Coefficients represent the estimated difference in the probability of depression relative to the reference group. Positive values indicate a higher probability of depression compared to recommended sleepers. The model adjusts for demographic and socioeconomic covariates, including age group, sex, race, BMI category, income, education, marital status, state, and survey year. Robust standard errors were used for statistical inference. Reported statistics include the coefficient (Coef.), standard error (Std. Err.), z-value, p-value, and 95% confidence interval (CI).

Participants with short sleep duration (≤5 hours) had a significantly higher probability of depression, with an ATE coefficient of 0.141 (Std. Err. = 0.002, z = 90.20, p < 0.001; 95% CI: 0.138–0.144). This indicates a 14.1 percentage-point increase in depression incidence compared to individuals who sleep the recommended number of hours.

Similarly, long sleep duration (≥9 hours) was also associated with a significantly elevated likelihood of depression, with an ATE coefficient of 0.129 (Std. Err. = 0.002, z = 65.99, p < 0.001; 95% CI: 0.125–0.133), representing a 12.9 percentage-point increase compared to the recommended group.

Individuals with recommended sleep had a baseline depression probability of 39.5% (Coef. = 0.395, Std. Err. = 0.0001, z = 581.66, p < 0.001; 95% CI: 0.394–0.397), which was markedly lower than that of both short and long sleepers.

These findings demonstrate a clear U-shaped association between sleep duration and depression risk, with both insufficient and excessive sleep durations linked to significantly higher odds of depression compared to the optimal 6–8 hours (S1 Fig 2 in S1 File).

Association between sleep duration and poor mental health days

Table 4 presents the ATE of sleep duration on the number of poor mental health days reported in the past 30 days, using IPW models. Individuals with recommended sleep (6–8 hours) served as the reference group.

Table 4. Average treatment effects of sleep duration on incidence of poor mental health days.

Outcome Sleep Categories Coef. Std. Err. z-value p-value 95% CI
Depression Short Sleep 5.319 0.035 150.51 <0.001 5.250 5.389
Long Sleep 2.456 0.043 56.72 <0.001 2.371 2.541
Recommended 12.129 0.015 800.76 <0.001 12.100 12.159
Open in a new tab

This table displays the estimated Average Treatment Effects (ATE) of sleep duration categories on the number of self-reported poor mental health days in the past 30 days, calculated using inverse probability weighting (IPW) models. The reference group is individuals reporting recommended sleep (6–8 hours per night). Coefficients represent the estimated mean difference in poor mental health days relative to the reference group. Positive values indicate a higher burden of poor mental health compared to recommended sleepers. The model adjusts for key demographic and socioeconomic covariates, including age group, sex, race/ethnicity, BMI category, income, education, marital status, state, and survey year. Reported values include the coefficient (Coef.), standard error (Std. Err.), z-value, p-value, and 95% confidence interval (CI).

Compared to the reference, individuals with short sleep duration (≤5 hours) reported a substantially higher number of poor mental health days, with an ATE coefficient of 5.319 (Std. Err. = 0.035, z = 150.51, p < 0.001; 95% CI: 5.250–5.389). Similarly, those with long sleep duration (≥9 hours) also reported significantly more poor mental health days, though to a lesser degree, with a coefficient of 2.456 (Std. Err. = 0.043, z = 56.72, p < 0.001; 95% CI: 2.371–2.541).

Individuals in the recommended sleep group reported a baseline average of 12.129 poor mental health days per month (Std. Err. = 0.015, z = 800.76, p < 0.001; 95% CI: 12.100–12.159), which was significantly lower than that of short and long sleepers.

These results highlight a clear inverse relationship between sleep duration and mental health burden. Both insufficient (≤5 hours) and excessive (≥9 hours) sleep durations are associated with an increased number of poor mental health days, reinforcing the mental health benefits of maintaining a recommended sleep schedule.

Association between sleep duration and poor physical health days

Table 5 presents the estimated ATE of sleep duration categories on the number of self-reported physically unhealthy days over the past 30 days, derived using the IPW models. The reference group in this analysis is individuals with recommended sleep duration (6–8 hours per night).

Table 5. Average treatment effects of sleep duration on incidence of poor physical health days.

