The importance of having fun: Daily play among adults with type 1 diabetes

Abstract

Although play has been linked to psychological and physical well-being in childhood, little work has examined benefits of play in adulthood. Play may be especially important in the context of coping with a chronic illness such as type 1 diabetes as self-care involves a difficult daily process. We hypothesized that daily play with one’s romantic partner would be linked to better mood, greater diabetes disclosure, greater support receipt, greater perceived coping effectiveness with one’s most important daily stressors, and better self-care regarding type 1 diabetes. We examined these hypotheses in a sample of 199 adults with type 1 diabetes. Participants completed daily diary measures of play with their partner and proposed outcomes for 14 days. Daily play was linked to better mood, greater diabetes disclosure to one’s partner, greater support received from one’s partner, and greater perceived coping effectiveness with the day’s most important diabetes and general stressors. However, daily play was unrelated to self-care. Mediation analyses further indicated that positive mood explained links from daily play to perceived coping effectiveness, and diabetes disclosure explained links from daily play to support. These findings indicate that having fun with one’s partner may have important psychological and relationship benefits for individuals with chronic illness. More work needs to be conducted to examine links from daily play to self-care. Implications are discussed.

Never, ever underestimate the importance of having fun.

—Pausch (2008)

Play is thought to be an intensely positive experience and an important aspect of a happy and healthy childhood. Empirical evidence has supported this claim, finding that play is associated with psychological and physical well-being in children and young animals (e.g., Hurwitz, 2003; Pellegrini & Smith, 1998; Power, 2000). However, few researchers have empirically examined play in adulthood (see Van Vleet & Feeney, 2015a for a review). Recent theoretical work has drawn attention to this neglected topic and advanced this research area by providing a definition of play behavior and outlining several important benefits likely to be accrued by play in adulthood (Van Vleet & Feeney, 2015b). The current work aims to build on this theoretical foundation by testing links from play activities shared with a relationship partner to emotional, relationship, and self-care outcomes among a sample of adults with type 1 diabetes—a group of people who must perform a number of difficult daily tasks as part of their self-care regimen. We expect that play will help reduce the emotional and cognitive toll of daily diabetes management and bond patients to partners, which may increase partners’ impact on patients’ well-being and self-care. Thus, play will have emotional, relationship, and, ultimately, diabetes management benefits for adults with type 1 diabetes.

Although there is debate about the definition of play, in the present investigation, we adopt a definition derived from theoretical work on the topic of play in adulthood that takes into consideration commonalities across the many proposed definitions. Specifically, we define play as an activity performed with a goal of amusement/fun; an enthusiastic, in-the-moment attitude; and a high degree of interaction either with the activity, itself, or with interaction partners (Van Vleet & Feeney, 2015a, 2015b). Defined as such, it is important to note that play can take many forms (e.g., cooperative, competitive, novel, familiar, planned, spontaneous). Illustrating this heterogeneity, when a sample of couples 65 years of age and older were provided with the definition of play above and asked to provide examples of activities they do with their partner that fit this definition, responses varied widely (e.g., “Playing cards and board games,” “Going to Bingo,” “Competing against each other and the contestants while watching Jeopardy and Wheel of Fortune,” “Playful bantering”; Van Vleet & Feeney, under review).

Theoretically, it is expected that play behavior in adulthood will accrue important benefits, just as play behavior in young mammals has been linked with important outcomes. First, play is likely to generate high levels of positive affect because it is an intensely positive experience (Van Vleet & Feeney, 2015b). Second, play should provide a momentary reprieve from one’s everyday stressors and responsibilities (Lauer & Lauer, 2002; Orthner, 1975, 1976; Van Vleet & Feeney, 2015b), thereby reducing stress. Third, play is thought to serve a bonding function between partners (Betcher, 1981; Johnson, 2003; Van Vleet & Feeney, 2015b), and so partners may feel more comfortable opening up to each other after participating in a shared play activity. Play also provides a safe context (free from negativity and seriousness) in which partners can relate to each other, express themselves, and become more adept at reading each other’s nonverbal and verbal expressions (Betcher, 1981). This may translate into a greater ability to discern cues and interpret expressions in more serious contexts, such as support interactions.

