Sleep, Health Outcomes and Body Weight (SHOW) study: a measurement burst design study on sleep and risk factors for obesity in black emerging adults in North Carolina, USA

Abstract

Introduction

Black emerging adults (18–28 years) have the highest risk of short sleep duration and obesity. This increased risk may be partly explained by greater stress levels, which may result from race-related stress (racial discrimination and heightened race-related vigilance) or living in more disadvantaged home and neighbourhood environments. Insufficient sleep may also impact obesity risk via several weight-related mechanisms including energy balance, appetite and food reward, cortisol profiles and hydration status. This paper describes the rationale, design and methods for the Sleep, Health Outcomes and Body Weight (SHOW) study. This study aims to prospectively assess the effects of sleep, race-related stress and home/neighbourhood environments on weight-related mechanisms and obesity markers (body weight, waist circumference and fat mass) in 150 black emerging adults.

Methods and analysis

The SHOW study follows a measurement burst design that includes 3, 7-day data collection bursts (baseline, 6-month and 12-month follow-ups). Sleep is measured with three methods: sleep diary, actigraphy and polysomnography. Energy balance over 7 days is based on resting and postprandial energy expenditure measured via indirect calorimetry, physical activity via accelerometry and self-reported and ad libitum energy intake methods. Self-reported methods and blood biomarkers assess fasting and postprandial appetite profiles and a behavioural-choice task measures food reward. Cortisol awakening response and diurnal cortisol profiles over 3 days are assessed via saliva samples and chronic cortisol exposure via a hair sample. Hydration markers are assessed with 24-hour urine collection over 3 days and fasting blood biomarkers. Race-related stress is self-reported over 7 days. Home and neighbourhood environments (via the Windshield Survey) is observer assessed.

Ethics and dissemination

Ethics approval was granted by the University of North Carolina at Greensboro’s Institutional Review Board. Study findings will be disseminated through peer-reviewed publications, presentations at scientific meetings and reports, briefs/infographics for lay and community audiences.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• This study focuses exclusively on black emerging adults, an understudied population despite compelling evidence that this population has the highest risk for both short sleep duration and obesity.

• This study includes comprehensive measures of habitual sleep via three methods (self-report, actigraphy and polysomnography) at home, which will provide the most comprehensive assessment of sleep in a sample of emerging adults.

• This study includes multiple self-reported, objective and biomarker measurements for up to 7 days over three bursts (baseline, 6-month follow-up and 12-month follow-up) to better characterise sleep and weight-related mechanisms over time. This approach allows us to assess both short-term (within 1 week) and long-term (12-month) changes (or stability) in multiple health behaviours (sleep, diet, physical activity) and weight-related mechanisms (energy balance, hydration status and cortisol profiles) which may contribute to a heightened risk of weight gain/obesity and cardiometabolic disease during emerging adulthood.

• Contextual (or social)-level factors are limited to self-reported and observer-assessed measures of the home and neighbourhood environments captured during participant home visits.

• The study sample will not permit meaningful comparisons by population subgroups such as chronotype, socioeconomics, substance and medication use, cultural identity/strength, and coping mechanisms in response to stress.

Introduction

Emerging adulthood (approximately ages 18–28 years) represents a developmental period often characterised by ‘frequent change and exploration’.¹ Adult health behaviours, including sleep, are often established during emerging adulthood, shaping the lifelong trajectory of cardiometabolic risk and well-being.² Additionally, evidence suggests that sleep duration gradually declines during emerging adulthood.³ Concurrent with these declines in sleep, obesity rates double from adolescence through early emerging adulthood,and then double again by late emerging adulthood.^4–6 The association between short sleep duration and obesity is also stronger among younger adults compared with older adults, suggesting that the short sleep–obesity association tends to wane with age.^7–14 Nonetheless, prospective evidence for sleep-obesity associations among emerging adults is currently scarce.¹⁵

Compared with peers in other racial groups, black emerging adults have the highest risk for short sleep duration^{3 16–18} and obesity,^{19 20} but sleep data are limited to self-report^{3 16–18} and actigraphy-based²¹ measures. In addition, the CARDIA study reported excessive weight gain (defined as >20 kg) in 15%–22% of black emerging adults (compared with 6%–9% of white emerging adults),²⁰ and a recent report noted that black emerging adults face the prospect of obesity a full decade before their white counterparts.¹⁹

A greater risk of short sleep and obesity in black emerging adults may be partly explained by greater stress levels,²² which may result from race-related stress (experiences of racial discrimination and heightened race-related vigilance) or living in more disadvantaged home and neighbourhood environments.²³ Heightened feelings/experiences of racial discrimination and race-related vigilance have been associated with a greater risk of short sleep duration and poor sleep quality^{17 24–28} and obesity/weight gain^29–33 in black adults. Living in more disadvantaged home and neighbourhood environments may also directly undermine sleep^34–39 and increase obesity/weight gain risk.^40–44 For example, living in a crowded household,⁴⁵ suboptimal housing conditions³⁹ and a neighbourhood perceived by residents as crime ridden, noisy and in structural disrepair^35–38 are associated with perceived short sleep duration and poorer sleep quality after adjusting for individual-level sociodemographic factors. These studies also noted that associations between disadvantaged neighbourhood environments and self-reported health were mediated by sleep quality.^35–37

Insufficient sleep (here defined as longer sleep-onset latency, shorter sleep duration, poorer sleep quality and efficiency or lower amounts of slow-wave sleep (SWS)) may impact obesity markers (here defined as body weight, waist circumference and body composition) via several weight-related mechanisms, including (1) energy balance (difference between energy intake and energy expenditure), (2) appetite and food reward, (3) cortisol profiles (cortisol awakening response and diurnal cortisol profiles based on acutely circulating cortisol levels in saliva; chronic stress/cortisol exposure via hair cortisol measures) and (4) hydration status. Body weight change is associated with an imbalance between the energy content of foods consumed and the amount of energy expended to maintain life (resting energy expenditure), digest foods consumed (thermic effect of food) and accomplish physical work (physical activity energy expenditure).⁴⁶ Weight gain and eventual development of obesity require that a positive energy balance (energy intake to exceed total energy expenditure) be maintained over time.⁴⁶ Habitual short sleep duration and imposed sleep restriction lead to increases in total energy intake,^47–49 which may lead to weight gain if sustained. Greater energy intake after sleep restriction may be driven by increased appetite (subjective feelings or haematological markers) or food reward (ie, eating without hunger and a greater desire to consume energy-dense foods).⁴⁷ Cortisol profiles^50–52 and hydration status⁵³ have also independently been associated with obesity. Short sleep duration and imposed sleep restriction are associated with flatter diurnal cortisol profiles,^54–62 which may signal impaired cortisol regulation and are associated with greater body weight and central obesity.^{50 52 63} Lastly, short sleep duration is associated with a higher risk of underhydration,⁶⁴ which in turn is associated with a higher risk of obesity.⁵³ However, no study has investigated stable markers of hydration status within the context of both sleep and obesity.