Outcome Sleep Categories Coef. Std. Err. z-value p-value 95% CI
Poor Physical Health days Short Sleep 4.379 0.033 132.91 <0.001 4.314 4.443
Long Sleep 3.194 0.041 77.02 <0.001 3.113 3.276
Recommended 10.723 0.015 729.34 <0.001 10.694 10.752
Open in a new tab

This table reports the estimated Average Treatment Effects (ATE) of sleep duration categories on the number of self-reported physically unhealthy days in the past 30 days, using inverse probability weighting (IPW) models. The reference group is individuals with recommended sleep duration (6–8 hours per night). Coefficients represent the estimated mean difference in physically unhealthy days compared to the recommended sleep group. Positive coefficients indicate a greater number of poor physical health days relative to recommended sleepers. The model was adjusted for demographic and socioeconomic variables, including age, sex, race/ethnicity, BMI, education level, income, marital status, geographic region, and survey year. Reported values include the coefficient (Coef.), standard error (Std. Err.), z-value, p-value, and 95% confidence interval (CI).

Compared to the recommended group, individuals with short sleep duration (≤5 hours) reported 4.379 additional physically unhealthy days per month (Std. Err. = 0.033, z = 132.91, p < 0.001; 95% CI: 4.314–4.443), indicating a substantial physical health burden associated with insufficient sleep. Similarly, those with long sleep duration (≥9 hours) also experienced significantly more physically unhealthy days than recommended sleepers, with an estimated increase of 3.194 days (Std. Err. = 0.041, z = 77.02, p < 0.001; 95% CI: 3.113–3.276).

Individuals in the recommended sleep group reported an average of 10.723 physically unhealthy days per month (Std. Err. = 0.015, z = 729.34, p < 0.001; 95% CI: 10.694–10.752), which was lower than both short and long sleep groups.

These findings suggest that deviations from recommended sleep—either too little or too much—are associated with a higher burden of physical health issues. The largest adverse effect was observed among short sleepers, reinforcing the importance of maintaining a sleep duration of 6–8 hours for optimal physical health.

Association between sleep duration and general health

Adjusted predictive margins for general health status varied substantially by sleep duration category (Table 6). Individuals reporting recommended sleep duration demonstrated the most favorable health profile overall. The adjusted probability of reporting excellent health was highest among adults with recommended sleep (0.077; 95% CI, 0.075–0.080), and was markedly lower among those with short sleep (0.049; 95% CI, 0.047–0.051) and long sleep (0.033; 95% CI, 0.031–0.036). A similar pattern was observed for reports of very good health, which were most common among individuals with recommended sleep (0.258; 95% CI, 0.254–0.261), declining among those with short sleep (0.197; 95% CI, 0.194–0.201) and long sleep (0.151; 95% CI, 0.143–0.158).

Table 6. Adjusted predicted probabilities of general health categories by sleep duration.

GENHLTH Sleep Category Predicted Probability Std. Err. z-value 95% CI
Excellent Recommended 0.077 0.001 64.41 0.075–0.080
Excellent Short 0.049 0.001 53.48 0.047–0.051
Excellent Long 0.033 0.001 26.68 0.031–0.036
Very Good Recommended 0.258 0.002 132.92 0.254–0.261
Very Good Short 0.197 0.002 112.67 0.194–0.201
Very Good Long 0.151 0.004 40.17 0.143–0.158
Good Recommended 0.335 0.002 176.22 0.331–0.338
Good Short 0.325 0.002 177.88 0.321–0.328
Good Long 0.300 0.003 99.75 0.294–0.306
Fair Recommended 0.231 0.002 136.64 0.227–0.234
Fair Short 0.280 0.002 145.79 0.276–0.284
Fair Long 0.313 0.003 102.38 0.307–0.319
Poor Recommended 0.100 0.001 79.22 0.097–0.102
Poor Short 0.149 0.002 95.67 0.146–0.152
Poor Long 0.203 0.005 40.52 0.194–0.213
Open in a new tab

Table presents adjusted predicted probabilities for each general health category by sleep duration derived from IPW models.

In contrast, the probability of reporting fair or poor health increased notably with departure from recommended sleep duration. Adults with long sleep demonstrated the highest predicted probability of fair health (0.313; 95% CI, 0.307–0.319), followed by those with short sleep (0.280; 95% CI, 0.276–0.284), and recommended sleepers (0.231; 95% CI, 0.227–0.234). This gradient was more pronounced for poor health: long sleepers exhibited a predicted probability of 0.203 (95% CI, 0.194–0.213), followed by short sleepers (0.149; 95% CI, 0.146–0.152) and recommended sleepers (0.100; 95% CI, s0.097–0.102).

Patterns for good health status were comparatively similar across groups, although recommended and short sleepers demonstrated slightly higher probabilities (0.335; 95% CI, 0.331–0.338 and 0.325; 95% CI, 0.321–0.328, respectively) than long sleepers (0.300; 95% CI, 0.294–0.306). Collectively, these findings indicate a graded, U-shaped association, wherein both short and long sleep are associated with less favorable self-rated general health compared with recommended sleep duration.