Finally, play should improve one’s problem-solving abilities, by enhancing one’s ability to think in an unconventional manner (Betcher, 1981; Lauer & Lauer, 2002). The broaden-and-build theory of positive emotions (Fredrickson, 1998, 2001, 2003) asserts that positive emotions, which are likely to be generated by play, elicit diverse thought-action repertoires that are advantageous for problem-solving. Relatedly, experimental work has found that positive emotional states predict better performance on a problem-solving task than negative or neutral emotional states (Isen, Daubman, & Nowicki, 1987; Isen, Johnson, Mertz, & Robinson, 1985).

Although theoretical work posits that play will elicit these benefits in healthy individuals, no work to date has examined such links in a sample of individuals coping with a chronic illness. Additionally, the chronic illness literature often focuses on understanding how illness-specific interactions with relationship partners influence health and well-being, while more positive interactions with partners that are not concentrated on the patient’s illness have been neglected. We hypothesize that the theorized benefits of play will extent to a sample of individuals coping with type 1 diabetes. Moreover, we expect that the theorized benefits of play will translate into positive diabetes outcomes, as the proposed benefits of play map on well to the resources needed to effectively cope with type 1 diabetes on a daily basis.

Coping with type 1 diabetes is a daily challenge that requires substantial effort and resources and can be cognitively and emotionally taxing. Individuals must regularly check blood glucose levels, administer the appropriate dosage of insulin based on food intake and exercise expenditure, maintain a healthy diet, regularly exercise, regulate one’s emotions, maintain attention to diabetes needs when other issues arise, and cope with daily stress related to diabetes management. All of these self-care responsibilities are also interconnected, with lapses in one aspect of self-care having downstream effects on other aspects of self-care.

Play may help individuals with type 1 diabetes better regulate their emotions when obstacles to self-care occur by equipping individuals with resources that are necessary for effective diabetes management. First, play should provide a boost in positive mood and a decline in negative mood. As mentioned earlier, positive affect has been linked to more effective problem-solving, which is an integral component of type 1 diabetes management, especially when obstacles to self-care arise. Thus, play may generate positive affect, which in turn leads to more effective problem-solving and self-care. Play may also lead to more effective diabetes problem-solving because it provides a venue for more unconscious processing of diabetes self-care information. Unconscious thought theory (e.g., Dijksterhuis & Nordgren, 2006) suggests that unconscious thought is more adaptive for making complex decisions, while conscious thought is more effective when making simple decisions. For instance, individuals faced with complex decisions performed better if they engaged in a distractor activity (presumably because this task offered an opportunity for unconscious thought) than if they spent the same amount of time consciously thinking about the information relevant to their decision. Given that self-care decisions often involve complex information, play may provide an opportunity for patients to unconsciously process such information and arrive at optimal self-care decisions.

Second, because play is thought to bond relationship partners and to elicit communication between partners, play may elicit greater diabetes disclosure to the partner. This more open illness-related communication may lead individuals to use their partner as a resource in their diabetes management. This may be particularly beneficial for individuals who have type 1 diabetes because there is a growing literature that indicates that collaborating with a close relationship partner in managing one’s illness is associated with a host of emotional, relationship, and health benefits (e.g., Afifi, Hutchinson, & Krouse, 2006; Badr & Acitelli, 2017; Helgeson, Jakubiak, Seltman, Hausmann, & Korytkowski, 2017; Van Vleet, Helgeson, Seltman, Korytkowski, & Hausmann, 2018a, 2018b; Zajdel, Helgeson, Seltman, Korytkowski, & Hausmann, 2018). As such, we expect that play will lead to greater diabetes disclosure to the partner, which will in turn elicit greater diabetes support from the partner and lead to better self-care. In sum, given that play is theorized to strengthen emotion regulation, improve problem-solving, open communication between partners, and improve diabetes support, play may be an important and unexpected vehicle to improved self-care for individuals with type 1 diabetes.

