The Role of Self-regulation Failures and Self-care in the Link Between Daily Sleep Quality and Blood Glucose Among Adults with Type 1 Diabetes

Abstract

Background

Sleep, a process that restores the body’s ability to self-regulate, may be one important factor affecting self-care behaviors and blood glucose (BG) levels. The link between sleep quality, self-care behaviors, and BG levels may occur by sleep-altering daily self-regulatory failures.

Purpose

This study examined whether the relation between sleep quality and self-care behaviors occurred through self-regulation failures and whether the relation between sleep quality and BG levels occurred through self-regulation failures and self-care behaviors sequentially.

Methods

One hundred and ninety-nine adults with type 1 diabetes (T1D) completed an online questionnaire for 14 days in which they reported sleep quality, self-regulation failures, and self-care behaviors. BG levels were gathered from glucometers. Analyses involved multilevel mediation models and focused on daily within-person and between-person variability of sleep quality.

Results

Better daily sleep quality was associated with higher self-care behaviors at both within-person and between-person levels, and self-regulation failures mediated the association between daily sleep quality and daily self-care behaviors at both within-person and between-person levels. Better daily sleep quality was associated with better BG levels at the within-person level and self-regulation behaviors and self-care behaviors sequentially mediated the association between daily sleep quality and daily BG levels at the within-person level.

Conclusion

This study provides a process account of the importance of daily sleep quality of adults with T1D, as well as one potential mechanism—self-regulation—that may explain the effect of sleep quality on diabetes outcomes.

Type 1 diabetes (T1D) is a chronic autoimmune disorder in which a person’s pancreas stops producing insulin. The goal of daily management is to mimic a functioning pancreas [1, 2] by engaging in a complex daily self-regulation process that requires checking blood glucose (BG) levels, injecting insulin, monitoring diet and exercise, and adjusting insulin based on these activities [3]. Such self-care behaviors are related to numerous self-regulatory factors in late adolescents, including one’s daily sleep quality, as well as one’s ability to regulate one’s cognitions, emotions, and behaviors [4–6]. Sleep in adults with T1D has not been well studied in relation to self-regulatory factors.

Sleep, a process that restores the body’s ability to self-regulate, may be one important factor affecting self-care behaviors [7–9]. It has been well established that sleep is important for glycemic control, because sleep issues may adversely affect insulin sensitivity, disease progression, and development of complications [10, 11]. However, the majority of diabetes-related sleep studies have been conducted with people with type 2 diabetes [12–14], with comparatively few researchers focusing on sleep for adults with T1D [15, 16]. In addition, previous diabetes-related sleep studies predominantly focus on the between-person (interindividual) effects of sleep quality [12–14]. Fewer studies have focused on both within-person (intraindividual) and between-person variability predicting diabetes outcomes [17]. Sleep quality and diabetes outcomes (i.e., self-care behaviors and BG levels) are likely to vary on a daily basis. Therefore, the present study addressed these limitations of previous research by examining within-person and between-person associations of daily sleep quality with daily diabetes outcomes among adults with T1D.

Sleep quality, an individual’s overall perceived satisfaction with sleep [18], is essential for metabolic health [19, 20], yet little attention has been focused on the benefits of sleep as a potential variable in improving self-care and subsequent BG levels in persons with T1D. Unfortunately, persons with T1D experience unique barriers to good sleep quality as a result of the symptoms (e.g., hyperglycemia, hypoglycemia, glucose variability) and management of T1D [21]. Persons with T1D are generally more likely to report sleep issues than individuals without T1D [22]. In one study, 35% of persons with T1D reported poor subjective sleep quality compared with 20% of participants without T1D [11]. Furthermore, persons with T1D face substantially higher risk for obstructive sleep apnea (17.2%) than those without T1D (5.15%) [22]. These sleep issues among persons with T1D may adversely affect the development of complications, and diabetes-related stress experiences [11, 19]. Insulin sensitivity has been shown to diminish after only a single night of partial sleep restriction compared to a night of normal sleep duration [10]. In 2017, the American Diabetes Association Standards of Medical Care in Diabetes established new recommendations to evaluate sleep patterns and sleep duration while evaluating how well persons with T1D adhere to their medication regimens [23] and routine sleep assessment recommendations continues presently [24].