Discussion

In this large, nationally representative sample of over 300,000 U.S. adults, we observed significant associations between sleep duration and key health outcomes, including depression, poor mental health days, poor physical health days, and self-rated general health. Using IPW to adjust for confounding, we found that both short sleep (≤5 hours) and long sleep (≥9 hours) were independently associated with worse health outcomes compared to individuals who reported recommended sleep duration (6–8 hours).

Our findings reinforce previous literature that short sleep duration is a strong predictor of adverse mental and physical health outcomes [21,45–47]. Adults reporting ≤5 hours of sleep had significantly higher incidence of depression (+14.1%, 95% CI: 13.8%–14.4%), longer duration of poor mental health (5.3 days), and poor physical health (4.4 days) in the past month. Furthermore, 44.8% of short sleepers rated their general health as fair or poor. These outcomes are in line with existing evidence that sleep deprivation can dysregulate emotional [48,49], cognitive [48,50], and immune [48,51,52] functioning.

Interestingly, and consistent with emerging literature [53,54], we also found that long sleep duration (≥9 hours) was associated with higher burden of depression and poorer general health compared to the reference group. Though traditionally underexplored, prolonged sleep has increasingly been implicated in poor health outcomes, potentially reflecting underlying undiagnosed health conditions or behavioral and social determinants of health [55–57]. For example, those in the ≥ 9-hour group had a 12.9% increase in depression incidence and reported worse mental and physical health than the 6–8-hour group, albeit with slightly lower burden than short sleepers. This U-shaped association highlights that both insufficient and excessive sleep are detrimental to health.

Importantly, we accounted for the potential influence of the COVID-19 pandemic by including survey year and state fixed effects in our analysis. The pandemic period did not significantly modify the observed associations, suggesting that the link between sleep duration and health outcomes persisted regardless of broader societal disruptions.

Moreover, our findings regarding the number of poor mental health days are consistent with national trends [58]. A recent analysis by Udupa et al. (2023) reported an increase in the burden of poor mental health days among U.S. adults, with notable disparities by sex, age, and income [59]. Our study corroborates these trends and extends the evidence by linking these poor mental health days directly to sleep duration [60], further supporting calls for prioritizing sleep health in population-level interventions.

The inclusion of general health as an outcome in this study strengthens its contribution to literature [61–63]. The strong association between suboptimal sleep and self-rated poor general health reflects sleep’s critical role in broader wellness perceptions and functional status. These findings emphasize that interventions targeting sleep hygiene may not only mitigate psychiatric and physical morbidity but also improve perceived quality of life.

Public health implications

Our findings carry substantial public health and policy implications, underscoring the critical role of optimal sleep duration, neither too short nor too long, in promoting mental, physical, and general health. The results reveal a U-shaped association where both short sleep (≤5 hours) and long sleep (≥9 hours) are linked with increased incidence of depression, more mentally and physically unhealthy days, and worse self-rated general health. This suggests that public health messaging should shift from simply promoting “more sleep” to advocating for an optimal range, particularly 6–8 hours, as supported by organizations such as the CDC and the American Academy of Sleep Medicine.

Given the high prevalence of short sleep in the U.S., especially among working adults and socioeconomically disadvantaged populations, routine screening for sleep patterns should be integrated into primary care and behavioral health visits. In addition, the growing body of evidence, reinforced by our findings, points to a need to assess for long sleep as a potential marker of underlying health risk, rather than dismissing it as benign.

Policy interventions could include public education campaigns promoting the 6–8-hour sleep guideline, employer-based initiatives that encourage healthy work–life balance and limit excessive shift work, and insurance coverage for evidence-based sleep interventions like cognitive-behavioral therapy for insomnia (CBT-I). Technology-assisted tools such as sleep tracking apps and behavioral nudges via telehealth may help scale these efforts.

Finally, our findings are consistent with national estimates, including data from Udupa et al. (2023), showing an increase in poor mental health days among U.S. adults. Addressing sleep as a modifiable behavioral target could be a scalable and cost-effective strategy to help reverse these mental health trends, particularly when tailored to high-risk subgroups.

Strengths and limitations

This study possesses several methodological strengths that enhance its validity, robustness, and generalizability. Leveraging a large, nationally representative sample from the Behavioral Risk Factor Surveillance System (BRFSS), our findings are applicable to diverse adult populations across the United States. The categorization of sleep duration into three distinct groups (short: ≤ 5 hours, recommended: 6–8 hours, and long: ≥ 9 hours) enabled a nuanced assessment of sleep-health relationships beyond the conventional binary classification. Importantly, the application of IPW allowed for adjustment of key sociodemographic and behavioral confounders and strengthened the causal interpretation of the average treatment effects of sleep duration on depression, self-rated general health, and mentally and physically unhealthy days. We also accounted for state and survey year fixed effects, including adjustment for the COVID-19 period, ensuring temporal and regional influences were considered.

