Childhood Environments and their Relationship with Sleep and Ambulatory Blood Pressure in College Students

Abstract

Objective:

Investigate whether psychosocial risk in the childhood family environment moderates the relationship between childhood socioeconomic status (SES) and sleep, and the relationship between childhood SES and ambulatory blood pressure (ABP) in college students, two factors that are linked to future risk for cardiovascular disease.

Participants:

124 American college students.

Methods:

Childhood SES and psychosocial risk in childhood family environments were measured by self-report instruments. Sleep was measured with self-report and actigraphy (over 5 days) and ABP over a 2-day period.

Results:

Linear regressions adjusting for age, sex, current SES, and current depressive symptoms indicated that SES and psychosocial risk in family environments during childhood interact to inform sleep quality, actigraphy derived wake after sleep onset (WASO), actigraphy derived Sleep Efficiency (SE) and ABP.

Conclusions:

Psychosocial risk in the childhood family environment may offset previously documented relationships between childhood SES and health-relevant outcomes in college students.

Introduction

Greater exposure to stress early in the lifespan may affect the development of health behaviors and processes and the programming of physiological systems.^1-2 In turn, these patterns and physiological programming may persist into adulthood,^3-4 and may partially explain links between early life environments and health throughout the lifespan. One source of possible stress exposure during childhood relates to socioeconomic status (SES).

Previous studies suggest that childhood SES relates to health throughout the lifespan such that individuals raised in low SES environments have worse health in adulthood and lower life expectancies compared to individuals raised in high SES environments, and this relationship is largely independent of adult SES.^5-8 Low SES environments are associated with greater loads of both physical and psychological stressors.^9-10 For example, individuals raised in low SES may be exposed to higher levels of noise, pollution, crime, and physical crowding compared to those individuals raised in high SES environments.¹¹ Low SES environments are also characterized by less control and more unpredictability, all while having fewer resources to manage these stressors.⁶ Early life exposure to adversity and stress may prime and sensitize individuals to stress experienced later in life.¹² A related body of work indicates that the dynamic of the family environment during childhood may have independent effects on health across the lifespan.¹³ Specifically, family environments which are characterized by high levels of adversity and low levels of warmth and affection can negatively impact adult health by shaping psychosocial functioning.^13-14

One focus of the literature investigating the relationship between adversity in childhood and later health centers on risk for cardiovascular disease (CVD).^15-18 The findings from this body of work are somewhat inconsistent, and the results appear to be dependent on the type of adversity considered or the method of measurement. For example, one investigation did not find evidence of a relationship between various types of childhood psychosocial adversity and future risk for CVD,¹⁹ while a prospective research study found that individuals from socioeconomically disadvantaged backgrounds during childhood were more likely to develop a CVD risk factor or have a CVD event in adulthood.²⁰ There has been less focus on psychosocial risk in childhood family environments as a predictor of future CVD risk, however one study found that, greater psychosocial risk in the family environment associated with greater CVD risk in adulthood (e.g. hypertension, physical inactivity).²¹ Importantly, this body of work has not considered the interactive relationships between socioeconomic disadvantage during childhood and psychosocial risk in the family environment as a predictor of risk factors for CVD.

Poor sleep may be one pathway that links early life exposure (e.g., low SES and risky family environments) to greater risk for CVD in later life, given that sleep has previously associated with CVD mortality and risk.²² Low childhood SES predicts worse sleep quality in childhood,²³ and in adulthood independent of SES levels in adulthood.^24-26 Furthermore, the well documented relationship between low SES and poor health is in part accounted for by differences in sleep patterns and outcomes.²⁷ There is also evidence early life trauma and abuse affects sleep quality into adulthood. For example, greater reports of emotional abuse in childhood associated with more sleep complaints in older age,²⁸ and greater reports of physical, emotional and sexual abuse in childhood was associated with global sleep pathology in a national sample of adults in the United States.²⁹ However, less is known about how the overall psychosocial nature of the family environment during childhood may independently inform sleep. Compromised sleep patterns and poor sleep quality linked to early life adversity may limit the potential for sleep to buffer the negative effects of stress and may exacerbate the consequences of current and future stress.³⁰