The specific aims of the Sleep, Health Outcomes and Body Weight (SHOW) study are (1) to examine the prospective effects of insufficient sleep on weight-related mechanisms and obesity markers; (2) to examine the prospective effects of race-related stress and living in more disadvantaged home and neighbourhood environments on insufficient sleep variables, weight-related mechanisms (primarily, cortisol profiles) and obesity markers and (3) to assess the mediating role of sleep on associations among race-related stress and living in more disadvantaged home and neighbourhood environments with weight-related mechanisms (primarily, cortisol profiles) and obesity markers in black emerging adults ages 18–28 years. We will also explore the influence of individual-level sociodemographic and health-related factors (eg, chronotype, substance and medication use, cultural identity/strength, and coping mechanisms) as potential moderators of observed associations between sleep and race-related stress with weight-related mechanisms and obesity markers. The purpose of this paper is to describe the protocol for the SHOW study.

Methods and analysis

Study participants

We aim to enrol 150 emerging adults ages 18–28 years old who self-identify as black or African American into the SHOW study. Other exclusion criteria include (1) not able to speak and read English fluently because this may limit the participants’ ability to provide informed consent and complete study questionnaires which have not all been validated in other languages; (2) currently trying to gain or lose weight, had weight loss surgery or other surgery on the digestive tract or are currently taking diuretics because these will have a direct impact on our study outcomes; (3) planning to move outside of a 64 km (40 mile) radius of the research site within the next 12 months (over the course of the data collection period) because this may limit the participants’ ability to travel to the research site for study assessments/laboratory visits and increase the risk of missing data and attrition; (4) currently pregnant or planning to become pregnant over the next 12 months (females only) because this will have a direct impact on our study outcomes; (5) food allergies due to the use of food intake protocols (eg, standard breakfast, ad libitum lunch intake), although we are able to accommodate some dietary restrictions such as vegetarianism and (6) has a current or previous clinical eating disorder diagnosis or experiences recent/current symptoms of eating disorders because the type of measures in this study (eg, obesity markers, energy intake) may be ‘triggering’ for these individuals. The latter will be screened using the Eating Attitudes Test (EAT-26),⁶⁵ a 26-item questionnaire that screens for a history of clinical eating disorders and recent/current symptoms of eating disorders. Anyone with a total score ≥20 or an answer of ‘yes’ to the behavioural questions will be excluded from participation in this study and provided with a list of local resources/dieticians to contact for follow-up. Lastly, we are not excluding participants based on medication use (except for diuretics), however, the type and dosage of medications must be stable for at least 3 months prior to enrolling into this study. Interested individuals complete an initial screening questionnaire over the phone to determine eligibility. Participants who meet all eligibility criteria attend a preliminary session to obtain written informed consent.

Recruitment efforts, patient and public involvement

Pilot data collection for the SHOW study took place from May 2022 to May 2023 and included 15 participants who completed a baseline and 6-month visit detailed in the SHOW study protocol. This pilot study provided evidence on completion rates for all measurements, which ranged from 93% to 100% for all outcome measurements. Furthermore, feedback was collected from pilot study participants throughout data collection on study procedures and potential burden of the protocol, and modifications were made to the SHOW study protocol following this feedback.

Recruitment strategies for the SHOW study include advertisements in local establishments and working with local consultants that provide culturally sensitive research participant recruitment and retention services. Furthermore, a study-specific community advisory board (CAB) with a roster of ~10–12 black emerging adults will be formed following established guidelines^{66 67} to offer strategies on engaging and recruiting participants in research, building community partnerships and knowledge dissemination efforts as the SHOW study progresses. For instance, CAB members will assist with creating recruitment messaging that promotes the importance of sleep health and with the design of lay summaries and infographics that discuss study-related activities, progress and findings. CAB members are compensated for their assistance.

Study design and procedures

The SHOW study follows a measurement burst design that includes data collection over three bursts (baseline, 6-month follow-up and 12-month follow-up), each lasting 7 days (figure 1). Data collected over 7 days include (1) actigraphy-based measures of physical activity and sleep, and self-reported sleep measures; (2) self-reported energy intake on days 1–4 and objective measures on days 5–7; (3) total energy expenditure based on 7-day physical activity, thermic effect of food and resting energy expenditure measures; (4) water intake measured with a smart water bottle; (5) self-reported race-related stress and (6) self-reported screen time. Data collected over 3 days include: (1) salivary-based cortisol profiles and (2) urine-based hydration status. Data collected over 2 days include (1) polysomnography-based sleep at home. Data collected for 1 day include (1) obesity markers; (2) self-reported and haematological measures of appetite and hydration status; (3) food reward; (4) chronic cortisol levels via a hair sample; (5) observer-assessed measures of home and neighbourhood environments and (6) self-reported measures of sociodemographic (eg, current employment/jobs, household income, relationship status) and health-related factors (eg, substance and medication use, perceived stress, depression and anxiety).

Figure 1.

Overview of study design. *Two in-lab visits: day 0 for a 90 min visit and day 5 for a 6-hour visit.

The frequency and timing of study measurements completed at each measurement burst are presented in table 1.

Table 1.

The frequency and timing of study measurements at each burst

	Preliminary session	Day 0	Day 1	Day 2	Day 3	Day 4	Day 5	Day 6	Day 7
Questionnaires
Self-reported sociodemographics, chronotype, social desirability, coping mechanisms, cultural strength, sensitivity to reward and eating behaviours*	X
Self-reported substance use, medication use and other sociodemographics that may change over time, food insecurity, physical activity, young adult activities and responsibilities, perceived stress and mental health outcomes*		X
Self-reported screen time use (habitual and daily measures)*		X	X	X	X	X	X	X	X
Self-reported frequency of race-related discriminatory experiences and feelings of heightened race-related vigilance (habitual and daily measures)*		X	X	X	X	X	X	X	X
Sleep
Polysomnography-based sleep*							†	†	†
Actigraphy-based sleep*			X	X	X	X	X	X	X
Self-reported sleep*			X	X	X	X	X	X	X
Home and neighbourhood environments
Observer-assessed home and neighbourhood environments (Windshield Survey)*							‡	‡
Obesity and cardiometabolic markers
Objective measures of weight, height, blood pressure, waist circumference and fat mass§		X
Fasting blood markers of cardiometabolic risk (eg, insulin, glucose)§							X
Energy and fluid intake
Self-reported measures of food and caloric beverage intake with a log*			X	X	X	X
Objective measures of food and fluid intake with a food menu*							X	X	X
Objective measures of water intake with a smart water bottle*			X	X	X	X	X	X	X
Energy expenditure
Actigraphy-based physical activity*			X	X	X	X	X	X	X
Thermic effect of food measured with indirect calorimetry§							X
Resting energy expenditure measured with indirect calorimetry§							X
Appetite and food reward
Self-reported feelings of appetite via Visual Analogue Scales§							X
Blood markers of appetite (eg, ghrelin, leptin)§							X
Food reward via a computer-based behavioural task§							X
Cortisol profiles
Cortisol awakening response with saliva samples*			†	†	†	†		X	X
Diurnal cortisol profiles with saliva samples*			†	†	†	†		X	X
Chronic cortisol levels with a hair sample§		X
Hydration status
Urinary hydration markers (24-hour urine collection)*			†	†	†	†		X	X
Blood markers of hydration (eg, copeptin)§							X

X indicates that this measurement takes place during this day.