Nevertheless, several limitations warrant consideration. First, sleep duration was self-reported, introducing the possibility of recall bias or misclassification. Recent evidence suggests that a proportion of individuals who self-report long sleep may actually obtain insufficient sleep when measured objectively, indicating potential misclassification among long sleepers and the need for cautious interpretation of this category. Objective sleep assessments, such as actigraphy or polysomnography, would yield more precise measures but were unavailable in this dataset. Second, the cross-sectional design limits causal inference and cannot fully address the bidirectionality between sleep and health outcomes. For instance, while short or long sleep durations may increase the risk of depression or poorer physical health, these conditions may also disrupt sleep patterns, leading to cyclical feedback loops. Third, while we adjusted for a broad array of covariates, residual confounding by unmeasured variables, such as undiagnosed sleep disorders, chronic stress, or genetic predispositions, cannot be entirely ruled out. Lastly, the BRFSS sleep module was administered only in selected years, potentially affecting temporal trend analysis despite our efforts to stratify by survey year.

Future research should employ longitudinal and experimental designs to clarify causality and explore potential biological and behavioral mechanisms underlying the associations observed. Such work may also identify vulnerable subpopulations who would benefit most from targeted sleep health interventions.

Conclusion

This study provides robust evidence that both short (≤5 hours) and long (≥9 hours) sleep durations are significantly associated with adverse health outcomes compared to the recommended 6–8 hours of sleep. Individuals in the short and long sleep categories exhibited a higher incidence of depression, more days of poor mental and physical health, and worse overall self-rated general health. These findings reinforce the importance of maintaining optimal sleep duration as a key determinant of population mental and physical well-being. Public health efforts should prioritize strategies to promote sleep health, identify at-risk individuals, and address underlying factors contributing to inadequate or excessive sleep. Tailored interventions, health education campaigns, and clinical screening for sleep patterns may collectively improve individual outcomes and reduce the burden of sleep-related health disparities across the United States.

Supporting information

S1 File. Supplementary figures and tables.

Contains Supplementary Figure 1 (Trends in Sleep Duration Categories), Supplementary Figure 2 (Prevalence of Depression by Sleep Duration), and Supplementary Tables 1–5 describing baseline characteristics, treatment effects, mentally and physically unhealthy days, and predicted general health probabilities.

(DOCX)

pone.0321347.s001.docx (183.7KB, docx)

Data Availability

All relevant data are within the manuscript and its supporting information files.

Funding Statement

This project was supported (in part) by the National Institute on Minority Health and Health Disparities of the National Institutes of Health under Award Number 2U54MD007597 to W.S. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.” For clarification, although Dr. William Southerland (W.S.) is not an author on this manuscript, the second author, Dr. Oluwasegun A. Akinyemi, is supported under his NIH center grant. The Article Processing Charge (APC) for this manuscript is being covered through that grant, which is why the funding must be disclosed under Dr. Southerland’s award number.

References

Decision Letter 0

Edward Chiyaka

6 Jun 2025

Dear Dr. Ndebele-Ngwenya,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

==============================

In addition to the comments raised by the reviewers, please also address the following in your revision: 

  • Please ensure that all citations adhere to the formatting style specified in the journal’s author guidelines especially in text citations.

  • In the Methods section, authors stated: "Data from 2016 to 2023 were used for this study, ensuring a broad temporal scope to assess trends and associations between sleep duration." However, the analysis does not include any temporal stratification or trend assessment to justify the use of an eight-year dataset. The authors should either conduct appropriate time-series or stratified analyses to examine changes in sleep duration or related variables over time, or clarify why the temporal element was not explored despite the dataset's extended duration. A visual or statistical analysis of trends would significantly enhance the contribution of the study.

  • While the Results section references the use of propensity score matching (PSM), the methodology is not adequately described in the Methods section. A clear explanation of the PSM process should be added, including: the rationale for using PSM, variables included in the matching algorithm, matching technique employed (e.g., nearest neighbor, caliper, 1:1 matching) and balance diagnostics used to assess matching quality. This information is essential for replicability and for readers to evaluate the robustness of the causal inferences drawn from the matched comparisons.

