Introduction

Approximately 47 million United States (US) women have prediabetes, an often asymptomatic condition characterized by having elevated blood sugar levels (hemoglobin A1C test: 5.7–6.4%) that do not yet meet the cutoff for diabetes (hemoglobin A1C test: ≥6.5%) [1]. An estimated 70% of prediabetes cases are expected to progress to diabetes [2, 3]. In addition to the clinical consequences of prediabetes and diabetes, these conditions respectively cost the US around $43.4 billion and $360.6 billion each year in terms of medical expenses and missed productivity [4]. While the burden of prediabetes and diabetes in the United States is notable, this very burden varies between sociodemographic groups and geographic areas [5–8].

One such example of a US diabetes burden disparity is that there is a 17% higher diabetes prevalence among rural populations as compared to urban populations [9–11]. This increased prevalence of diabetes may be attributable to the management of prediabetes in rural settings. For example, prediabetes cases among rural US women may be more likely to progress to diabetes due to greater obesity and physical inactivity rates in US rural areas [12]. Additionally, other modifiable factors may also contribute to this observed inequity. Like obesity and physical inactivity, short sleep duration (SSD) (<7 hours/night), for example, has also been linked with prediabetes to diabetes progression [13, 14]. A systematic review conducted across different populations, including South Korean, Finnish, and American individuals, indicated that SSD was associated with a 59% (95% CI: 29%-97%) increased risk of progression from prediabetes to diabetes [13]. Thus, understanding the prevalence of SSD among individuals with prediabetes is critical for the identification of modifiable risk factors for diabetes development.

Despite a high diabetes burden in US rural women and an association between SSD and progression to diabetes, there is little information on SSD estimates for US women in rural areas with prediabetes. Existing US research points towards high levels of short sleep duration (38.5%, 95% CI: 37.5%-39.6%) among rural residents nationwide, but this finding may not apply to rural women with prediabetes [15]. Additionally, the few US observational and interventional studies that have examined SSD among individuals with diabetes do not provide estimates specific to women or rural residents [16–18]. As such information can inform evidence-based diabetes prevention for women in medically underserved rural areas, we conducted an observational study using cross-sectional data from 2016–2020 to ascertain SSD estimates for US women with prediabetes by rural/urban residence status. We also determined if associations between rural/urban residence status and SSD were present prior to and following adjustment for prediabetes- and sleep-related sociodemographic factors.

Methods

Results

We included 20,997 women with prediabetes in our study (Table 1). 33.7% of these women resided in rural US areas. Women with prediabetes in rural areas were comparatively more likely to be White (rural % White: 83.5%, urban % White: 66.4%), have a lower income (rural % <$25,000: 42.3%, urban % <$25,000: 30.6%), and have completed fewer years of education (rural % college graduate: 14.8%, urban % college graduate: 27.1%) than their urban counterparts.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Characteristics of women with prediabetes participating in the 2016–2020 Behavioral Risk Factor Surveillance System (BRFSS) surveys.

https://doi.org/10.1371/journal.pone.0284304.t001

Between 2016–2020, unadjusted SSD prevalence was found to be similar between rural (35.5%, 95% CI: 33.0%-38.0%) and urban women (35.4%, 95% CI: 33.7%-37.1%). Using unadjusted logistic regression, rural residence was not associated with SSD among US women with prediabetes (Odds Ratio (OR): 1.00, 95% CI: 0.87–1.14) (Table 2). Following adjustment for sociodemographic factors, there continued to be no association between rural residence and SSD (OR: 1.06, 95% CI: 0.92–1.22). Furthermore, among US women with prediabetes, being aged 18–44 years (OR: 2.22, 95% CI: 1.75–2.82) and aged 45–64 years (OR: 1.64, 95% CI: 1.41–1.92) was associated with significantly higher odds of short sleep duration compared to being aged 65 years and up, being Black (OR: 1.41, 95% CI: 1.17–1.70) with significantly higher odds of short sleep duration compared to being White, and earning <$50,000 with significantly higher odds of short sleep duration compared to earning ≥$50,000.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Associations between rural/urban residence and short sleep duration among US women with prediabetes in the 2016–2020 Behavioral Risk Factor Surveillance System (BRFSS) surveys².

https://doi.org/10.1371/journal.pone.0284304.t002

Discussion

We determined levels of SSD by rural/urban residence status in a nationally representative sample of US women with prediabetes. Regardless of residence status, around 35% of US women with prediabetes had SSD. Prevalence of and odds of SSD were similar between rural and urban women in unadjusted and adjusted analyses.