The link between sleep quality and diabetes outcomes such as blood glucose and self-care behaviors may occur by sleep altering daily self-regulation (Turner SL, unpublished data). Self-regulation is conceptualized as the modulation of emotions, behaviors, and cognitions toward a particular goal [25, 26]. In this study, we focused on self-regulation failures, which are defined as failures in self-regulation involving cognitive, behavioral, and emotional control in the context of monitoring blood glucose levels. Poorer sleep is associated with a range of self-regulation failures including memory failures, failures to initiate a behavior and motivate to engage in a behavior [27]. Persons with T1D must manage competing demands on their time and resources to complete diabetes management tasks [1, 8]. Poor sleep quality may disrupt one’s ability to regulate emotions, cognitions, and behaviors. Self-regulation failures may further disrupt daily diabetes-related self-care behaviors, which are themselves self-regulatory in nature [26]. The relation between sleep and self-regulation failures is especially important for persons with T1D because self-regulation failures are important in diabetes-related self-care behaviors such as remembering to check blood glucose levels, initiating behaviors to address hypo- and hyperglycemia, such as injecting insulin and monitoring diet and exercise [3]. Thus, poor sleep quality may affect daily blood glucose levels through daily self-regulation failures, which impedes the person’s ability to engage in self-care behaviors, leading to poor blood glucose [26].

This study examined potential underlying within-person and between-person mechanisms between sleep quality and diabetes outcomes (self-care behaviors and BG levels) by focusing on the mediating role of self-regulation failures. Persons with T1D completed a 14-day diary that assessed their daily sleep quality, self-regulation failures, self-care behaviors, and BG levels. First, we examined whether sleep quality of persons with T1D affected self-care behaviors in terms of daily variability (within-person effects, or deviation from one’s average across the 14 days) and average effects (between-person effects, averaged across the 14 days [28]) and then explored the potential mediating role of daily self-regulation failures in these relations. We hypothesized that better sleep quality would be associated with higher self-care behavior at both the daily and average levels. We also hypothesized that better sleep quality would be associated with fewer self-regulation failures at both the daily and average levels and that fewer self-regulation failures would explain the link of better sleep quality to better self-care. Then, we examined whether better sleep quality was associated with BG levels at both the daily and average levels and explored the potential, sequential mediating roles of self-regulation failures and self-care behaviors on the relation between daily sleep quality and daily BG levels at both the daily and average levels. We hypothesized that better sleep quality would be associated with healthier/better BG levels and that this association would be explained by lower self-regulation failures and better diabetes self-care behavior at both the daily and average levels.

Method

Participants

All study procedures were approved by each of the two university’s institutional review boards (At the Utah site: IRB_00076031, at the Pittsburgh site: IRBSTUDY2015_00000440). All the participants provided written informed consent. Persons with T1D were recruited from two university-affiliated endocrinology clinics in Pittsburgh, PA and Salt Lake City, UT. Individuals were eligible if they had a diagnosis of type 1 diabetes for at least 1 year, were taking insulin for type 1 diabetes within 1 year of diagnosis, spoke English as their primary language, and were married or in a cohabitating relationship for at least 1 year. At the Utah site, of the 319 persons with T1D approached, 66 were ineligible, and 118 declined to participate. Of the remaining 135 persons with T1D, 107 were scheduled and included in the study. At the Pittsburgh site, of the 202 contacted by the project director, 47 were ineligible (including 2 found ineligible after they started study procedures), 57 declined participation, and 6 could not be reached, resulting in 92 persons with T1D included in the study. The final sample included 199 persons with T1D who were eligible, enrolled, and completed study measures across both sites. Demographic data for persons with T1D are shown in Table 1.

Table 1.