  • In the Results section, Table 1 reports a total sample size (N) of 163,724, which appears inconsistent compared to the matched samples reported in the same table. Please clarify how this N was derived and ensure internal consistency across the manuscript. If Table 1 is meant to reflect the unmatched sample, explicitly label it as such and explain any exclusions or data processing steps that led to this specific N.

  • Include line numbers.

==============================

Please submit your revised manuscript by Jul 21 2025 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org . When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

  • A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

  • A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

  • An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols . Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols .

We look forward to receiving your revised manuscript.

Kind regards,

Edward Chiyaka, Ph.D., MSc

Academic Editor

PLOS ONE

Journal Requirements:

1. When submitting your revision, we need you to address these additional requirements.-->--> -->-->Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at -->-->https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and -->-->https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf-->--> -->-->2. We note that your Data Availability Statement is currently as follows: All relevant data are within the manuscript and its supporting information files.-->--> -->-->Please confirm at this time whether or not your submission contains all raw data required to replicate the results of your study. Authors must share the “minimal data set” for their submission. PLOS defines the minimal data set to consist of the data required to replicate all study findings reported in the article, as well as related metadata and methods (https://journals.plos.org/plosone/s/data-availability#loc-minimal-data-set-definition).-->--> -->-->For example, authors should submit the following data:-->--> -->-->- The values behind the means, standard deviations and other measures reported;-->-->- The values used to build graphs;-->-->- The points extracted from images for analysis.-->--> -->-->Authors do not need to submit their entire data set if only a portion of the data was used in the reported study.-->--> -->-->If your submission does not contain these data, please either upload them as Supporting Information files or deposit them to a stable, public repository and provide us with the relevant URLs, DOIs, or accession numbers. For a list of recommended repositories, please see https://journals.plos.org/plosone/s/recommended-repositories.-->--> -->-->If there are ethical or legal restrictions on sharing a de-identified data set, please explain them in detail (e.g., data contain potentially sensitive information, data are owned by a third-party organization, etc.) and who has imposed them (e.g., an ethics committee). Please also provide contact information for a data access committee, ethics committee, or other institutional body to which data requests may be sent. If data are owned by a third party, please indicate how others may request data access.-->--> -->-->3. When completing the data availability statement of the submission form, you indicated that you will make your data available on acceptance. We strongly recommend all authors decide on a data sharing plan before acceptance, as the process can be lengthy and hold up publication timelines. Please note that, though access restrictions are acceptable now, your entire data will need to be made freely accessible if your manuscript is accepted for publication. This policy applies to all data except where public deposition would breach compliance with the protocol approved by your research ethics board. If you are unable to adhere to our open data policy, please kindly revise your statement to explain your reasoning and we will seek the editor's input on an exemption. Please be assured that, once you have provided your new statement, the assessment of your exemption will not hold up the peer review process.-->--> -->-->4. PLOS requires an ORCID iD for the corresponding author in Editorial Manager on papers submitted after December 6th, 2016. Please ensure that you have an ORCID iD and that it is validated in Editorial Manager. To do this, go to ‘Update my Information’ (in the upper left-hand corner of the main menu), and click on the Fetch/Validate link next to the ORCID field. This will take you to the ORCID site and allow you to create a new iD or authenticate a pre-existing iD in Editorial Manager.-->?>

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

Reviewer #1: Yes

Reviewer #2: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously? -->?>

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available??>

The PLOS Data policy

Reviewer #1: Yes

Reviewer #2: No

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English??>

Reviewer #1: Yes

Reviewer #2: Yes

**********

Reviewer #1:  This is an automated report for PONE-D-25-11133. This report was solicited by the PLOS One editorial team and provided by ScreenIT.

ScreenIT is an independent group of scientists developing automated tools that analyze academic papers. A set of automated tools screened your submitted manuscript and provided the report below. Each tool was created by your academic colleagues with the goal of helping authors. The tools look for factors that are important for transparency, rigor and reproducibility, and we hope that the report might help you to improve reporting in your manuscript. Within the report you will find links to more information about the items that the tools check. These links include helpful papers, websites, or videos that explain why the item is important. While our screening tools aim to improve and maintain quality standards they may, on occasion, miss nuances specific to your study type or flag something incorrectly. Each tool has limitations that are described on the ScreenIT website. The tools screen the main file for the paper; they are not able to screen supplements stored in separate files. Please note that the Academic Editor had access to these comments while making a decision on your manuscript. The Academic Editor may ask that issues flagged in this report be addressed. If you would like to provide feedback on the ScreenIT tool, please email the team at ScreenIt@bih-charite.de. If you have questions or concerns about the review process, please contact the PLOS One office at plosone@plos.org.