We expect that the benefits of play will translate into better self-care for individuals with type 1 diabetes. However, it could also be argued that play may distract or detract from one’s self-care responsibilities for a couple of reasons. First, one could imagine that an individual would become so fully absorbed in a play activity that he/she may lose track of time and forget to carry out one of the many aspects of diabetes management. Second, play may be used intentionally as a way to avoid one’s diabetes and general problems. If this were the case, play may provide boosts in mood, but have no effects on diabetes outcomes. This is the first work to examine the extent to which play is associated with benefits or detriments to self-care.

The goals of this work were to investigate links from individuals’ daily play to several important theorized outcomes in adults who have type 1 diabetes. Initial work on play behavior in adulthood has established links from play to psychological and relational well-being in healthy adults (Vanderbleek, Robinson, Casado-Kehoe, & Young, 2011; Van Vleet & Feeney, under review), but no work to date has examined potential benefits of play in a chronically ill population. In addition, research has not examined how play is linked to outcomes on a daily basis. Given the in-the-moment approach inherent to play, daily diary assessments of play may be particularly well-suited for capturing links from play to theorized outcomes than more retrospective assessments.

Our hypotheses were as follows. First, daily play will be linked to better mood— greater daily positive mood and lower daily negative mood. Second, daily play will be linked to greater disclosure about diabetes to the partner. Third, play will be linked to greater diabetes support received from the partner. Fourth, as a proxy of diabetes problem-solving, we expected that daily play will be related to feeling that one handled the day’s stressors effectively. Finally, for these same reasons, we hypothesized that daily play will be related to better self-care. In testing these hypotheses, we examined links between daily play to proposed outcomes both concurrently and longitudinally. We also examined whether positive affect and diabetes disclosure explained the relations of daily play to outcomes.

Method

Participants

Participants were 199 patients with diabetes, recruited from two different sites (Pittsburgh and Salt Lake City). The majority of the sample was married (92%) and the remainder were cohabiting, with the vast majority (97%) in heterosexual relationships. Demographic information for the sample is displayed in Table 1. The mean glycemic levels for patients were above the American Diabetes Association’s recommended guideline of 7.0% for adults (M = 7.57, SD = 1.06).

Table 1.

Demographic information.

	M (SD)	Range	%
Age	46.81 (13.95)	25–74	—
Gendera (% women)	—	—	52
Race (% NH White)	—	—	90
Ethnicity (% Hispanic)	—	—	6
Education (<associate’s degree)	—	—	28
Relationship lengthb	19.36 (14.56)	1.00–52.00	—
Length of diagnosis	26.97 (13.88)	3.10–60.63	—
Pump use (%)	—	—	69
CGM use (%)	—	—	43

Note. CGM = continuous blood glucose monitor; NH = non-Hispanic.

^aGender coded “0” (men) and “1” (women). No patient selected a nonbinary gender (“Other”), so this variable was dichotomized.

^bRelationship length denotes amount of time patient reported living with current partner in years (whether married or not).

Procedure

Patients were recruited from university-affiliated endocrinology clinics in Pittsburgh, Pennsylvania and Salt Lake City, Utah. Patients were eligible if they were 25 years of age or older, were diagnosed with type 1 diabetes for at least 1 year, were currently on an insulin regimen for type 1 diabetes, had been placed on an insulin regimen within a year of their diagnosis, spoke English as their primary language, and were either married or living with their romantic partner in a cohabiting relationship for at least 1 year.

At the Pittsburgh site, patients were first approached by their diabetes care provider and asked their permission to release their contact information to the project director. If patients agreed, the study director later called the patient to provide study information, screen him/her for eligibility, and obtain contact information for the romantic partner (who was later contacted and screened for eligibility). If both partners were eligible and agreed to participate, the couple was enrolled in the study. Of the 206 patients approached in clinic, 4 declined contact with the study director. Of the 202 contacted by the study director, 47 were ineligible (including 2 who were found to be ineligible after starting study procedures), 57 declined, and 6 could not be reached to determine eligibility. Thus, 92 couples were enrolled at the Pittsburgh site.