Blood pressure (BP) was first established as a risk factor for future CVD risk based on findings from a 30-year longitudinal study in the United States, the Framingham Heart Study (FHS). The data from the FHS indicated a continuous graded relationship between BP and future risk for CVD.³¹ In previous work, socioeconomic disadvantage has predicted both resting BP and BP reactivity in response to psychological stress in lab settings.^32-33 Blood pressure can also be measured outside of the laboratory using ambulatory assessments. Ambulatory blood pressure (ABP) provides a cumulative measure of exposure to high blood pressure levels and is thought to reflect multiple factors including differences in real-life environments, daily behaviors, and lifestyles. Previous research found that ABP was a better predictor of clinical and subclinical endpoints of CVD compared to measures of resting BP.³⁴ The relationship between socioeconomic status and BP appears to extend outside of the lab, with socioeconomic disadvantage predicting higher average ambulatory blood pressure, in adolescence and adulthood.^35-36 Given that ABP has been identified as a predictor of future CVD risk, it is important to understand the independent and interactive relationships between different dimensions of the childhood environment (i.e. socioeconomic disadvantage and risk in family environments) and this CVD risk factor.

There is little research testing a possible interactive effect of childhood SES and psychosocial risk in childhood family environments on sleep quality, and to date, no research has considered whether this interaction may inform patterns of ABP. One previous study has examined the independent and interactive contributions of different components of the childhood environments as predictors sleep quality in college students. This research found that students exposed to both low childhood SES and a family environment high in psychosocial risk had the worst sleep quality, while students from high SES backgrounds and nurturing family environments reported the best sleep quality.³⁷ Furthermore, a less risky family environment associated with better sleep quality in students from both low and high childhood SES, highlighting the need to consider psychosocial risk in childhood family environments when examining relationships between childhood SES and sleep later in the lifespan.

It is important to note that the majority of the extant literature focused on relationships between childhood environments and health-relevant outcomes focuses on these relationships during middle to later adulthood. Previous studies suggest that childhood adversity predicts health-relevant outcomes including compromised sleep quality in college students,^37-39 and given that sleep quality in college students is particularly compromised compared to other age groups,^30,40 more research is needed to better understand how childhood environments inform sleep quality during this period. In a similar manner, elevated blood pressure in adolescence and early adulthood predict CVD risk into adulthood.⁴¹ As such, it is important to identify predictors of blood pressure patterns in college years in order to identify at-risk individuals.

The present study seeks to replicate previous work showing that risk in family environments moderates the relationship between childhood SES and subjective sleep quality in college students,³⁷ and extend this work by examining in a sample of college students whether the interaction between childhood SES and psychosocial risk in childhood family environments predicts 1) actigraphy derived measures of sleep and 2) average ABP, two outcomes which are related to risk for future CVD.

Methods:

Participants (N=124) were students enrolled in introductory psychology courses at a state university in the United States. Students were asked to participate in research studies as part of their coursework. Students who did not wish to participate in studies were given other options to complete their course requirements. Students who elected to participate in this research were asked to undergo a five-day monitoring period and complete beginning and end of day sleep diaries. Course credit was awarded upon completion of the monitoring period and questionnaires. The study was made available to all students in the participant pool and students were excluded if they did not have access to a smartphone, had shift work sleep disturbances, reported any sleep disorders, or were taking any medications that could affect their sleep. All materials and procedures were approved by the university’s institutional review board, and informed consent was obtained from all participants.

Procedures

Participants had an initial lab appointment in which they completed questionnaires and were outfitted with a wrist accelerometer. Participants were instructed to wear the accelerometer (Actigraph GT9X Link, Penascola, FL) watch on their non-dominant wrist) continuously over a seven-day monitoring period, while also answering daily surveys about sleep upon waking and going to bed. During this visit, they were trained to use an automated ABP monitor (Oscar 2 oscillometric monitor; Suntech Medical, Inc, Morrissville, NC).⁴² During two days of the monitoring period they wore the ABP cuff from the time they completed their morning survey until they completed their end of the day survey. Participants then returned to the lab after completion of the monitoring period to return devices and receive course credit for their participation.