Reviewer #2:  The paper examined the impact of short sleep duration on depression, self-reported poor mental health days, and poor physical health days using large-scale data from the Behavioral Risk Factor Surveillance System.

The paper is well written and results offer robust evidence of the detrimental impact of short sleep on physical and mental health. I recommend the authors address the following comments to make the paper more comprehensive and increase readability.

1) The period in which data was collected includes the COVID pandemic. Can the authors address how their findings were impacted by this time period and perhaps add an analysis that excludes these time period which likely biases the results?

2) The discussion should add more information regarding the prevalence of main study outcomes with respect to the current survey (e.g. does the number of poor mental health days in the US as measured by other surveys match the current results? See Udupa et al. 2023)

3) It would be very informative to add the impact of the co-variates (sex ,age etc.) on the outcome in each of the models (can be done as supplementary).

4) The analysis of physical health is a bit confusing - the survey asks about general physical activity (being active or not) yet the result present poor physical health days. Can the authors explain this difference?

5) Please add the exact phrasing used for the self-reported sleep duration question. Also, please add the number of participants to the study abstract.

6) Last, the study does not include any visual elements other than tables – can the authors find a way to visualize the results?

**********

what does this mean? ). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy

Reviewer #1: No

Reviewer #2: No

**********

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/ . PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org . Please note that Supporting Information files do not need this step.

Attachment

Submitted filename: report_10.1101+2025.03.06.25323545.pdf

pone.0321347.s002.pdf (588.9KB, pdf)
PLoS One. 2026 Jan 14;21(1):e0321347. doi: 10.1371/journal.pone.0321347.r002

Author response to Decision Letter 1

11 Aug 2025

We sincerely thank the reviewers for their thoughtful comments, detailed reviews, and valuable suggestions. We have implemented all the requested revisions and have outlined them below. In addition, we made a few other changes that we believe further enhance the clarity, relevance, and overall quality of the manuscript.

Summary of Major Revisions:

1. New Analytical Approach:

We revised the analytic design from a binary classification of sleep duration (short vs. recommended) to a three-category classification:

o Short Sleep (≤5 hours),

o Recommended Sleep (6–8 hours), and

o Long Sleep (≥9 hours).

This allows for a more nuanced evaluation of the association between sleep and health outcomes.

2. Expanded Health Outcomes (Now Four):

We now include General Health Status (fair/poor vs. good/very good/excellent) as a fourth outcome in addition to:

o Depression,

o Poor mental health days, and

o Poor physical health days.

3. Revised Sample Size and Flow Diagram:

The analytic sample is now N = 318,000. A detailed study flow diagram (Figure 1) has been added to illustrate inclusion/exclusion criteria and missing data handling.

4. Precise Variable Definitions:

We now clearly provide the exact BRFSS questions and operational definitions used to derive each variable and outcome, as requested.

5. Revised Manuscript Title:

The title has been updated to:

“Associations Between Sleep Duration and Depression, Mental Health, Physical Health, and General Health in U.S. Adults: A Population-Based Study”

to better reflect the revised scope and objectives.

Point-by-Point Responses

Editorial Comments

1. Citation Formatting

“Please ensure that all citations adhere to the formatting style specified in the journal’s author guidelines especially in-text citations.”

Response:

All references have been reformatted in accordance with PLOS ONE citation guidelines. In-text citations now use the correct numerical format.

2. Temporal Trend Justification

“The analysis does not include any temporal stratification or trend assessment... A visual or statistical analysis of trends would significantly enhance the contribution of the study.”

Response:

We agree with this recommendation and have now included a temporal trend analysis to examine the changes in sleep duration and health outcomes across 2016–2023. A new section has been added to the Methods and Results, and the table is available in Table 2 and Supplementary Figure .

3. Propensity Score Matching (IPW) Methodology Description

“The methodology is not adequately described… include rationale, variables, matching technique, and diagnostics.”

Response:

We have added a new subsection to the Methods section titled “Propensity Score Matching (Inverse Probability Weighting – IPW)”. This subsection outlines the rationale for using IPW to reduce selection bias inherent in observational data. We detail the covariates included in the propensity score model (age, sex, race, education, BMI, income, etc.), the matching algorithm used (1:1 nearest neighbor without caliper), and the approach for assessing covariate balance (standardized mean differences). In addition, we have provided the full Stata syntax used for implementing the IPW procedure to enhance transparency and reproducibility.

4. Sample Size Discrepancy (Table 1 vs. Text)

“Clarify how the N = 163,724 was derived and ensure internal consistency.”