The recruitment procedure was similar at the Salt Lake City site, with the exception that participants were first approached by the study research assistants in clinic, who provided study information and screened patients for eligibility. All other recruitment procedures were the same as the Pittsburgh site. Of the 319 patients approached and screened for eligibility, 66 were deemed ineligible and 118 declined to participate. Of the remaining 135 couples, 107 were enrolled and 28 were unable to be contacted following their clinic visit. The final sample across sites included 199 patients and their partners.

As part of a larger investigation, patients completed an online survey and participated in a laboratory visit during which they completed a series of questionnaires and interviews with their partners. Following the laboratory visit, patients completed a 14-day daily diary protocol, which is the focus of the current investigation. During this 14-day period, patients completed an online questionnaire every evening before going to bed. Patients were e-mailed links to their questionnaire each evening. If they had not completed the survey by 9:00 p.m., reminders were sent via patients’ preferred method of e-mail, text, or phone call. A final reminder was e-mailed at 6:00 am the following morning if the previous night’s survey had not been completed. To accommodate those working third shift, access to questionnaires remained open until 10:00 am the following morning. Compliance was excellent for completing the daily diary questionnaires. On average, nearly all of the 14-days’ questionnaires were completed (M = 13.76).

Measures

Daily play.

To measure daily play, patients completed a single item created for study purposes (“Did you do something fun and light-hearted with your partner today?”). This item was designed to reflect the definition of play behavior. Patients responded by selecting either 0 = “no” or 1 = “yes.” The mean was .56 (SD = .50), indicating that participants engaged in shared play on approximately half of the days.

Mood.

To assess daily mood, patients rated their experience of 18 emotions each day (Cranford et al., 2006). Nine items reflected positive mood (e.g., joyful, happy, amused; λ = .95). Reliability was calculated following procedures outlined by Raudenbush and Bryk (2002). Three items reflected depressed mood (sad, discouraged, hopeless), three measured anxious mood (anxious, on edge, uneasy), and three measured angry mood (annoyed, angry, resentful). A composite of the depressed, anxious, and angry mood items was created to reflect negative mood (λ = .91).

Disclosure.

Patients also completed a single diabetes disclosure item (“How much did you tell your partner about things that happened with your diabetes today?”). Patients responded on a 5-point scale from “Nothing” to “A lot.”

Support.

To assess daily support, patients reported the extent to which their partners provided emotional support (2 items; “was there for me by giving me his/her undivided attention,” “tried to understand my situation”) and instrumental support (3 items; “helped me figure out how to take care of my diabetes,” “suggested things that might help me manage my diabetes,” “reminded me of the things I needed to do to manage my diabetes”) on a 5-point scale from 1 = “not at all” to 5 = “a lot.” A principal components analysis of the 5 items included in the emotional support and instrumental support scales followed by varimax rotation indicated the items loaded on a single factor. Thus, a composite was calculated, indicative of diabetes support received from the partner (λ = .96).

Perceived coping effectiveness.

To measure how patients felt they handled their most stressful diabetes event of the day, patients first described their most stressful diabetes-related event and then rated how well they handled it (“How well did you handle this event?”). This same method was used to assess perceived coping effectiveness with the day’s most stressful non-diabetes event. Responses were made on a 5-point scale from “very badly” to “very well”.

Self-care.

Patients reported their daily self-care using a shortened 7-item version of the Self Care Inventory (Berg et al., 2014). Patients rated how well they followed recommendations in the past 24 hr on a scale from 1 (“did not do it”) to 5 (“did it exactly as recommended”) for: 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, for those using a pump or continuous blood glucose monitor (CGM), using my pump (programming the pump, making sure there is enough insulin) or continuous monitor (wearing the sensor) correctly. The average of the 7 items was used as a measure of daily self-care (λ = .97).