Demographic Information for Persons with T1D

	Persons with T1D
	M (SD)	Range
Age	46.82 (13.95)	25.85–74.89
Gender (% Women)	52.3%	--
Race (%White)	92.5%	--
Ethnicity (% Hispanic)	6%	--
Daily Sleep Quality	3.17 (0.60)	1.75–4.93
Pittsburgh Sleep Quality Index	3.60 (2.55)	0–12
Length of Diagnosis	26.97 (13.88)	3.10–60.63
Pump Use (%)	68.7%	--
CGM Use (%)	43.4%	--
HbA1c	7.57 (1.06)	4.9–11.20

Procedure

At the University of Utah, trained recruiters approached individuals who were diagnosed with T1D and were at least 25 years of age in clinic. Interested participants were provided information about the study (verbally and with a brochure) and were asked to provide their preferred contact information to discuss eligibility and enrollment. At the University of Pittsburgh, persons with T1D were approached by their regular diabetes care provider and were asked for permission to release their name and contact information to the project coordinator. The project coordinator then contacted persons with T1D who agreed to be contacted and provided them with more information about the study, assessed interest in the study, and discussed eligibility and enrollment. If individuals were eligible and interested in the study, the study team then obtained permission to independently contact their romantic partner for eligibility screening. If both members of a couple met study criteria and agreed to participate, the couple was enrolled in the study and scheduled for a laboratory visit.

Prior to attending the in-person laboratory session, persons with T1D provided their informed consent for completing a brief at-home questionnaire online. At the laboratory visit, persons with T1D provided informed consent for completing all other study procedures. During the laboratory visit, participants completed cognitive assessments, an interview, and a couples’ interaction task that were part of the larger study on couples coping with T1D. Following the laboratory visit, participants completed a short online questionnaire every evening before going to bed for 14 days. The present study used only persons with T1D’s daily diary portion of the study because the focus was on persons with T1D’s daily diabetes outcomes. On average, persons with T1D completed 13.82 days of the diary.

Persons with T1D were compensated individually for their participation and for mileage for traveling to the laboratory. Persons with T1D were compensated up to $225 for completing all of the parts of the study ($100 for the initial survey and lab assessment, $7.14 per diary day completed up to $100, and $25 for returning a study-owned glucometer in a prepaid/preaddressed envelope).

Measures

Daily diary measure

Daily sleep quality

To capture daily sleep quality, participants rated the prior night’s sleep using the following item: “How satisfied/dissatisfied were you with your sleep last night?” on a 1 (not at all satisfied) to 5 (completely satisfied) scale, used by Williams et al. (2013) to capture daily sleep quality [29]. This item was modeled after a similar item on the Pittsburgh Sleep Quality Index (a longer, validated global self-report of sleep quality [30]), with the wording changed to reflect the daily level. In the present study, our aggregate daily sleep quality item from the diary correlated with global scores on the Pittsburgh Sleep Quality Index, where lower scores denote better sleep quality (r = −.54, p < .001 [30]). On average across the days of the diary, persons with T1D reported they were somewhat satisfied with their sleep (M = 3.17, SD = .98).

Daily self-regulation failures

At the end of each day, persons with T1D reported on their experience of eight failures in self-regulation involving cognitive, behavioral, and emotional control in the context of monitoring BG, a crucial and difficult daily behavior [3, 31], using a 1 (strongly disagree) to 5 (strongly agree) scale. Higher scores represent a higher level of self-regulation failures. These items were developed by Berg et al. [8] to identify key components of regulation failures involving planning, initiation, memory, and emotional control: (a) I kept putting off my BG testing, (b) I had a lot going on and had a hard time figuring out the best time or place to do my BG tests, (c) I kept meaning to test my BG, but in the end it didn’t quite happen the way it was supposed to, (d) Each time I was about to test my BG, I got distracted by something else, (e) Testing my blood glucose kept slipping my mind, (f) I figured that if I skipped some of my BG testing, it wouldn’t be a big deal, (g) I was so involved in doing something else I was enjoying that I didn’t stop to test my BG when I was supposed to, and (h) I was in a bad mood today and didn’t really care about testing my BG. An average score was used. Interitem consistency reliability of the eight items was calculated via Hierarchical Linear Modeling (HLM) random intercept models, with both time and item treated as nested levels, and was excellent (λ _00. =.96).