*Measures completed at home.

†Participants choose one of the days (from days 1 to 4) for saliva and urine collection and two of the days (from days 5 to 6 or days 6 to 7) for polysomnogram-based sleep measures.

‡Measurement completed on the first night of the home visit (either day 5 or day 6).

§Measures completed at the research site/UNCG.

UNCG, University of North Carolina at Greensboro.

During the preliminary session (baseline only), participants provide informed consent, complete questionnaires (see below for more details) and rate 171 food images that vary in both fat content and taste with the following question ‘How often do you consume this food item?’. The top 16 rated food items will be used to personalise the Leeds Food Preference Questionnaire (LFPQ) for each participant based on preference/familiarity.⁶⁸ Day 0 starts before each 7-day burst of data collection and is a 90 min in-laboratory visit. Participants are instructed to refrain from smoking, vaping, caffeine, eating and exercising for at least 2 hours prior to this visit. Anthropometrics (ie, height, weight, waist and hip circumference), resting blood pressure, body composition and a hair sample (if a participant is able and willing to provide a sample) are taken, and participants complete additional questionnaires. Day 1 represents the start of ‘burst’ data collection, where participants begin to wear 2 accelerometers, use a smart water bottle to track water intake and complete self-reported daily measures (via paper or online versions) of accelerometer wear time, sleep, race-related stress and screen time. Participants are also instructed to fill out a food diary on days 1–4 and select one of these initial days of data collection for 24-hour urine and saliva collection (ie, urinary hydration markers and salivary cortisol profiles will be measured over a 24-hour period during days 1–4).

Data collection for day 5 includes a 6-hour in-laboratory protocol (figure 2). Participants are instructed to arrive at the laboratory following a 12-hour, overnight fast from food (although ad libitum water intake is allowed) and refrain from smoking, vaping and caffeine the morning of the visit, as well as alcohol intake and vigorous physical activity participation for 24 hours prior to this visit. First, resting energy expenditure is measured via indirect calorimetry. After participants consume 500 mL of water, an intravenous catheter is inserted into the participants’ radial, hand or antecubital vein by a trained research technician. Blood samples are collected with a 22-gauge or 24-gauge butterfly needle and a vacutainer system. Universal precautions and Occupational Safety and Health Administration (OSHA) guidelines are always followed for blood handling. A total of 10 mL of blood is collected into a serum separator tube (~3 mL) and 2 EDTA plasma tubes (3 mL per tube) at each of the seven times points starting with a fasting sample (figure 2) for a total of 7 serum and 14 plasma tubes. A standard breakfast is then served (64 g biscuit, 10 g of butter, 12oz of orange juice and 65 g of egg bites (vegetarian or ham and cheese); energy content: 531–536 kcal), and participants are given 15 min to consume this meal in its entirety. After breakfast, blood draws and measures of self-reported appetite and thirst via Visual Analogue Scales are repeated every 30 min for 3 hours. During this same 3-hour time frame, postprandial energy expenditure is measured for 30 min during the second half of each hour to assess the thermic effect of food via indirect calorimetry. After the last blood draw is taken (ie, 3 hours postbreakfast), the intravenous catheter is removed, and participants complete the LFPQ to measure food reward. Participants then self-select food items that they may want to consume during lunch from a validated menu,⁶⁹ ad libitum quantities of each item (ie, 2 portions of each item are provided, but more can be requested) is served, and participants have 30 min to ‘eat as much or as little as you want’. Participants also self-select food and caloric beverage items that they may want to consume for the remainder of day 5, and days 6 and 7 at home and instructed to only consume the items provided to them during these days. Following lunch, participants repeat the LFPQ. Lastly, participants bring home prelabelled vials to collect saliva samples on five occasions (on awakening, 30 and 45 min postawakening, 8 hours postawakening and immediately prior to bedtime) each day on days 6 and 7 with participants recording the clock time for each saliva collection. In addition, all participants receive opaque urine containers to collect 24-hour urine produced on days 6 and 7.

Figure 2.

Outline of in-laboratory protocol for day 5. PSG, polysomnogram; UNCG, University of North Carolina at Greensboro; VAS, Visual Analogue Scale.

Two research assistants travel to the participants’ home ~2 hours prior to their usual bedtime on days 5–6 or 6–7 based on participants’ availability/preference. The research assistants set up the polysomnogram for sleep monitoring and collect data on home and neighbourhood environments by completing the Windshield Survey.⁶⁷ On day 8, participants remove the accelerometers on awakening and return all study equipment, any remaining food items/containers, and saliva and urine samples to the research site. Each participant receives up to US$300 per data collection burst and a US$100 incentive for completing the entire study for a total of US$1000 over 12 months.

Questionnaires: preliminary session (baseline only)

Participants complete a sociodemographic and medical questionnaire that collects information on age, biological sex, gender identity, nationality (parental and self), address of current and permanent residence (if different than current), years of education, current employment/jobs and shift work schedule (if applicable), household income and the number of individuals that this income supports, the use of public assistance programmes (eg, Medicaid, food stamps), relationship status, medical history and current medication use (prescribed and over the counter). They also complete the Three-Factor Eating Questionnaire to assess eating behaviours⁷⁰; the Sensitivity to Punishment and Sensitivity to Reward Questionnaire to assess the degree of anxiety and impulsivity in response to scenarios of punishment and reward⁷¹; the Morningness-Eveningness Questionnaire to assess chronotype⁷²; the short-form Marlowe-Crowne Social Desirability Scale⁷³ to determine the extent to which social desirability affects responses to self-reported measures; as well as the Giscombé Superwoman Schema in females⁷⁴ and the John Henryism Active Coping Scale in males⁷⁵ to assess coping mechanisms and effortful strength in response to stress. Lastly, participants complete a compilation of questionnaires to assess the construct of Black Cultural Strength.⁷⁶ Specifically, the work by Johnson and Carter⁷⁶ proposed that positive and affirming components of black racial identity, racial socialisation, racism-related coping, communalism and cultural spirituality—together, defined as a construct of black cultural strength—are positively associated with psychosocial health⁷⁶ and reduce the negative effects of racism and discrimination on perceived stress.⁷⁷