Covariates.

Patients also completed measures of demographic variables (age, sex, race, education, income) and method of insulin delivery (CGM, pump, injections), which were examined as potential covariates in analyses. To control for alternative explanations for daily play’s links to outcomes, participants also completed an abbreviated 27-item measure of the big five personality characteristics (openness to experience, conscientiousness, extroversion, agreeableness, and neuroticism) developed for the Midlife Development Inventory (Lachman & Weaver, 1997). Variables that were significantly linked to play were controlled in subsequent analyses.

Overview of analyses

Data were analyzed using multilevel modeling as days (Level 1) were nested within participants (Level 2). First, we examined same-day associations between daily play and outcomes. Following methods outlined by Bolger and Laurenceau (2013), we included both the within-person centered daily play variable (or person-centered) and the between-person centered daily play variable (grand-centered) in all models. Because we expect stronger relations between days that are closer together, we allowed for the autocorrelation of errors for the repeated measure. The unstructured covariance matrix was chosen for random effects. All models allowed for a random intercept and a random effect for the within-person effect of play.

Second, we examined links from daily play to changes in daily outcomes with lagged analyses. These analyses assessed whether play on Day n predicted outcomes on Day n, controlling for the previous day’s outcome (Day n − 1).

To assess the extent to which positive affect and diabetes disclosure mediated links from daily play to support, perceived coping effectiveness and self-care, mediation analyses were conducted in MPlus following the procedures outlined by Bolger and Laurenceau (2013). We followed the guidelines for lower level mediation of daily diary data, because all variables included in the analyses were Level 1 variables. Thus, within-person centered variables were used in mediation analyses (independent, mediator, and dependent variables), as recommended by Bolger and Laurenceau (2013).

Results

Descriptive statistics and zero-order correlations among all study variables are shown in Table 2.

Table 2.

Descriptive statistics and zero-order correlations among study variables.

	Mean	SD	1	2	3	4	5	6	7	8
Daily play	0.55	0.50	—
Positive mood	2.81	0.77	.36***	—
Negative mood	1.49	0.58	− 17***	−.33***	—
Disclosure	2.46	1.31	41***	.22***	− 11***	—
Support	2.23	1.14	.49***	.29***	− 15***	.78***	—
PCE-diabetes	3.44	0.92	.09***	.38***	−.38***	.08***	.06***	—
PCE-general	3.54	0.93	− .03	.30***	−.33***	−.03+	.01	.59***	—
Self-care	4.42	0.67	.05**	14***	− 26***	.06**	−.04*	17***	25***	—

Note. N = 199; Correlations in table reflect aggregates across 14-day daily diary period. SD standard deviation; PCE = perceived coping effectiveness.

⁺p ≤ .10;

^*p ≤ .05;

^**p ≤ .01;

^***p ≤ .001.

Before conducting the multilevel models described earlier, we first partitioned the variance of daily play into its within-person and between-person components to determine whether there was adequate fluctuation in daily shared play to warrant the multilevel analyses. The unconditional model revealed that 83% of the variability in daily play was accounted for by within-person effects, and 17% of the variability in daily play was accounted for by between-person effects. Thus, there were substantial daily fluctuations in play.

Additionally, relations from play to personality variables (openness, conscientiousness, extroversion, agreeableness, and neuroticism) and possible covariates (age, sex, time since diagnosis, life satisfaction, income, education, site¹ (Pittsburgh vs. Salt Lake City)) were examined by conducting generalized mixed models in SPSS, with the personality variables and potential covariates as Level 2 predictors of daily reports of play. Separate models were run for each potential covariate. This procedure indicated that patients who reported more daily play across the 14 days also reported higher openness to experience, lower conscientiousness, marginally lower neuroticism, marginally higher life satisfaction, and were more likely to use CGM for insulin delivery. When we entered all five possible covariates into a single analysis to predict daily play, only three remained significant: openness to experience, conscientiousness, and CGM. Thus, these three covariates were statistically controlled in all subsequent analyses.²

Concurrent associations

Within-person effects.