Daily self-care behaviors.

To measure diabetes self-care behaviors, at the end of each day, persons with T1D rated their self-care behaviors using six items from the short modified Self Care Inventory (SCI) [8]. Persons with T1D rated how well they followed recommendations in the past 24 hr on a 1 (did not do it) to 5 (did it exactly as recommended) scale: checking blood glucose with meter, administering insulin dose as recommended, adjusting insulin based on blood glucose values, having quick-acting sugar to treat reactions, eating the proper foods or counting all carbohydrates, and using a pump (e.g., programming the pump, making sure there is enough insulin) or continuous monitor (e.g., wearing the sensor) correctly for those using pumps or continuous glucose monitor (CGM). In the present study, an average score was used. Interitem consistency reliability of the six items was calculated via Hierarchical Linear Modeling (HLM) random intercept models, with both time and item treated as nested levels, and was excellent (λ ₀₀ = .97)

Daily blood glucose.

Persons with T1D were given OneTouch Verio IQ glucometers to use in place of their regular meter for the 2-week duration of the daily diary portion of the assessment. In this study, we used both BG mean and average daily risk range (ADRR), which was calculated using McCall and Kovatchev’s [32] average risk for high and low BG Index. The reason we included both metrics is that although BG mean has been widely used, it has limitations. That is, the scaling of BG is neither linear nor equivalent at the low and high ends of the scale with a greater range of high BG values (180–600+ mg/dl) than low (∼20–70 mg/dl)). The ADRR is calculated using at least 14 days within a month of self-monitoring blood glucose (at least three readings/day) and is designed to be equally sensitive to hypoglycemic and hyperglycemic BG deviations to optimize to predict glycemic extremes.

Analytic Strategy

The present study utilized multilevel mediation models [33], which extend the classic mediation model to clustered data, in Mplus (version 8) via the “type is two level” specification [34]. In this study, clustered data involved repeated measures collected from the same individuals over 14 days. We employed a two-level multilevel model to address the study’s central questions. Level 1 represented variability due to within-person repeated measures for persons with T1D, and Level 2 represented between-person variability across persons with T1D [28]. At each level, we specify a mediation model allowing for distinct mediation at Level 1 and Level 2, a mediation model of within person variation and a mediation model of between people variation. This is equivalent to a mediation model where sleep quality, self-regulation failures, and self-care behaviors were person centered and the between person means (removed due to centering) are also included as a mediation channel. First, we used single mediator multilevel mediation models to examine whether the daily associations between sleep quality was associated with self-care behaviors and whether this association was mediated through self-regulation failures at both within-person and between-person levels. Second, we used sequential multiple mediator multilevel mediation models to examine whether self-regulation failures and self-care behaviors mediated daily relationships between sleep quality and BG levels sequentially at both within-person and between-person levels.

Results

First, we ran unconditional models to calculate intraclass correlation coefficients (ICCs) and examined within- and between-person variability in daily sleep quality, self-regulation failures, self-care behaviors, BG mean, and ADDR across the 14 diary days. There was both within-person (67.57%) and between-person variability (32.43%) in daily sleep quality. There was also both within- and between-person variability in measures of daily self-regulation failures (37.61% within; 62.39% between), daily self-care behaviors (31.87% within; 68.13% between), daily BG mean (69.39% within; 30.61% between), and daily ADRR (69.22% within; 30.78% between).