Questionnaires: day 0 and daily measures during each data collection burst

On day 0, participants are asked to self-report habitual substance use (eg, smoking, alcohol, vaping), daily home activities and responsibilities (eg, caregiving, family food shopping and meal preparation, household cleaning), as well as any sociodemographic and medical outcomes that may change from baseline to the 6-month and 12-month follow-up time points (eg, medication use, current employment/jobs, current address and household income, relationship status). Furthermore, participants complete the Centre for Epidemiologic Studies Depression Scale to assess feelings of depression over the past week⁷⁸; the State-Trait Anxiety Inventory to assess feelings of anxiety over the past 2 weeks⁷⁹; the Perceived Stress Scale to assess overall stress levels over the past month⁸⁰; the short-form International Physical Activity Questionnaire to assess physical activity participation over the past 7 days⁸¹; the short-form US Household Food Security Survey Module to identify experiences of food (in)security over the last 30 days⁸²; and the Assessment of Sleep Environment (ASE) to assess the degree to which the physical (home) environment impacts sleep.⁸³

Three additional questionnaires are completed on day 0 and every evening during each data collection burst (ie, over 7 days) to assess both habitual and daily measures. First, a Screen Time Questionnaire is used to assess screen time/use for five different categories of devices (television, connected devices, laptop/computer, smartphone and tablet)⁸⁴ over the past 6 months (habitual measure) and that day (daily measure). Next, the Racism and Life Experiences Scale (RaLES)⁸⁵ and the Heightened Vigilance Scale⁸⁶ are administered on day 0 to assess the frequency of race-related discriminatory experiences and feelings of heightened race-related vigilance (collectively, providing a measure of race-related stress) over the past year (baseline) and over the past 6 months (6-month and 12-month follow-up time points). For daily measures, participants are asked to report whether each item on the RaLES and Heightened Vigilance Scale happened that day (yes/no) and the number of items endorsed are summed to get daily discrimination and heightened vigilance scores.⁸⁷

Sleep

Three methods are used to collect sleep data during each data collection burst: actigraphy, sleep diary and in-home polysomnography. Because actigraphy and sleep diaries are non-invasive and have minimal participant burden, they are used over the entire 7-day burst to capture usual sleep and wake patterns. In-home polysomnography is more labour intensive and cumbersome but provides unique information on sleep stage duration and sleep-disordered breathing via Apnea/Hypopnea Index (AHI) measures that cannot be captured with other methods.⁸⁸ Participants complete two nights of in-home polysomnography on either days 5–6 or 6–7 based on their availability and to align with the salivary cortisol and urinary hydration measures which take place on days 6–7.

Participants are asked to wear an Actigraph GTx9 Link accelerometer (Actigraph, Pensacola, Florida, USA) on their non-dominant wrist to assess sleep over the 7-day burst. Participants are instructed to remove the device for water-based activities and periods of non-wear time >10 min are reported in an accelerometer log sheet. Accelerometer data are collected at a sampling rate of 30 Hz and aggregated to 60 s epoch files for analysis by the Actilife software (currently V.6.13.5). At this time, the Sadeh sleep scoring algorithm is used to derive sleep duration (minutes/day), sleep timing (bed times and wake times; clock time), sleep-onset latency (minutes) and sleep efficiency (sleep duration/time in bed; %), which is validated using the bed times and wake times from the accelerometer log sheet. The Sadeh algorithm was chosen because it was originally validated on a sample of adolescents and emerging adults.⁸⁹ Additional algorithms or processing methods to derive sleep variables using the raw accelerometry data may be used at the conclusion of data collection given that processing and analysis recommendations for accelerometry-derived data are rapidly advancing.

The Karolinska Sleep Diary^90–92 is used to capture subjective sleep measures on awakening over the 7-day burst. This questionnaire asks participants to report their bed times and wake times, time required to fall asleep (indicator of sleep latency), the number of awakenings throughout the night and to rate different measures of perceived sleep quality (eg, ‘how did you sleep?’; ‘Ease of falling asleep?’) and daytime sleepiness (eg, ‘feeling refreshed after wakening?’; ‘ease of waking up?’) on a 5-point Likert scale.

The Nox A1 full polysomnogram system (Nox Medical USA, Suwanee, Georgia, USA) is used to measure sleep duration (minutes/day), sleep latency (elapsed time between bedtime identified with the accelerometer data and log and 10 min of stage 1 or 20 s of any other sleep stage), sleep efficiency (sleep duration divided by time in bed based on bed times and wake times identified with the accelerometer data and log; %), wake after sleep onset in minutes, absolute (minutes) and relative (%) time spent in each sleep stage (stage 1, stage 2, SWS and rapid eye movement (REM) sleep), as well as severity of sleep-disordered breathing based on the number of apneas and hypopneas per hour of sleep (AHI<5 is ‘normal sleep’, 5–14 is mild sleep-disordered breathing, 15–30 is moderate sleep-disordered breathing and >30 is severe sleep-disordered breathing). This system is lightweight and has been shown to be feasible for unattended at-home sleep monitoring compared with in-laboratory sleep monitoring.⁹³ The 10–20 placement system is followed to place six electroencephalogram (EEG) electrodes (O1, O2, C3, C4, F3, F4), two electrooculogram (EOG) electrodes and three reference electrodes (1 forehead and 2 mastoids). Additionally, three adhesive electrodes are placed on the chin to monitor muscle tone via electromyography (EMG) signalling, a nasal canula is placed below the nose to monitor respiratory flow, a two-lead ECG wire is placed under the right clavicle and over the left seventh rib, and a pulse oximeter is placed over the wrist and index finger to monitor blood oxygen saturation during sleep. Gauze and medical tape are used to secure electrodes in place with Ten20 conductive paste. All electrodes/wires are attached to the Nox A1 device, which is then secured to participants with two adjustable belts. An alternative approach to the traditional EEG montage offered by Nox Medical is the Self Applied Somnography (SAS) setup (Nox Medical, Reykjavik, Iceland), which is a simplified EEG montage that positions 10 stick-on electrodes along the forehead and near the eyes to capture prefrontal cortex EEG (AF8, AF4, GND, AFz, AF3, AF7) and EOG (E1, E2, E3, E4) activity without recording the mastoid channels.^{94 95} Chin EMG is also omitted, although an EMG signal derived from the eyes is available for analysis.^{94 95} Since EEG setups may fail to accommodate thick hair types or various hairstyles, offering the SAS setup as an alternative, but valid,^{94 95} approach to in-home sleep monitoring may promote a more culturally inclusive method for individuals who are unable to or prefer not to use the traditional PSG setup with Ten20 paste adhesion.