Measures

Actigraphy measures of Sleep

Actigraphy is a cost-effective and non-invasive method of measuring sleep quality that involves the use of a portable device that records movement data over an extended period of time. Actigraphy has been shown to be a reliable measure of sleep in the natural environment.⁴³ Actigraphy measurement is an objective measure of sleep, in contrast to subjective, self-report measures like sleep diaries or validated scales. Subjective measures may suffer from inaccuracies in overestimating or underestimating sleep-time and wakefulness or ratings, or self-report bias in individuals reporting socially desirable answers that do not accurately reflect their sleep. As such, sleep was measured with self-report measures and actigraphy.

Participants were instructed to wear an accelerometer (Actigraph GT9X Link, Penascola, FL) on their non-dominant wrist over a week-long period. Data were collected using the zero-crossing mode in one-minute epochs and sleep parameters were estimated using a medium detection threshold and the Cole-Kripke algorithm.⁴⁴ Only participants with five nights (weekdays) of valid data were included in subsequent analyses as prior research indicates that a period of five days is ideal to obtain an accurate reflection of an individual’s sleep-wake pattern.⁴⁵ Based on the given criteria, four participants were excluded from analyses. Nighttime sleep periods were determined using self-reported sleep onset and offset times from the sleep diaries in combination with visual inspection of activity levels. Total sleep time was the sum of the number of sleep minutes during the night-time sleep period. Wake after sleep onset (WASO) was calculated as the number of minutes spent awake within the total night-time sleep period. Sleep efficiency (SE) was calculated as the number of sleep minutes divided by the total duration of time between sleep onset and sleep offset, multiplied by 100.

Sleep diaries

Participants were sent two daily questionnaires to track sleep: one each morning upon waking up and one each night before sleep. These sleep questionnaires were delivered by text message to participants’ mobile phones using the Qualtrics platform. These questionnaires asked participants to record their sleep and wake times, time spent in bed, time spent asleep, and any sleep disturbances. Before going to sleep, participants were asked to record any naps taken, any substances or medication use, whether they worked and whether they exercised during that day.

Ambulatory Blood Pressure

ABP was assessed using a two-day protocol (two weekdays). Participants wore the oscillometric Oscar 2 ABP monitor (SunTech Medical, Inc, Morrisville, NC). On each monitoring day, the cuff worn on the upper arm inflated every two hours during waking hours and the monitor recorded blood pressure. Text messages were sent to participants the night before each of these days and again in the morning to remind them to place the cuff on their arm. The Oscar 2 has been validated to the standards of international protocols.⁴² Data from the Oscar 2 ABP monitor was downloaded when participants returned to the lab. Using measurements from the 2 days of blood measure monitoring, we created an average systolic ABP score and an average diastolic ABP score. Six participants did not have complete ABP data and so were excluded from these analyses.

Questionnaire Measures

Subjective Sleep Quality

Sleep quality over the past 30 days was measured with the PSQI, a measure of global sleep quality over the past 30 days.⁴⁶ The 19-item scale was used to derive seven component scores: sleep quality, sleep latency, sleep duration, habitual sleep efficiency, sleep disturbances, sleep medication, and daytime dysfunction. A global PSQI score was produced from the sum of the seven component scores with a possible range of 0-21, where higher scores indicate poor sleep quality. Global PSQI scores were calculated for each participant (M [SD] = 6.58 [3.21]). A global PSQI score of 5 or more reflects poor sleep, and the average global PSQI score in this sample was above this cutoff. Internal consistency in this sample between the seven component scores was α = .77. Global sleep quality among college students is commonly measured by the PSQI.⁴⁷

Subjective childhood SES

Subjective SES during childhood was measured with the MacArthur scale of subjective childhood SES. The scale consists of a 10 rung ladder visual analogue scale on which participants rank where they feel their family stood during their childhood relative to other families in the United States.⁴⁸ Scores ranged from 1 (lowest SES) to 10 (highest SES) (M [SD] = 6.42 [1.75]). The scale explains that families with more money, education, and better jobs are represented by the top of the ladder, whereas the bottom of the ladder represents families who were worse off, had the least amount of money or education, and had less-respected jobs or were unemployed.