Results for concurrent links from daily play to outcomes are presented in Table 3. Significant within-person effects indicated that on days that patients reported shared play with their partners, they reported more positive and less negative mood. They also reported receiving more support from their partner and disclosing more about their diabetes to their partner. Similarly, they reported feeling that they handled their most stressful diabetes and non-diabetes event of the day better. However, daily fluctuations in play were unrelated to self-care.

Table 3.

Concurrent multilevel model results.

	Positive mood	Negative mood	Disclosure	Support	PCE-diabetes	PCE-general	Self-care
Intercept	1.89***	1.42***	2.40***	2.62***	2.89***	2.87***	3.57***
CGM	−0.02	0.00	0.10	0.00	0.00	−0.04	0.03
Openness	0.26**	0.07	−0.02	−0.11	0.24**	0.20*	0.01
Conscientiousness	0.05	−0.02	0.05	0.00	−0.04	0.06	0.24**
Day	−0.01***	−0.01***	−0.02***	−0.01***	−0.01*	−0.02***	0.00
WI play	0.12***	−0.07***	0.17***	0.16***	0.06*	0.06*	−0.01
BW play	0.40***	−0.13*	0.74***	0.84***	0.05	−0.03	0.09

Note. N = 199; PCE = perceived coping effectiveness; CGM continuous blood glucose monitor; WI = within-person variability; BW = between-person variability. For self-care, analysis indicated that the final Hessian matrix was not positive definite although convergence criteria had been satisfied. We removed the random slope for within-person variability in play and changed the random effect covariance matrix to identity. When this was done, the model converged and ran without errors. Results presented in the table reflect the model without the random slope for within-person variability in play.

⁺p ≤ .10;

^*p ≤ .05;

^**p ≤ .01;

^***p ≤ .001.

Between-person effects.

Significant between-person play effects indicated that patients who engaged in more play across the 14 days reported more positive and less negative mood, greater support, and greater diabetes disclosure. Between-person variability in daily play was unrelated to perceived coping effectiveness with either stressor or self-care.

Lagged effects

Within-person effects.

Results from lagged multilevel models are presented in Table 4. Similar to concurrent findings, significant within-person variability in play was linked to next-day better mood (higher positive and lower negative mood), greater support, more disclosure, and greater perceived coping effectiveness with their most stressful diabetes and non-diabetes events of the day. No significant within-person play effects were found for self-care.

Table 4.

Lagged multilevel model results.

	Positive mood	Negative mood	Disclosure	Support	PCE-diabetes	PCE-general	Self-care
Intercept	1.11***	1.14***	2.45***	1.66***	2.09***	2.47***	2.40***
CGM	0.01	−0.01	0.13	0.00	0.01	−0.05	0.02
Openness	0.17**	0.04	−0.03	−0.07	0.17*	0.15	0.02
Conscientiousness	0.03	−0.01	0.06	0.04	0.01	−0.01	0.15*
Day	0.00	−0.01**	− 0.02**	−0.01+	−0.01+	0.01*	0.00
Previous day DV	0.38***	0.21***	−0.02	0.30***	0.25***	0.21***	0.33***
WI play	0.11***	− 0.07***	0.18***	0.16***	0.07*	0.04	−0.01
BW play	0.25***	−0.10*	0.77***	0.62***	0.05	0.02	0.07

Note. N = 199; PCE = perceived coping effectiveness; CGM = continuous blood glucose monitor; DV = dependent variable; WI = within-person variability; BW = between-person variability. For self-care, the the analysis indicated that the final Hessian matrix was not positive definite although convergence criteria had been satisfied. We removed the random slope for within-person variability in play, and changed the random effect covariance matrix to identity. When this was done the models converged and ran without errors. Results presented in the table reflect the model without the random slope for within-person variability in play.

⁺p ≤ .10;

^*p ≤ .05;

^**p ≤ .01;

^***p ≤ .001.