Data collected by the Nox A1 system are downloaded onto the Noxturnal software (currently US V.6.3.0) for analysis. Automated algorithms to analyse AHI and sleep stage duration are embedded within the Noxturnal software and have shown good reliability with manually scored recordings based on the American Academy of Sleep Medicine (AASM) guidelines.^{96 97} Therefore, the automated algorithms are used to initially score the recordings and then all polysomnogram recordings are independently reviewed by one staff member following the same AASM criteria.⁹⁷ Any discrepancies in scoring between the automated algorithm and the staff member are reviewed by a second staff member and resolved by mutual agreement (between staff members).

Home and neighbourhood environments

In addition to the ASE questionnaire⁸³ completed by participants on day 0, research staff who complete the home visits for polysomnogram setup also fill out a Windshield Survey,⁹⁸ which includes 12 items with Likert scales to capture data on participants’ home and neighbourhood environments. Example items include perceived noise in the neighbourhood around the home at night (0=can’t rate, 1=very quiet and 5=very noisy), perceived safety of the home’s interior (0=can’t rate, 1=obviously dangerous and 5=above average safety; reversed scored), perceived cleanliness of the home (0=can’t rate, 1=very dirty and 5=clean; reversed scored) and safety of the neighbourhood around the home at night (0=can’t rate, 1=very safe/crime free and 5=very unsafe/high risk). Scores for each item are summed, with higher scores indicating more disadvantaged home and neighbourhood environments. This survey is a measure of direct neighbourhood observation developed by the Conduct Problems Prevention Research Group.⁹⁸ It has successfully been used to measure home and neighbourhood noise levels, safety and upkeep in large longitudinal studies including the Family Life Project.^{99 100}

Obesity markers

All obesity marker measurements take place during the day 0 visit. Body height is measured to the nearest 0.1 cm using a wall-mounted stadiometer (SECA, Chino, California, USA) and body weight is measured to the nearest 0.1 kg with a Tanita WB-800S Plus digital scale (Arlington Heights, Illinois, USA) without shoes. Waist and hip circumferences are measured to the nearest 0.5 cm with a Gulick II anthropometric tape (Country Technology, Gays Mills, Wisconsin, USA). Using the NHANES III protocol,¹⁰¹ waist circumference is measured at the top of the iliac crest and hip circumference is measured around the widest part of the hips/buttocks. A third measurement of waist or hip circumference is taken if the initial two measurements differ by more than 0.5 cm. Body composition (fat mass, fat-free mass and bone mineral density) is measured with dual X-ray absorptiometry (DXA) (Lunar Prodigy, General Electric, Madison, Wisconsin, USA). Female participants must complete a pregnancy test to confirm no pregnancy prior to completing the DXA scan. If a participant exceeds the height or weight limits for the DXA (ie, >6 feet tall or >300 lbs), fat mass and fat-free mass are measured via air displacement plethysmography with the BOD POD (Cosmed, Concord, California, USA). Standard manufacturer calibration and measurement procedures are followed for both the DXA and BOD POD measurements. Lastly, the SHOW study aims to assess additional measures of cardiometabolic risk, including resting blood pressure measured with the Omron HEM-907XL Professional Blood Pressure Machine (Omron Healthcare, Lake Forest, Illinois, USA), as well as fasting serum blood biomarkers (ie, total, Low-density cholesterol (LDL) and high-density cholesterol (HDL) cholesterol, glucose, and insulin) related to metabolic syndrome and insulin resistance.

Energy balance

Daily energy balance over the 7-day data collection burst is calculated by subtracting total energy intake from energy expenditure (based on resting, postprandial and physical activity energy expenditure measures).

On days 1–4 of each burst, energy and macronutrient intake (from foods and caloric beverages) is self-reported with a food diary.¹⁰² On days 5–7 of each burst, energy and macronutrient intake is measured with a validated food menu.⁶⁹ This food menu contains 60 food and beverage options (eg, lasagna, yoghurt, pizza, juice, cereal, granola bar, 8 different types of fruits and vegetables, crackers, cookies). Participants self-select items from this menu that they may want to consume during lunch on day 5 (consumed inside the laboratory), as well as for the remainder of day 5, and days 6 and 7 (consumed outside of the laboratory). Two portions of each selected food item are individually prepared, weighed with a digital food portioning scale (AvaWeigh Commercial Scales, Lancaster, Pennsylvania, USA) and served (inside the laboratory) or packed into separate containers/bottles and placed into a lunch box container for participants to bring home with them (outside of the laboratory). Participants are instructed to bring the remaining food items/leftovers, wrappings and containers/bottles back to the laboratory following the end of the measurement period. All remaining items are weighed following the in-laboratory lunch and out-of-laboratory measurement period. Energy and macronutrient intake for days 1–7 of each data collection burst is determined and analysed with the Nutritionist Pro diet analysis software (Axxya Systems, Redmond, Washington, USA).

Water intake on days 1–7 is measured with the PÜL SmartCap water bottle (Hyduro, San Diego, California, USA). Participants are instructed to drink water (only) from the bottle and the smart cap uses proprietary, real-time tracking technology to monitor the amount of fluid exiting the cap (consumed by the participants). Data on water intake are retrieved at the end of the data collection burst by connecting the smart cap to the PÜL App.

Resting energy expenditure following a 12-hour overnight fast and the thermic effect of food following standard breakfast intake are measured with the Quark RMR indirect calorimeter with a plexiglass hood (Cosmed, Concord, California, USA) on day 5. On arrival at the laboratory, participants are instructed to rest quietly on a bed for 20 min prior to assessing resting metabolic rate for 30 min (including a 5 min habituation period). Following standard breakfast intake, 30 min measurement periods performed every hour over 3 hours are completed to assess post-prandial energy expenditure. A mean value (kcal/min) is calculated for each 30 min measurement period and extrapolated over 60 min for each hour of measurement. The thermic effect of food is then determined by subtracting resting energy expenditure from postprandial energy expenditure during each hour of measurement, as well as over the entire 180 min measurement period.¹⁰³

Physical activity energy expenditure is derived from data collected with the waist-worn Actigraph GTx9 Link accelerometer (ActiGraph, Pensacola, Florida, USA) over the 7-day burst. Like the wrist-worn device, participants are instructed to remove the waist-worn device for water-based activities and periods of non-wear time >10 min are reported in an accelerometer log sheet. Accelerometer data are collected at a sampling rate of 30 Hz and aggregated to 60 s epoch files for analysis by the Actilife software (currently V.6.13.5). A waist, rather than wrist, placement was chosen because accelerometer counts (and derived physical activity values) are substantially higher for wrist-mounted compared with waist-mounted devices.¹⁰⁴ At this time, the Freedson VM3 Combination algorithm¹⁰⁵ is used to derive daily physical activity energy expenditure (kcal/day). Additional algorithms or processing methods to derive physical activity energy expenditure may also be used at the conclusion of data collection.