Risk in early family environments

The Risky Families questionnaire was used to measure levels of psychosocial risk in childhood family environments.¹³ Participants indicate exposure during childhood to physical, mental, and emotional neglect or abuse, as well as levels of warmth and affection they received from family during this same period. Using a 5-point Likert scale (1 = not at all and 5 = very often), participants rated the frequency of exposure to certain events or situations that occurred in their home environment from the ages 5-15. For example, the measure asks, “How often did a parent or other adult in the household make you feel that you were loved, supported, and cared for?” and “How often did a parent or other adult in the household push, grab, shove or slap you?” Positively worded items are reverse-coded and the 13 items are summed (possible range 5-65) to quantify overall risk in early family environment (M [SD] = 30.92 [10.27]).

Current depressive symptoms

We used the Hospital Anxiety and Depression Scale (HADS) to obtain a measure of current depressive symptoms.⁴⁹ The measure includes a seven-question depression subscale (HADS-D) Participants rate each statement on a scale 0 (absence of symptoms) to 3 (most acute symptoms), providing a possible HADS-D subscale total of 0 to 21, where higher scores indicate higher levels of depressive symptoms. The HADS-D subscale had good internal validity (a=.72).

Current Subjective SES

Current subjective SES was measured with the MacArthur scale of subjective SES.⁵⁰ Participants were asked to place an “X” on a rung of the same ladder visual described above to indicate where they felt they currently stood relative to others in society with regard to occupation, money, and education (M [SD] = 6.22 [1.65])

Data analyses

Analyses were conducted using SPSS (version 24; IBM, Armonk, NY). Hierarchical linear regression models were used to conduct the main analyses. Continuous covariates were centered prior to use in statistical models. In all reported linear regression models, age, sex, and depressive symptoms were included as covariates, given their relationship to sleep and BP.^51-58 To test our hypothesis, we created an interaction term between subjective childhood SES and risk in early family environments and examined whether this interaction term predicted sleep outcomes and average ambulatory blood pressure. We used the Johnson-Neyman technique⁵⁹ to identify specific values of risk in family environments for which the relationship between childhood SES and our outcomes was statistically significant.

Results

Descriptive statistics are listed in Table 1 and bivariate correlations between main variables of interest are listed in Table 2. As expected, we observed statistically significant correlations between depressive symptoms and other predictors and outcomes including risk in family environments, childhood SES, subjective global sleep quality, WASO, and SE. Also, as expected, we observed significant correlations between childhood SES and each of our outcomes and significant correlations between risk in family environments and each of our outcomes.

Table 1.

Descriptive Statistics (N=120).

	Mean	SD	Observed Range
Age	20.07	5.16	17-56
Sex (% female)	74.6%
Subjective SES (range:1-10)	5.93	1.58	2-10
Current depressive symptoms (range: 0-21)	3.90	2.86	0-16
Risky Family Environment (range:13-65)	30.78	10.19	16-50
Childhood Subjective SES (range:1-10)	5.83	2.73	1-10
PSQI Global Sleep Quality (range: 0-21)	6.70	2.21	2-12
Actigraphy average sleep efficiency	89.35	4.62	75.20-97.27
Actigraphy average WASO (minutes)	39.75	14.43	9.67-76.33
Actigraphy average total sleep time (minutes)	362.50	78.18	221.14-622.40
Average systolic ABP	120.50	12.97	100.57-155.23
Average diastolic ABP	70.32	6.41	61.71-87.86

Note: SES= Socioeconomic status, PSQI= Pittsburgh Sleep Quality Index, WASO= Wake after sleep onset, ABP= Ambulatory Blood Pressure.

Table 2.

Correlation Matrix for Key Variables. (N=120)

Variable	1	2	3	4	5	6	7	8	9	10	11
1. Age	-	−.14	−.03	−.21*	−.16	.08	−.05	.07	−.12	.17	.10
2. Sex		-	−.15	.18	−.06	−.09	.16	−.12	−.02	−.13	−.05
3. Depressive symptoms			-	.26**	.20*	.20*	.12**	.11*	.20*	.23	.18
4. Childhood SES				-	−.12*	−.28**	.20*	−.34*	.03	−.52**	−.53**
5. Risk in family environment					-	.23**	−.41**	.38**	−.09	.44**	.34**
6. PSQI global sleep quality						-	−.31**	.35**	−.10	.32**	.28**
7. Sleep Efficiency							-	−.77**	−.45**	−.41**	−.44**
8. Wake After Sleep Onset								-	−.10	.61**	.60**
9. Total Sleep Time									-	−.02	−.09
10. Systolic ABP										-	.89**
11. Diastolic ABP											-

Note:

^*p < .05 (two-tailed).