Previous US research examining SSD among individuals with prediabetes has yielded varied results [16–18]. In a study following American Indians and Alaskan natives with prediabetes between 2006–2009, 41.3% of individuals in the study self-reported having SSD (≤6 hours of sleep/night) [16]. Using National Health and Nutrition Examination Survey data from a similar time frame (2005–2008), investigators observed that 56.9% of participants in the survey slept ≤6 hours/night [17]. More contemporary research coming from the Restoring Insulin Secretion Consortium randomized controlled trials (2013–2017) conducted in Illinois, Indiana, Washington, and California found that 31.9% of participants in the trial reported sleeping <6 hours/night [18].

Differences between our findings and those from previous US research may be attributed to differing definitions of SSD and improvements in sleep duration over time as part of US preventative health initiatives (i.e., Healthy People 2020) [15]. Moreover, comparison of our inferences around sociodemographic factors with inferences from earlier studies is also challenging as those existing studies do not provide estimates of associations between sociodemographic factors and SSD [16–18]. However, our findings align with CDC estimates of SSD for the general US population showing greater levels of SSD among those aged <65 years and Blacks having the second highest SSD levels (43.5%, 95% CI: 42.3–44.6%) after Native Hawaiian/Pacific Islanders [26]. Future study examining the reasons as to why women with prediabetes from certain sociodemographic groups experience increased SSD is needed. Our study expands on prior US short sleep and prediabetes research by providing previously unknown SSD estimates for women and showing that rural women with prediabetes have SSD levels that are as high as their urban counterparts.

While our finding of similar SSD levels in rural and urban women may be somewhat surprising, it may be explained by findings from prior research on sleep disrupting factors such as noise and light pollution [27–29]. There is evidence that noise pollution does not significantly differ between rural and urban areas [27] and that some rural areas have higher levels of light pollution than urban areas [28, 29]. Thus, rural women may experience multiple factors that predispose them to SSD, potentially explaining our findings.

Although US estimates of SSD specific to women without prediabetes are unavailable, CDC statistics indicate that 32.2% (95% CI: 31.7–32.7%) of women in the general US population have SSD, which is lower that what we found for rural and urban women with prediabetes [26]. High levels of SSD among women with prediabetes present a potential opportunity to reduce prediabetes to diabetes progression. Efforts to improve SSD among rural women in medically underserved areas may be especially challenging due to access to care-related factors, such as primary care providers and sleep specialists. For example, the limited number of primary care providers in rural areas may contribute to rural residents being unaware of the risk associated with SSD and prediabetes to diabetes progression [30]. In addition, there is evidence that between 2010–2021 a lack of clinics and sleep specialists in rural areas led to lower receipt of sleep care along US rural compared to urban residents [31, 32]. Although telehealth could be a viable sleep care option for rural residents, potential obstacles include the low availability of broadband required for telehealth in rural areas and the necessity of in-person overnight sleep diagnostic tests [33, 34]. Further efforts to improve SSD among women with prediabetes, especially efforts in conjunction with improving other known diabetes risk factors, are warranted and could be instrumental in reducing the high diabetes burden in rural areas.

Our study has several limitations. Sleep duration is self-reported and, as such, it is possible there may be some self-report and non-response bias. However, studies have found that self-reported sleep duration is highly correlated with measurements from actigraphy, and self-reported sleep has been widely used in existing US research looking at SSD among people with prediabetes [16–18, 35]. Furthermore, it is likely that if any of the mentioned biases were present, they would be non-differential and lead to more conservative estimates [15]. We are unable to include SSD estimates for women with prediabetes who are institutionalized, but, as individuals who are institutionalized represent a small fraction of the US population, our findings should be generalizable to the majority of US women with prediabetes [36, 37].

Conclusion

While SSD estimates among US women with prediabetes did not vary by rural/urban residence status, there was, nonetheless, a high level of SSD in rural areas. Additional focus on strategies to improve access to sleep care among women in rural areas, particularly those from certain sociodemographic backgrounds, in tandem with managing other diabetes risk factors may be helpful to reducing prediabetes to diabetes progression in this medically underserved high diabetes burden population.