Appetite and food reward

Fasting and postmeal (measured every 30 min for 3 hours following standard breakfast intake on day 5) appetite and thirst sensations are recorded with 100 mm computerised visual analogue scales.¹⁰⁶ The following four questions are asked at every time point to assess feelings of appetite¹⁰⁷: desire to eat (‘How strong is your desire to eat?’; very weak—very strong), hunger (‘How hungry do you feel?’; not hungry at all—As hungry as I have ever felt), fullness (‘How full do you feel?’; not full at all—very full), and prospective food consumption (‘How much food do you think you could eat?’; nothing at all—a large amount). The following six questions are asked at every time point to assess feelings of thirst¹⁰⁸: feelings of thirst (‘How thirsty do you feel right now?’; not at all thirsty—very thirsty), pleasantness of water intake (‘How pleasant would it be to drink some water right now?’; very unpleasant—very pleasant), mouth dryness (‘How dry does your mouth feel right now?’; not at all dry—very dry), mouth taste (‘How would you describe the taste in your mouth?’; normal—very unpleasant), stomach fullness (‘How full does your stomach feel right now?’; not at all full—very full) and stomach sickness (‘How sick to your stomach do you feel right now?’; not at all sick—very sick).

Fasting and postmeal (30, 60, 90, 120, 150 and 180 min poststandard breakfast intake) blood samples are processed for analysis of total and active ghrelin, leptin, peptide YY (PYY) and glucagon-like peptide-1 (GLP-1). Specifically, dipeptidyl peptidase-4 (DPP-4) and aprotinin are added to the first precooled EDTA plasma tube (3 mL) to assess PYY and GLP-1 (labelled ‘satiety hormones’). Pefabloc (AEBSF) is added to the second precooled EDTA plasma tube (3 mL) to assess active and total ghrelin levels (labelled ‘hunger hormones’). This sample is further treated by acidifying the plasma using hydrochloric acid. Leptin is analysed using a serum tube. The serum and plasma tubes are centrifuged at 3000 rpm for 12–15 min at 4°C. The samples are then divided into multiple aliquots of ~500 µL to assist with the freeze/thaw cycle. Once aliquoting is complete, samples are immediately stored in a −80°C freezer. For both self-reported and haematological appetite markers, as well as self-reported thirst measures, postbreakfast area under the curve will be calculated with the trapezoid method, as previously described.¹⁰⁹

Food reward is assessed pre-ad libitum and post-ad libitum lunch intake on day 5 with the LFPQ.^{110 111} This validated, computer-based behavioural task measures food wanting and liking for 16 food items that vary in both fat content and taste: high-fat savoury (eg, pizza), low-fat savoury (eg, carrots), high-fat sweet (eg, ice cream) and low-fat sweet (eg, strawberries). The 16 food items presented to each participant were chosen based on personal preferences/familiarity during the preliminary session. During the forced choice portion of this task, each food image from a given category is presented with every other image from a different category in turn. For each pair of food images presented, participants are asked to select the food item they would ‘most want to eat now’. A standardised implicit wanting score for each food item is calculated as a function of the reaction time in selecting the food item adjusted for the frequency of choice for food items selected in each category.¹¹¹ Participants are also asked to rate the extent to which they ‘like’ (‘How pleasant would it be to experience a mouthful of this food now?’) or ‘want’ (‘How much do you want to eat this food now?’) each randomly presented food item using a 100 mm Visual Analogue Scale, with the score for each item (within each category) used as a measure of explicit liking and wanting, respectively.

Salivary cortisol

Participants are asked to provide saliva samples on a predetermined day among days 1–4, days 6 and 7. On each day of collection, participants are asked to provide five samples at the following time points: immediately after waking, 30 min after waking, 45 min after waking, ~8 hours after waking (late afternoon) and before bed. Per cortisol awakening response recommendations, participants are instructed to take the three morning samples before eating, drinking or brushing their teeth.¹¹² Each saliva sample is collected via passive drool into pre-labelled polypropylene storage vials. Participants are instructed to record the clock time of each sample collected and to store the samples in a home freezer (−20℃) until returned to the laboratory. On return, saliva samples are stored in a −80℃ freezer until assayed.

Salivary cortisol will be assessed with ELISA kits following manufacturer instructions and using recommended quality control. All samples from a given participant will be assayed on the same kit lot and on the same plate whenever possible. Samples will be measured in duplicate and if the coefficient of variation is greater than 20%, samples will be reassayed. Cortisol awakening response will be calculated as the area under the curve for the awakening and 30 and 45 min postawakening samples¹¹³ with respect to ground and with respect to increase. Diurnal cortisol slope will be quantified as the value of the slope coefficient of a linear ordinary least squares regression that is fit to three saliva samples from the cortisol profile: the first awakening sample, 8 hours after awakening (late afternoon) and before bed (nmol/L/hour).

Hair cortisol

Participants are asked if they are willing and able to provide a hair sample to examine chronic stress (~3 months) cortisol measure. On day 0, participants able to or willing to provide a hair sample complete a Hair Care form which includes information about shampoo frequency and hair chemical use (eg, hair dyes, bleach, chemical straighteners or permanently waved).¹¹⁴ The form will be attached to the hair sample and considered when the sample is analysed. Before the hair sample is taken, a research staff reviews the procedures with participants. Participants are then asked to hold a mirror to view the location and amount of hair to be sampled during the collection process. A 3 cm, 5 mg hair sample (~20 strands) is obtained from the vertex area (posterior head) using hair shears. The hair sample is cut as close to the scalp as possible. Hair samples are then placed on foil with the scalp end marked on the foil and secured with painter’s tape and placed in an individually labelled envelope and stored in a 2℃–4℃ fridge. Hair samples are stable at room temperature for years.¹¹⁵ Hair samples will be shipped in batches to the Biobehavioral Laboratory at the University of North Carolina at Chapel Hill for analysis.

Hydration status

Both urinary and haematological hydration markers will be collected. Participants are provided with 3 L horizontal containers (VWR, Radnor, Pennsylvania, USA) in reusable opaque bags to bring home with them and to collect all urine produced over a 24-hour period for one preselected day among days 1–4 and days 6–7. Male participants directly void into the container’s 10 cm diameter opening. Female participants are given a graduated specimen pan (Model DYND36600, Medline Industries, Northfield, Illinois, USA) to void all urine, and then transfer to the 3 L containers. Each 24-hour urine sample is measured for urine volume, urine osmolality, urine-specific gravity (USG) and urine colour. Urine volume is measured to the nearest 0.001 kg using a digital scale (Ranger 3000, OHAUS Corporation, Parsippany, New York, USA) where 0.001 kg is equal to 1 mL. Urine osmolality is determined in duplicate using freezing point depression (Advanced Instruments, Norwood, Massachusetts, USA). However, if the first two osmolality measures vary by ≥3 mOsm/kg then a third measure is taken and averaged. If the first three measures differ by ≥3 mOsm/kg then two additional measures are taken (five measures total), and the median is reported. Urine-specific gravity is measured using a refractometer (Reichert AR200, Reichert Technologies, Buffalo, New York, USA). Urine colour is assessed using a validated 8-point urine colour chart.¹¹⁶ A fasting serum tube will be used for copeptin and serum osmolality analyses. Copeptin will be assessed with ELISA kits using an Epoch microplate spectrophotometer plate reader (BioTek, Winooski, Vermont, USA). Serum osmolality will be assessed in duplicate using the freezing point depression method (Advanced Instruments, Norwood, Massachusetts, USA) and will follow the same procedures as the urine osmolality measures if the first two measures differ by ≥3 mOsm/kg.