Between-person effects.

Significant between-person play effects were nearly identical. Between-person variability in daily play was linked to more positive and less negative mood, greater support, and greater diabetes disclosure, but was unrelated to perceived coping effectiveness for the patient’s non-diabetes stressor of the day or self-care.

Ancillary analyses.

To address an alternative explanation for our findings, that positive mood predicts daily play, we conducted additional lagged multilevel modeling analyses in which play was the dependent variable, within-person and between-person variability in positive mood were entered as independent variables, and the previous day’s play was controlled. Results indicated that between-person variability in positive affect, but not within-person variability in positive affect, was linked to daily play. This pattern suggests that it is not within-person variability in positive affect that is driving the results of this work.

Mediation

Next, we examined positive mood and diabetes disclosure as potential mediators of the links from daily play to support, perceived coping effectiveness, and self-care. Results indicated that positive affect mediated links from daily play to perceived coping effectiveness for one’s diabetes stressor (β = .06, SE = .01, p < .0001, 95% mediated effect) and general stressor (β = .05, SE = .01, p < .0001, 86% of effect mediated), but did not mediate links from daily play to support or self-care. Diabetes disclosure mediated links from daily play to support (β = .06, SE = .01, p < .0001, 43% of effect mediated), but not perceived coping effectiveness with one’s diabetes or general stressors, or self-care.

Discussion

This is the first study to examine links from daily play to important emotional, relationship, and self-care outcomes among adults with chronic illness, in this case, type 1 diabetes. On days that individuals participated in a play activity with their partners, they reported better mood, they told their partner more about their diabetes issues, they felt their partner was more supportive, and they felt they handled their most pressing diabetes and non-diabetes stressors better than on days they did not play. Daily play, however, was unrelated to self-care. All of these effects, with the exception of perceived coping effectiveness with one’s non-diabetes stressor of the day, remained when controlling for the previous day’s reports of the outcome. Our findings also indicated that positive mood explained links from daily play to perceived coping effectiveness, and that diabetes disclosure explained links from daily play to support. Additionally, individuals who reported greater play across the daily diary period reported more positive mood, greater diabetes disclosure, and greater support receipt than individuals who reported less play. As a whole, these results suggest that daily play is associated with a positive emotional and relational climate and with feeling one is more effective in handling diabetes stressors but not necessarily better self-care.

These findings are consistent with the little existing work on play in adulthood that suggests shared play has important emotional and relationship consequences (Aune & Wong, 2002; Baxter, 1992; Vanderbleek et al., 2011). These patterns replicate findings from work linking play behavior observed in the laboratory to greater relational well-being in a sample of healthy newlyweds and older adults (Van Vleet & Feeney, under review) and extend them through its use of intensive longitudinal methods and examination of potential benefits of play in a chronically ill sample.

Findings were less conclusive for self-care outcomes. Although play predicted improved perceived problem-solving (i.e., handling one’s stressors effectively) and more open diabetes communication between partners—which we viewed as important elements of diabetes management—this did not translate into better self-care. There may be a number of reasons why we did not find this link. First, the paths to self-care may be more indirect than the paths to the other outcomes examined. Therefore, play may take a longer time to exert downstream effects on self-care. In future work, researchers should consider following couples over longer periods of time to assess gradual changes in self-care and diabetes outcomes as a function of play.

However, play was not associated with detriments to self-care. It is important to take note of this because, although we expected play to be beneficial to self-care, it could also be argued that the immersive, in-the-moment attitude inherent to play may distract from one’s diabetes management responsibilities. Our data suggest that this was not the case.

A viable alternative explanation for why play was unrelated to self-care, however, is that patients may intentionally use play as a way to emotionally cope with their diabetes problems, but that play does not directly address the source of such problems. In line with this explanation, play was related to better mood but was unrelated to self-care. Although play was related to greater perceived coping effectiveness with one’s diabetes stressor of the day, our perceived coping effectiveness measure did not specify whether such coping was emotion-focused or problem-focused. Therefore, this alternative explanation cannot be ruled out. It is also possible that benefits and detriments of play behavior canceled each other out when it comes to self-care.