^**p < .01 (two-tailed); Sex: 1=male, 2= female; SES: Socioeconomic status; PSQI: Pittsburgh Sleep Quality Index; ABP: Ambulatory Blood Pressure

For adults, systolic ABP below 130 is considered normal and a diastolic ABP below 80 is considered normal.⁶⁰ While there is limited research on ABP in college students, the average ABP observed in this research was similar to average values of ABP in a previous investigation in college students.⁶¹ As noted previously, a score of 5 or greater on the PSQI global sleep quality scale reflects poor sleep quality, and in this sample the average PSQI global sleep quality score was M [SD] = 6.58 [3.21], indicating that this sample of college students have poor sleep quality on average, similar to numerous prior research in college student samples.^37,62-63 Average scores on the risky family questionnaire and childhood subjective SES measure were also similar to those observed previously in college samples.³⁷

Childhood SES by risky family environments predicting sleep outcomes

To see if our data replicated the findings reported in previous research, we created an interaction term between childhood SES and risk in early family environments. We entered the covariates of age, sex, subjective SES and current depressive symptoms, the childhood SES measure, and the risky family measure into block one of a hierarchical linear regression model, and the interaction term in block two. We used this model to predict each of our sleep outcomes, and average ABP. This interaction term was a significant predictor of self-reported sleep quality (β = −.36, t(111) = −4.32, p <.001, R² change = 0.12), actigraphy measured WASO (β = −.28, t(111) = −3.40, p <.001, R² change = 0.07), and actigraphy measured SE (β =.22, t(111) = 2.60, p =.01, R² change = 0.05). The interaction term was not a significant predictor of total sleep time (β = .03=.00, t(111) = .01, p =.99). The results of the regression models for sleep outcomes are listed in Table 3. The patterns of these interactions for WASO and SE are displayed in Figure 1.

Table 3.

Linear regression models with early family environments predicting sleep outcomes

	PSQI global sleep quality (N=118)		WASO (N=118)		Sleep Efficiency (N=118)
	β	t	β	t	β	t
Age	.09	1.07	.10	1.19	−.11	−1.28
Sex	.02	.22	−.05	−.64	.10	1.10
Subjective SES	.13	1.51	.05	.57	−.04	−.49
Depressive Symptoms	.17*	2.03	−.06	−.67	.02	.27
Childhood SES	−.24**	−2.70	−.27**	−3.07	.10	1.08
Risk in Family Environment	.18*	2.15	.34**	4.20	−.36**	−4.21
Childhood SES x Risk in Family Environment	−.36**	−4.32	−.28**	−3.40	.22**	2.60
R 2	.12		.07		.05

Notes:

^*p < .05

^**p <.001. Standardized estimates are displayed. Sex is coded as 1 = Male and 2 = Female.

SES= Socioeconomic Status, PSQI=Pittsburgh Sleep Quality Index, WASO= Wake after sleep onset.

Figure 1.

Average WASO and average SE as a function of reported childhood SES and psychosocial risk in childhood family environments. High values for each variable represent 1 SD above the mean, and low values for each variable represent 1 SD below the mean. Results from linear regression analyses controlling for age, sex, subjective SES, and depressive symptoms. Note: SD= Standard Deviation; SES= Socioeconomic status, WASO= Wake After Sleep Onset, SE= Sleep Efficiency.

The Johnson-Neyman technique⁵⁹ was used to identify at which value on the Risky Families Questionnaire (range 13-65) the relationship between childhood SES and our sleep outcomes became statistically significant. These analyses revealed that the association between childhood SES and the PSQI global sleep quality measure was statistically significant for individuals with scores on the Risky family measure above 28. The relationship between childhood SES and WASO and the relationship between childhood SES and SE was statistically significant for individuals with scores on the Risky Families Questionnaire above 27 and 35 respectively.

Childhood SES by Risk in Family Environments predicting Ambulatory Blood Pressure

In separate regression models with the same covariates, we tested whether the interaction between childhood SES and risk in childhood family environments would predict average ABP. This interaction term was a significant predictor of average ambulatory systolic blood pressure (β = −.20, t(107) = −2.49, p =.01, R² change = 0.04) and average ambulatory diastolic blood pressure (β = −.24, t(108) = −3.24, p <.01, R² change = 0.05). Table 4 reports the full regression model for systolic and diastolic ABP. Figure 2a and Figure 2b display the pattern of these interactions.