Link Between Daily Sleep and Self-care Behaviors and Mediating Role of Self-regulation Failures

We examined the relations between daily sleep quality and daily self-care behaviors, and the potential mediating role of daily self-regulation failures with both the within-person effect and between-person effect simultaneously (see Fig. 1 for full model results). Analyses revealed a within-person effect of persons with T1D’s daily sleep quality on their daily self-care behaviors, indicating that on days when persons with T1D reported having better daily sleep quality the prior evening than their average, they reported higher daily self-care behaviors on that day (B= .02, SE = .01, p = .01, confidence interval [CI][.01; .04]). Furthermore, a significant within-person effect of persons with T1D’s daily sleep quality on daily self-regulation failures was found, such that on days when persons with T1D reported having better daily sleep quality the prior evening than their average, they reported lower daily self-regulation failures on that day (B= −.03, SE = .01, p = .008, CI [−.06; −.01]). Also, analyses revealed a significant within-person effect of persons with T1D’s daily self-regulation failures on daily self-care behaviors, such that on days when persons with T1D reported higher daily self-regulation failures than their average, they reported lower daily self-care behaviors (B = −.29, SE = .04, p = < .001, CI [−.38; −.20]). In addition, the indirect within-person effect was significant, such that daily self-regulation failures partially mediated the association between daily sleep quality the prior evening and daily self-care behaviors on that day. That is, on days when persons with T1D reported better daily sleep quality than their average, they reported lower daily self-regulation failures, leading to an increase in daily self-care behaviors (B = .01, SE = .004, p = .02, CI [.002; .02]; see Fig. 1 for within-person effect results).

Fig. 1.

Multilevel mediation model to examine the relationship between sleep quality and self-care behaviors and self-regulation failures as a mediator. Within-person effect and between-person effect were examined simultaneously. Figure reports unstandardized coefficients with standard errors in parentheses and the 95% confidence intervals (CI) in brackets. *p < .05, ^**p < .01, ^***p < .001.

Analyses of between person effects were identical to the within-person effects noted above: (a) persons with T1D who reported better sleep quality than the average persons with T1D reported better self-care behaviors (B= .16, SE = .05, p = .001, CI [.07; .26]), (b) persons with T1D who reported better sleep quality than the average persons with T1D reported lower self-regulation failures (B = −.21, SE = .08, p = .006, CI [−.36; −.06]), (c) persons with T1D who reported higher self-regulation failures on average reported lower self-care behaviors on average (B = −.55, SE = .06, p < .001, CI [−.66; −.43]), and (d) the indirect between-person effect was significant such that persons with T1D with better sleep quality on average reported lower self-regulation failures on average, leading to better self-care behaviors on average (B = .12, SE =.04, p = .01, CI [.03; .20]; see Fig. 1 for between-person effect results.)

Link Between Daily Sleep and BG Levels Mediated through Self-regulation Failures and Self-care Behaviors Sequentially

We examined the relation between daily sleep quality and daily BG mean and the sequential mediating roles of daily self-regulation failures and self-care behaviors for within-person and between-person effects simultaneously (see Fig. 2 for full model results). Analyses revealed a within-person effect of persons with T1D’s daily sleep quality on their daily BG mean, indicating that on days when persons with T1D reported having better daily sleep quality the prior evening than their average, they had lower daily BG mean that day (B = −4.04, SE = 1.35, p = .003, CI [−6.68; −1.40]). Furthermore, a significant within-person effect of persons with T1D’s daily sleep quality on daily self-regulation failures was found such that on days when persons with T1D reported having better daily sleep quality the prior evening than their average, they reported lower daily self-regulation failures (B = −.03, SE = .01, p = .008, CI [−.06; −.01]). Also, analyses revealed a significant within-person effect of persons with T1D’s daily self-regulation failures on daily self-care behaviors, such that on days when persons with T1D reported higher daily self-regulation failures than their average, they reported lower self-care behaviors (B = −.29, SE = .04, p = < .001, CI [−.38; −.20]). Furthermore, analyses revealed a significant within-person effects of persons with T1D’s daily self-care behaviors on daily BG levels, such that on days when persons with T1D reported higher daily self-care behaviors than their average, they had healthier BG levels (B = −20.01, SE = 3.89, p < .001, CI [−27.63; −12.40]). We examined the sequential multiple mediating roles of self-regulation failures and self-care behaviors between sleep quality and BG mean for within-person effect. Analyses revealed significant indirect within-person effects. A chain of daily self-regulation failures and daily self-care behaviors partially mediated the association between daily sleep quality and daily BG mean. In other words, on days persons with T1D reported better sleep quality than their average, they reported lower self-regulation failures and then higher daily self-care behaviors leading to a decrease in BG mean (B = −.18, SE = .08, p = .02, CI [−.34; −.03]; see Fig. 2 for within-person effect results).