Response:

We have clarified that Table 1 originally reflected the matched sample. It has now been updated to explicitly indicate whether the characteristics pertain to the matched or unmatched population, and we have included the full analytic sample size of 318,000. Additionally, a study flowchart has been added to illustrate the sample derivation and exclusion criteria. A comprehensive description of all study variables, stratified by sleep category, is now provided. As we employed inverse probability weighting (IPW) rather than traditional propensity score matching, we did not construct separate matched and control groups; instead, the entire weighted sample was used for analysis.

5. Add Line Numbers

“Include line numbers.”

Response:

Line numbers have been included in both the clean and tracked-changes versions of the revised manuscript.

Reviewer #2 Comments

Comment 1: COVID-19 Bias Potential

“The period in which data was collected includes the COVID pandemic. Can the authors address how their findings were impacted…?”

Response:

We conducted a stratified analysis comparing the pre-pandemic (2016–2019) and pandemic (2020–2023) periods to assess potential variation in associations. Results remained consistent across both timeframes. Additionally, we included year and state fixed effects in our models to account for temporal and geographic variability. A binary variable for the COVID-19 pandemic period (“COVID”) was also created and included in the models to control for the specific effects of the pandemic years. These adjustments are detailed in the revised Methodology and Results section.

Comment 2: Benchmark Outcomes with National Survey (e.g., Udupa et al. 2023)

“Add more info regarding the prevalence of outcomes compared to other surveys…”

Response:

We have expanded the Discussion section to include comparisons of our findings with Udupa et al. (2023) and other national BRFSS-based studies. Our estimates are consistent with prior literature, reinforcing external validity.

Comment 3: Impact of Covariates

“Add the impact of covariates (sex, age, etc.) in the models – can be done in supplementary.”

Response:

We appreciate the reviewer’s insightful suggestion. While we agree that examining the independent effects of covariates such as sex, age, education, income, and race on each health outcome would be valuable, we have chosen not to include these detailed estimates in the current manuscript for several reasons.

First, incorporating these results across all four outcome variables would significantly expand the number of tables (we currently present six tables), potentially overwhelming readers and detracting from the study’s main focus—examining associations between sleep duration and health outcomes.

Second, the purpose of this study was to evaluate sleep duration as the primary exposure of interest, while adjusting for key covariates. Including full covariate-specific estimates may shift focus away from our central objective.

That said, we fully acknowledge the importance of these covariate effects and have adjusted for all suggested variables in our multivariable models. To address this important line of inquiry, we are actively planning a separate follow-up study that will comprehensively evaluate the independent impact of sociodemographic covariates on depression, mental health, physical health, and general health status using the BRFSS dataset.

We hope this rationale is acceptable and appreciate your understanding.

Comment 4: Clarify Use of ‘Poor Physical Health Days’ vs. Physical Activity

“The survey asks about physical activity, yet you report poor physical health days.”

Response:

We clarified that our outcome was BRFSS-reported number of poor physical health days, not physical inactivity. The specific question wording is now quoted in the Methods section.

Comment 5: Add Exact Question Wording & Abstract Sample Size

“Please add the exact phrasing used for sleep duration… and participant number in the abstract.”

Response:

We have added the precise BRFSS item:

“On average, how many hours of sleep do you get in a 24-hour period?”

to the Methods section, and the final sample size (N = 318,000) is now included in the Abstract.

Comment 6: Add Visualizations Beyond Tables

“Can the authors find a way to visualize the results?”

Response:

We have included the following visualizations to support our findings and enhance interpretability:

• Supplementary Figure 1: A line graph illustrating the weighted proportion of each sleep duration category (Short, Recommended, Long) among U.S. adults across BRFSS survey years 2016–2023.

• Supplementary Figure 2: A bar chart displaying the estimated prevalence of self-reported depression across sleep categories, derived from inverse probability weighting (IPW) models.

These figures highlight both temporal trends in sleep patterns and the adjusted associations between sleep duration and depression.

• A flow diagram showing the analytic sample derivation (Figure 1).

Additional ScreenIT & Editorial Compliance Updates

• Study Flow Chart: Added (Supplementary Figure 1).

• Inclusion/Exclusion Criteria: Individuals were excluded if they had missing data on sleep duration, any study outcomes, or covariates. Additionally, participants under the age of 18 were excluded from the analysis.

• Sex as a Variable: Included in descriptive stats and multivariable models.

• Open Data: Minimal dataset will be provided upon acceptance. De-identified BRFSS data and codebook are publicly available.

• ORCID ID for Corresponding Author: Updated in Editorial Manager profile.