Table 4.

Linear regression models with early family environments predicting ambulatory blood pressure (ABP).

	Systolic ABP (N=114)		Diastolic ABP (N=114)
	β	t	β	t
Age	.09	1.13	.09	1.14
Sex	.01	.14	.09	1.19
Subjective SES	−.07	−.81	−.13	−1.73
Depressive Symptoms	−.02	−.22	−.04	−.46
Childhood SES	−.37**	−4.32	−.45**	−5.68
Risk in Family Environment	.32**	3.98	.32**	4.34
Childhood SES x Risk in Family Environment	−.20*	−2.49	−.24**	−3.24
R 2	.34		.44

Notes:

^*p < .05

^**p <.001. Standardized estimates are displayed. Sex is coded as 1 = Male and 2 = Female.

SES= Socioeconomic Status; ABP= Ambulatory Blood Pressure.

Figure 2.

Average ambulatory systolic and diastolic blood pressure as a function of reported childhood SES and psychosocial risk in childhood family environments. High values for each variable represent 1 SD above the mean, and low values for each variable represent 1 SD below the mean. Results from linear regression analyses controlling for age, sex, subjective SES, and depressive symptoms. Note: SD= standard deviation; SES= socioeconomic status; SBP= Systolic Blood Pressure, DBP= Diastolic Blood Pressure.

As with our sleep analyses, we used The Johnson-Neyman technique⁵⁹ to identify at what value on the Risky Family Questionnaire (range 13-65) the relationship between childhood SES and our ABP outcomes became statistically significant. These analyses revealed that the relationship between childhood SES and ambulatory systolic blood pressure and the relationship between childhood SES and ambulatory diastolic blood pressure was statistically significant for college students with scores on the Risky Family Questionnaire above 23 and 20 respectively.

Discussion

The findings from this research provide additional evidence of relationships between early life exposures (i.e. socioeconomic status and family environments) and markers of CVD risk in college students. Replicating results seen in previous work,³⁷ the association between childhood SES and subjective sleep quality was moderated by the psychosocial risk in childhood family environments. We extended this work methodologically, with our findings indicating that this interaction similarly informs actigraphy derived measures of sleep including WASO and SE. Our analyses indicated that the relationship between childhood SES and our sleep outcomes was significant only when college students reported above a certain level of risk in the family environment. These measures of sleep are not subject to self-report errors or biases which could be implicated with subjective measures of sleep.^43,64 It is interesting to note that these early life environments did not associate with differences in total sleep time for college students. Instead, these dimensions of an individual’s early life experience appear to shape the quality of the sleep, including one’s subjective assessment of sleep quality, sleep disruptions (WASO) and SE. These outcomes could predict worse health outcomes in the future. For example, previous work found that greater WASO at the onset of a longitudinal study predicted worsening depressive symptoms at a 5-year follow-up.⁶⁵

In addition, we found a similar pattern of results with the interaction of childhood SES and risk in childhood family environments as a predictor of average ABP over a two-day period. Specifically, there was a statistically significant relationship between childhood SES and average systolic and diastolic ABP only when college students reported above a certain level of risk in their childhood family environment.

These findings provide further support for the importance of consideration of multiple dimensions of early life environments in shaping health-relevant outcomes in college students. More specifically, our findings indicate that it is important to consider multiple dimensions of childhood environments when identifying students who may be at risk for future CVD. In future work, it will be important to consider whether these childhood exposures interact in a similar manner to inform additional behavioral and physiological markers related to CVD risk. The college experience presents a novel environment, with unfamiliar stressors and challenges. While all college students are likely to experience high levels of psychological stress, the college environment may be particularly stressful for individuals who were exposed to high levels of stress during childhood, and this may be reflected in poor sleep quality and higher ambulatory blood pressure as well as other indices of health.