Sample size and power considerations

We aim to include 150 individuals with 3 measures of obesity markers, 9 measures of salivary cortisol profiles and urinary hydration markers, and 21 measures of accelerometer-based sleep and energy balance per person across 3 bursts. Because we will only have three repeated measures of obesity markers, with only one time-invariant predictor for these models, these analyses will be our limiting factor with respect to power. Therefore, if we have power to detect effects of baseline sleep variables on changes in obesity markers, we will have power for our other analyses. We constructed a simulation model with 1000 replications from the following population-generating model: $y_{ti}=80+.5x_{1ti}+.3x_{2i}{x}_{1ti}+u_{0i}+u_{1i}{x}_{1ti}+r_{ti}$ , with $r_{ti}~N(0,2)$ , $VAR(u_{0i})=3.0$ , and $VAR(u_{1i})=.2$ . These parameters approximate the average weight of a North American adult (80 kg) with variance of 3.0 kg, an average weight gain of 0.5 kg per year with random variance of 0.2 kg, plus a 0.2 kg increase in weight per year for every 1 hour decrease in baseline sleep duration.^{8 117} The residual error variance reflects a measurement error of 0.5 kg for each burst. The most difficult effect to power is the interaction between time and baseline sleep variables. Of the 1000 simulated samples with n=150, the interaction term was significant 90% of the time with a two-sided α=0.05. We also conservatively allowed for a 15% attrition rate. Hence, we conducted a separate simulation to assess power after deleting 15% of 12-month bursts. Under this condition, we expect to have 85% power to detect a significant interaction term. Lastly, the use of multiple imputation techniques will help mitigate some loss of power due to missing data.

Analyses: exploratory phase

We will visualise data distributions, identify and correct or winsorise outliers, note any necessary transformations to data distributions, identify potential non-linear associations between variables and identify potential collinearity. We will use multilevel multiple imputation to handle any missing data. Results will also be compared between complete-case and imputed cases in sensitivity analyses to examine the impact of attrition and partial data completion. Potential confounders will include variables captured via questionnaires during the preliminary and day 0 sessions, such as age, biological sex and gender, chronotype, substance use, medication use (baseline and change from baseline to control for potential weight gain due to new medication use), mental health outcomes (eg, perceived stress, depression and anxiety), household responsibilities and income-to-needs ratio. We will consider confounders in the aim-specific analyses if their inclusion cause parameter estimates of the aim’s primary interest to change by ≥10%.¹¹⁸

We will also conduct initial analyses of sleep variables. First, we will examine distributions, visualisations and pairwise correlations (eg, scatterplots with LOESS trend lines) of sleep variables (eg, sleep duration, efficiency, timing, onset latency, stage duration). We will examine variability and robust measures of central tendency (eg, median) around the day-level data for each sleep variable. This will help us determine if the arithmetic mean over each burst is appropriate or an alternative (eg, winsorised mean) should be used. We will examine overlap in sleep variables through eigen analysis, including estimation of condition number.¹¹⁹ If empirical overlap is high, we will use the first principal component from principal component analysis in place of multiple sleep variables in sensitivity analyses and compare findings, including model predictive accuracy. If it is apparent that night-to-night variability is substantially heterogeneous through equal variances testing and visualisations, we will model the coefficient of variation of sleep variables across bursts separately or additionally in further sensitivity analyses. We will consider AHI<15 vs ≥15 (mild vs moderate-severe symptoms of sleep-disordered breathing) in sensitivity analyses. We will also use correlations to explore associations between social desirability scores and responses to self-report measures.

Analyses: study aims

To examine the prospective effects of insufficient sleep on weight-related mechanisms, we will assess within-burst prospective associations between each sleep variable on day t with energy balance, appetite, food reward, salivary cortisol profiles and hydration status on day t+1 to maintain temporal ordering. This same approach will be used to assess within-burst prospective associations between daily measures of race-related stress on sleep variables and weight-related mechanisms, with a primary focus on salivary cortisol profiles. For between-burst analyses, we will use generalised linear mixed models (GLMMs) to test for significant changes between baseline, 6-month and 12-month time points in sleep variables, weight-related mechanisms and race-related stress. An unstructured covariance matrix for the random effects will be used. We will account for any residual autocorrelation of adjacent time points not already accounted for by the random effects using an autoregressive correlation matrix for residuals. The fixed effect of sleep variables on subsequent weight-related mechanisms, as well as race-related stress on sleep and salivary cortisol profiles, will be of primary interest.

The prospective effects of baseline sleep variables and race-related stress on changes in obesity markers over the 12-month period will also be tested with a GLMM, modelling obesity markers at each burst as a function of sleep and race-related stress variables at baseline. The random intercept will adjust for individual differences in obesity markers at baseline (thus removing regression to the mean effects), and the random slope will reflect individual differences in the degree to which sleep/race-related stress predicts obesity marker changes. The random intercept and slope will be allowed to covary. The fixed effect of sleep and race-related stress on obesity markers is of primary interest. A significant main effect of sleep/race-related stress will indicate an association between these predictors and obesity markers at baseline. A significant interaction between sleep/race-related stress and time will reflect a prospective effect of these predictors on obesity marker changes. We will again use GLMM to assess whether measures of home and neighbourhood environments (mean score on the Windshield Survey and ASE) at baseline predict subsequent changes in sleep variables, weight-related mechanisms (primarily, hair and salivary cortisol profiles), and obesity markers over the 12-month period. Lastly, potential modifying effects of individual-level sociodemographic and health-related factors (eg, chronotype, substance and medication use, cultural identity/strength, and coping mechanisms) on observed associations between sleep and race-related stress with weight-related mechanisms and obesity markers will be explored in sensitivity analyses.