________________________________________

We hope the comprehensive revisions and additions have addressed all comments satisfactorily. We are grateful for the opportunity to revise and improve our work and look forward to your favorable consideration.

Sincerely,

Oluwasegun Akinyemi, MD, MSc. PhD

Senior Research Fellow

Howard University College of Medicine

Oluwasegun.akinyemi@howard.edu

(309) 255-8284

Attachment

Submitted filename: Response to Reviewers_Sleep7232025.docx

pone.0321347.s004.docx (22.4KB, docx)

Decision Letter 1

Valentina Alfonsi

30 Oct 2025

Associations Between Sleep Duration and Depression, Mental Health, Physical Health, and General Health in U.S. Adults: A Population-Based Study

PONE-D-25-11133R1

Dear Dr. Ndebele-Ngwenya,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice will be generated when your article is formally accepted. Please note, if your institution has a publishing partnership with PLOS and your article meets the relevant criteria, all or part of your publication costs will be covered. Please make sure your user information is up-to-date by logging into Editorial Manager at Editorial Manager®  and clicking the ‘Update My Information' link at the top of the page. For questions related to billing, please contact billing support .

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Valentina Alfonsi, Ph.D.

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

Reviewer #2: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions??>

Reviewer #2: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously? -->?>

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available??>

The PLOS Data policy

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English??>

Reviewer #2: Yes

**********

Reviewer #2: I thank the authors for their comprehensive response and revision of the manuscript. I have a few follow-up questions regarding the new analyses –

1) the sleep categories used in the study are suggested to be informed by the American Academy of Sleep Medicine and the Centers for Disease Control and Prevention, both endorse 7 to 9 hours of sleep as optimal for adult health. It is therefore unclear why 6 hours of sleep is not included in the short sleep category and 9 hours is considered long rather than recommended sleep. This is especially relevant when considering the young cohort in the study (18-25) which routinely sleep more than 8 hours when sleep is free from constrains (e.g. Motomura et al. 2017, Klerman et al. 2008). I would therefore recommend reanalyzing the data using the following categories which are more aligned with current guidelines –

• Short sleep: ≤6 hours,

• Recommended sleep: 7–9 hours,

• Long sleep: ≥10 hours.

2) When discussing long sleep please address current studies revealing a potential bias in self-reported sleep duration such that a proportion of self-reported long sleepers are in-fact getting insufficient sleep when measured objectively (e.g. Wang et al 2025, Health Data Science)

3) Table 1 – headlines refer to years instead of hours of sleep for all sleep categories. Please revise.

4) There is a typo in the abstract (adults’ instead of adults).

**********

what does this mean? ). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy

Reviewer #2: No

**********

Acceptance letter

Valentina Alfonsi

PONE-D-25-11133R1

PLOS ONE

Dear Dr. Ndebele-Ngwenya,

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now being handed over to our production team.

At this stage, our production department will prepare your paper for publication. This includes ensuring the following:

* All references, tables, and figures are properly cited

* All relevant supporting information is included in the manuscript submission,

* There are no issues that prevent the paper from being properly typeset

You will receive further instructions from the production team, including instructions on how to review your proof when it is ready. Please keep in mind that we are working through a large volume of accepted articles, so please give us a few days to review your paper and let you know the next and final steps.

Lastly, if your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

You will receive an invoice from PLOS for your publication fee after your manuscript has reached the completed accept phase. If you receive an email requesting payment before acceptance or for any other service, this may be a phishing scheme. Learn how to identify phishing emails and protect your accounts at https://explore.plos.org/phishing.

If we can help with anything else, please email us at customercare@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Valentina Alfonsi

Academic Editor

PLOS ONE

Associated Data

    This section collects any data citations, data availability statements, or supplementary materials included in this article.

    Supplementary Materials

    S1 File. Supplementary figures and tables.

    Contains Supplementary Figure 1 (Trends in Sleep Duration Categories), Supplementary Figure 2 (Prevalence of Depression by Sleep Duration), and Supplementary Tables 1–5 describing baseline characteristics, treatment effects, mentally and physically unhealthy days, and predicted general health probabilities.

    (DOCX)

    pone.0321347.s001.docx (183.7KB, docx)
    Attachment

    Submitted filename: report_10.1101+2025.03.06.25323545.pdf

    pone.0321347.s002.pdf (588.9KB, pdf)
    Attachment

    Submitted filename: Response to Reviewers_Sleep7232025.docx

    pone.0321347.s004.docx (22.4KB, docx)

    Data Availability Statement

    All relevant data are within the manuscript and its supporting information files.

    Articles from PLOS One are provided here courtesy of PLOS

    RESOURCES