With the current study design, we were able to examine whether differences in levels of psychosocial risk in the childhood family environment may offset the documented risk associated with being raised in a low SES environment. Our cross-sectional findings suggest that childhood family psychosocial risk may moderate the relationship between childhood SES and sleep, an important health behavior which is related to CVD risk, and the relationship between childhood SES and average ABP, another important predictor of future CVD risk. The pattern of findings for both outcomes suggest that the level of psychosocial risk in family environments during childhood may be particularly important for college students from low childhood SES backgrounds. These results extend a growing area of research documenting the enduring negative effects of a risky early family environment on social relationships, hostility, and reactivity in response to conflict,^13-14 and align with previous research indicating that family dynamics can moderate the association between childhood SES and other health-relevant outcomes.⁶⁶

This study has important limitations to note. First, given the cross-sectional design, we are unable to infer causality or directionality. Future work should utilize existing longitudinal data to examine whether childhood SES and psychosocial risk in childhood family environments predict changes in CVD risk prospectively over time. Second, our measure of SES was subjective, asking respondents to consider how the status of their family during childhood compared to other families with regard to parental income, occupation, and education attainment. We utilized this measure based on previous work which suggests that subjective measures of childhood SES are more closely associated with outcomes later in life than objective measures of childhood SES.⁶⁷ Subjective measures of SES may allow for individuals to reflect on more subtle SES indicators when reporting their SES that may not be captured by objective indicators of SES such as education or income. However, it remains possible that an individual’s recollection of their early family environment is influenced by faulty retrospective recall or by current factors including their mood, self-esteem and current relationships with their family members. Future research should consider the role of these factors in explaining the observed relationships.

Our sample was predominantly composed of Non-Hispanic White college students, reflecting the demographics of the region and the university in which the study was conducted. The racial makeup of our sample limits the generalizability of our results to all college students. Future research should seek to examine whether the patterns of these results differ by ethnicity, race or other demographic factors. While investigation of these relationships in college students is important given the high rates of sleep problems in this group, it will also be important to examine these patterns across the lifespan including in adolescence, another period that is characterized by changes in sleep-wake cycles, and into later adulthood. The pattern of findings in this research was specific to the college context, a period characterized by high psychological stress and transition. The findings highlight the cumulative risk associated with exposure to both low SES and risky family environments in childhood. Future work should examine these relationships across different high stress contexts (e.g., caregiving or postpartum) as well as test whether this pattern is apparent in low stress contexts.

Future work should utilize ecological momentary assessment (EMA) to understand how these dimensions of the childhood environment inform daily life patterns of psychological stress and social interactions, and to investigate how these patterns map onto markers of CVD risk, including the ones measured in this research, and extending to other physiological systems including the endocrine and immune system. EMA research could identify targets which could offset the observed relationships. For example, EMA data may indicate that college students raised in low SES and high risk family environments have smaller social networks or may perceive less social support in their daily lives. If this was the case, interventions may be developed to provide opportunities for students to increase the size of their social network or may provide opportunities to access and utilize social support. Such interventions could be implemented through university health centers and student support services. Furthermore, based on these findings, targeted interventions focused on sleep hygiene (i.e. habits that promote good sleep) and education about the relationship between sleep and health may be particularly beneficial for college students who were raised in low SES and high risk family environments.

Sleep is an important factor for both academic success and physical and psychological health in college students.⁶⁸ While college students generally report low sleep quality, our work indicates that early life experiences are important predictors of sleep quality in this population. Compromised sleep quality may be one pathway that contributes to adverse mental and physical health in the future. Separately, if the observed patterns of blood pressure persist into adulthood for students from low childhood SES environments and more risky childhood family environments, these students will be at increased risk for CVD in later adulthood.

Finally, more research is needed to investigate cross-lagged relationships between sleep and ambulatory blood pressure. Cross-lagged analyses could elucidate directionality in the relationship between sleep and ambulatory blood pressure, and this more nuanced understanding would allow more targeted interventions. For example, if these analyses suggest that poor sleep predicts subsequent high ABP, sleep may be a more effective target for interventions. In the current research, ABP was measured for 2 days while sleep was measured for one week. Ideally, for these cross-lagged analyses, sleep and ambulatory blood pressure should be collected concurrently for at least one week, and should include measurement of nocturnal ABP. The knowledge obtained from future work in this area could be used to inform interventions in early life that aim to offset the risk that is traditionally associated with being raised in low SES and high risk environments.

Data Availability Statement:

The data that support the findings of this study are available from the corresponding author upon reasonable request.