Fig. 2.

Multilevel mediation model to examine the relationship between sleep quality and BG mean and self-regulation failures and self-care behaviors as sequential multiple mediators. Within-person effect and between-person effect were examined simultaneously. Figure reports unstandardized coefficients with standard errors in parentheses and the 95% confidence intervals (CI) in brackets. *p < .05, ^**p < .01, ^***p < .001

Analyses revealed that there was no between-person effect of persons with T1D’s sleep quality on their BG mean (B = −4.64, SE = 4.83, p = .19, CI [−14.10; 4.82]) and no significant between-person indirect effect (B = −1.28, SE = .86, p = .14, CI [−2.97; .41]; see Fig. 2 for between-person effect results).

Additional analyses were conducted to understand how sleep quality was associated with another metric of BG levels (i.e., ADRR) and underlying mechanisms of this association. We found similar findings to those examined between sleep quality and BG mean. We addressed whether models were consistent across treatment regimen (CGM or daily injections) as the regulatory demands experienced by those on CGM could be lessened. Further as length of diagnosis frequently affects blood glucose, we controlled for this variable assessed via persons with T1D’s report. With these covariates included, we replicated the results.

Discussion

To our knowledge, this is the first study to examine within-person and between-person associations between daily sleep quality and diabetes outcomes and possible mediating variables among adults with T1D. In the context of a daily diary study, we found better daily sleep quality was linked to better self-care behaviors, and fewer self-regulation failures was one mechanism that explained this relation. We also found evidence that better sleep quality was linked to lower BG levels by enabling persons with T1D to experience fewer self-regulation failures and engage in more self-care behaviors on a daily basis. These effects were found within subjects, meaning that we were able to explain why sleep quality on a given day was linked to better BG levels on that day within persons.

Many of the previous studies on the effect of sleep quality on diabetes outcomes have been conducted in individuals with type 2 diabetes. Among individuals with type 2 diabetes, the number of sleep awakenings using objective measures of sleep (Actigraph-wGT3X) predicts self-care along with diabetes distress, fatigue and daytime sleepiness [35]. Similar to our findings, Chasens et al. [36] found that impaired sleep quality was associated with decreased self-management behaviors and negative emotional symptoms. Poor sleep quality may lead to late-night eating habits and encourage the use of caffeine, a chemical known to increase blood glucose levels and duration of elevated blood glucose levels in adults with type 1 and type 2 diabetes [37]. A meta-analysis of individuals with type 2 diabetes also identified a relation between poor sleep quality and higher HbA1C [14]. Our findings suggest that the type 2 diabetes literature related to sleep quality, self-care, and glucose levels applies to T1D.

A significant contribution of the present study was to explore potential mediators of the link between poor sleep quality and blood glucose levels. The results are consistent with the interpretation that daily poor sleep quality impairs self-regulation so that failures impede self-care behaviors and lead to higher blood glucose values. The self-regulation failures measured in the present study represent a range of failures in planning, initiation of behavior, memory, and control over emotions with regard to BG checking [8]. Such failures are key in understanding the dynamics of a broad range of aspects of an individual’s ability to regulate the self in relation to diabetes over time (Turner SL, unpublished data). The results are consistent with other research suggesting that greater sleep is associated with increases in a number of facets of self-regulation, including self-monitoring as well as goal setting, which would benefit self-care behaviors [38].