To assess aim 3, we will use a causal mediation approach that accounts for time-varying confounding in the mediational pathway,^{120 121} predicting weight-related mechanisms and obesity markers at time t+1 from sleep variables at time t, and predicting sleep variables and obesity markers at time t from race-related stress and home/neighbourhood environments at time t−1. We will first construct inverse probability weights separately for the mediator and outcome models (thereby controlling for a different set of confounding effects for the exposure-mediator and mediator-outcome pathways). After computing weights that adjust for confounding, the macro will apply them to a set of marginal structural models to estimate causal mediation effects. We will use robust SEs (sandwich estimators) to adjust for use of weights and non-independence of the repeated measures.

Ethics and dissemination

The SHOW study protocol was approved by the University of North Carolina at Greensboro’s Institutional Review Board (#IRB-FY23-271; PI: JM). This study was funded in May 2023 and data collection began in November 2023. Data collection is expected to be completed by October 2027, at which point data analyses and dissemination efforts will be begin. Study findings will be disseminated through peer-reviewed publications, presentations at scientific meetings, as well as reports, briefs and infographics that will be developed for lay and community audiences. Notably, CAB members will assist with creating and reviewing these dissemination documents designed for lay and community audiences that will promote the importance of sleep health and share summaries of study-related findings.

Black emerging adults who experience greater race-related stress, insufficient sleep or live in more disadvantaged home and neighbourhood environments may be more vulnerable to obesity and cardiometabolic disease. Research is needed to determine how these factors impact weight-related mechanisms and obesity markers in this population; data can then be used to inform intervention efforts to reduce sleep and obesity disparities during a key developmental period when many lifelong health behaviours are established. Therefore, the primary aim of the SHOW study is to understand the prospective effects of sleep, race-related stress, as well as home and neighbourhood environments on weight-related mechanisms and obesity markers in black emerging adults. This study will also explore the influence of individual-level sociodemographic and health-related factors (eg, chronotype, substance and medication use, cultural identity/strength, and coping mechanisms) as potential moderators of observed associations between sleep and race-related stress with weight-related mechanisms and obesity markers in this population.

The SHOW study will address several gaps within the scientific and medical literature. First, this study includes measures of hydration status, an important component of body weight regulation that is often overlooked in the energy balance and obesity literature.⁴⁶ Hydration status has been inversely associated with obesity and mortality risk.⁵³ While the literature on associations between sleep and hydration is currently scarce,¹²² recent studies have reported that habitual short sleep duration is associated with a greater risk of underhydration,⁶⁴ and imposed underhydration within a laboratory setting was associated with lower REM sleep duration.¹²³ A primary fluid regulatory hormone, arginine vasopressin, follows a diurnal rhythm within the brain and cerebrospinal fluid and progressively increases during sleep to offset water losses via the reabsorption of water in the kidneys. Insufficient sleep may disrupt this pattern and, therefore, impact hydration status.^{122 124} Copeptin (a prohormone derived from the same precursor peptide as arginine vasopressin) and 24-hour urine osmolality are commonly used as stable and sensitive markers of hydration status¹²⁵ and are included measures in the SHOW study.

Second, the SHOW study includes multiple self-reported, objective and biomarker measurements for up to 7 days over 3 bursts (baseline, 6-month follow-up and 12-month follow-up) to better characterise sleep and weight-related mechanisms over time. This approach allows us to assess both short-term (within 1 week) and long-term (12-month) changes (or stability) in multiple health behaviours (sleep, diet and physical activity) and weight-related mechanisms (energy balance, hydration status and cortisol profiles) which may contribute to a heightened risk of weight gain/obesity and cardiometabolic disease during emerging adulthood. Furthermore, the SHOW study includes comprehensive measures of habitual sleep via three methods (self-report, actigraphy and polysomnography) at home. This multimethod approach allows us to capture information on habitual sleep patterns over a 7-day burst, as well as unique information about sleep stage duration and sleep-disordered breathing over 2 days with polysomnography, thus providing the most comprehensive assessment of sleep in a sample of emerging adults.

Lastly, the SHOW study focuses exclusively on black emerging adults, an understudied population despite compelling evidence that this population has the highest risk for both short sleep duration^{3 16–18} and obesity.^{19 20} This disparity has been observed for many years, yet there is still very little evidence for the mechanisms or factors that contribute to this disparity. Hence, our within-race investigation of black emerging adults will allow us to study pathways and mechanisms linking sleep with obesity markers, as well as explore the influence of certain sociodemographic factors (eg, socioeconomic status) that often confound associations between race and health outcomes in multi-race studies in the United States.

The social-ecological model of sleep and health first described by Grandner et al ¹²⁶ was used to inform our research aims. This model suggests that individual-level factors (eg, behaviours, beliefs, biology and health) are the proximal drivers of sleep.^{34 126 127} These factors are embedded within social-level factors (eg, home and neighbourhood environments, relationships, culture, socioeconomics), which themselves are embedded within societal-level factors (eg, public policy and geography).^{34 126 127} This model also highlights the downstream consequences of insufficient sleep on adverse health outcomes, such as obesity, cardiometabolic disease and mortality.¹²⁶ Furthermore, Grandner et al have highlighted key factors that influence racial disparities in sleep (eg, perceived racial discrimination, household responsibilities)³⁴ and the sleep–obesity association (eg, energy intake, haematological markers of appetite and cardiometabolic disease risk)¹²⁷; data which are being measured in the SHOW study. Therefore, the SHOW study will address multiple gaps in the sleep and chronic disease risk literature by capturing both individual-level and contextual-level factors that may have a ‘downstream’ impact on overall sleep quantity and quality, as well as obesity and cardiometabolic disease risk, in black emerging adults. Doing so will help to identify targets for sleep-related interventions to reduce the risk of obesity and cardiometabolic diseases in black adults, particularly during a developmental period (emerging adulthood) when many health behaviours are more malleable but also become increasingly stable. At the same time, this research may identify contextual factors that lead to inequitable experiences of sleep which, although difficult to change, would highlight a need for additional research, policy and initiatives to address these contextual inequities.

In conclusion, the SHOW study will recruit 150 black emerging adults who will partake in three data collection bursts over a 12-month period, with each burst lasting 7 days. Multiple self-reported, objective and biomarker measurements will be collected during each burst in an effort to understand the prospective effects of sleep, race-related stress, as well as home and neighbourhood environments on weight-related mechanisms (ie, energy balance, appetite and food reward, cortisol profiles and hydration status) and obesity markers (ie, body weight, waist circumference and body composition) in this population. The within-race investigation of black emerging adults in this study will also allow us to distinguish the confounding role of certain sociodemographic factors from race, as well as explore the influence of additional health-related factors on observed associations between sleep and race-related stress with weight-related mechanisms and obesity markers within this understudied population. Insights provided by these data could contribute to our understanding of the role of sleep in obesity and cardiometabolic disease risk in black emerging adults and ultimately help to address individual and contextual-level factors that impact sleep, obesity and cardiometabolic disease risk to reduce health disparities in this population.

Ethics statements

Patient consent for publication

Not applicable.