The mediating roles of self-regulation failures and self-care behaviors on the relation between daily sleep quality and BG levels were only found at the within-person level, as opposed to the between-person level. The analyses performed in the present study, controlling for both between- and within-person level measures of each construct, suggest that it is daily fluctuations in these variables from one’s own average rather than between-person effects that are important for understanding these processes. The ICCs for both sleep and BG mean revealed that the majority of the variance existed within-persons rather than between persons, which could account for the results found here. Such results provide caution in interpreting studies that rely on individual difference measures of these constructs as reflective of daily processes as well. Future research is needed, however, to examine the time course of these processes at the daily level assessing sleep quality after rising and self-regulation failures, self-care behaviors, and BG levels throughout the day to capture how self-regulation failures may disrupt self-care behaviors and BG sequentially by utilizing ecological momentary assessment.

The results of the current study must be interpreted in the context of limitations. First, data were mainly self-report with the exception of blood glucose levels provided by glucometer. Future research would benefit from objective measures of sleep. For example, participants can wear an actigraph device, which objectively monitors sleep over the study period to give additional insights into how other metrics of sleep, such as sleep duration, sleep disturbances, and obstructive sleep apnea [39] relate to diabetes outcomes [40]. Actigraphy monitors are noninvasive, portable, and easy to wear for extended periods of time and are uniquely suited to capture night-to-night variability in sleep in persons with T1D simultaneously in an ecologically valid context (i.e., the persons with T1D’ homes) [41, 42]. In addition, participants completed the daily diary in the evening about the previous night’s sleep quality and that day’ self-regulation failures, self-care behaviors. Thus, the measurement of daily sleep quality may be more affected by retrospective bias than those of self-regulation failures and/or self-care behaviors. Furthermore, we examined self-regulation failures in the context of BG monitoring. Future research is needed to examine daily self-regulation failures in the context of other behaviors, such as diet and exercise. The fact that the results held after controlling for factors such as CGM use, which may reduce the self-regulatory demands associated with T1D may indicate that our measure of self-regulation failures is capturing self-regulation beyond behaviors associated with monitoring BG. Consistent with this idea, our measure of self-regulation failures has been related with broader measures of executive functioning in our work with adolescents [8].

These findings hold some useful clinical implications. Such process-oriented investigations between daily sleep quality and daily diabetes outcomes hold the potential to identify targets for prevention and intervention efforts to optimize daily sleep quality among persons with T1D. Clinicians should assess sleep quality and duration and provide counsel about sleep hygiene, strategies to manage dietary self-care when tired, and address insulin adjustments and/or behaviors that may help to minimize overnight hypo- and hyperglycemia. In particular, among persons with T1D who have sleep disorders, daily sleep quality can be monitored by clinicians or physicians using telemedicine, which is the delivery of health care services via virtual visits without an in-person visit. That is, clinicians or physicians and persons with T1D can share information about sleep quality in real time by telephone or via the Internet and can potentially help persons with T1D adhere to recommended treatment in an efficient and timely manner [43]. Improvements in daily sleep quality may provide important reductions in self-regulatory failures, better self-care, and lower BG levels [8].

In conclusion, this study provides a window into the importance of daily sleep quality of adults with T1D, as well as an illustration of self-regulation as an important underlying mechanism linking sleep quality to diabetes outcomes. More conclusive study is needed while addressing the limitations of our study. This study may encourage future research to identify daily antecedent factors that affect daily sleep quality of persons with T1D. This study also may provide a foundation for future longitudinal research on persons with T1D to examine the underlying physiological mechanisms of these associations as well.

Compliance with Ethical Standards

Conflict of Interest The authors declare that they have no conflict of interest.

Authors’ Contributions E.L.T. and C.A.B. developed the study concept and drafted the manuscript, in collaboration with all other co-authors. E.L.T. analyzed the data with analytic and methodological assistance from J.B. All authors provided critical revisions to the document, and then approved the final version of the manuscript for submission.

Ethical Approval All study procedures were approved by each of the two university’s institutional review boards (At the Utah site: IRB_00076031, at the Pittsburgh site: IRBSTUDY2015_00000440). All the participants provided written informed consent.