This investigation has several strengths worth noting. This work contributes substantially to the literature on play in adulthood. This is the first work on this topic to employ an intensive longitudinal design and to replicate links from play to emotional and relationship benefits in a chronically ill population. This work also contributes to the diabetes literature by drawing attention to the potential benefits of positive interactions with partners that are not focused on diabetes. Research has consistently found spousal support to be a strong predictor of self-care (Tang, Brown, Funnell, & Anderson, 2008; Whittemore, Melkus, & Gray, 2005), and a growing body of research has linked communal coping (e.g., Helgeson, Jakubiak, Van Vleet, & Zajdel, 2017) to self-care and well-being. Few other positive interactions between couple-members have been examined. Moreover, the positive interactions explored in empirical research tend to be diabetes-focused. However, we suspect there are many different ways that partners can contribute to patients’ well-being and self-care resources that are less direct.

Despite these strengths, readers should consider our findings within the context of the study limitations. Although this study’s daily diary methodology adds rigor to the existing literature on play in adulthood, the data are correlational in nature and thus causation cannot be inferred. A single item dichotomous measure of daily shared play was used and could be improved upon in future work. We urge future researchers to capture descriptive information about daily play activities, the context in which play activities occur, and to obtain participants’ perceptions of such activities. As mentioned earlier, we did not have descriptive information regarding what patients did with their partners that they considered to be “fun and light-hearted.” However, previous work involving older healthy couples found participants reported a great variety of activities that fit this description, including planned activities and events (e.g., going to Bingo, bowling), as well as shorter, spontaneous moments (e.g., banter, teasing). It will also be important for future research to assess both partners’ perceptions of play on a daily basis. We suspect that play will be most beneficial when both partners report they did something fun together that day. Furthermore, the single-item measure of play employed in this work focused on play activities shared with the partner and did not capture independent play activities or play activities that were shared with others besides one’s romantic partner. These are important steps for future work. We expect independent play and play with other relationship partners to also have benefits, but that they will be less impactful on outcomes specific to the spousal relationship.

The research on play in adulthood is in its infancy, and there are many avenues for future researchers to explore. First, causal links from play to outcomes need to be examined by experimentally manipulating play in the laboratory. Another important step for this area is to assess perceptions and outcomes of specific shared play episodes immediately after the activity. This will better allow researchers to examine potential mechanisms underlying links from play to its proposed outcomes. We suspect that the immediate increases in positive affect may be an important pathway through which play has longer term benefits over time, and our findings in the current work support this claim. Future work examining links from play to outcomes in chronically ill populations should also consider outcomes of play for the partner. We suspect that play activities shared together may help relieve stress, strengthen problem-solving abilities, and improve relationship quality for both partners.

In addition to considering general play—which is very heterogeneous in nature—it may also be useful for researchers to examine different forms that play may take. Previous work identified several dimensions on which play may vary (e.g., novelty, familiarity, cooperative, competitive). Specific play forms may differentially predict the theorized benefits of play. For instance, shared play that is more novel in nature may be especially likely to generate relationship excitement and reduce relationship boredom. Work by Aron and colleagues on shared novel and exciting activities support this claim (Aron & Aron, 1986; Aron, Norman, Aron, McKenna, & Heyman, 2000; Reissman, Aron, & Bergen, 1993), as does work by Malouff and colleagues (Malouff et al., 2012; Malouff, Mundy, Galea, & Bothma, 2015). Shared novel play activities may serve not only as a means of sparking excitement in the relationship, but may also serve communicative and relationship maintenance functions.

As play is defined in this work, we expect that play will generally have positive effects. However, it will be important for future work to consider the boundary conditions for such effects. For instance, obsessive play (e.g., compulsive gaming) or play that is done to avoid important issues (e.g., illness issues, relational problems) may come at a cost because it interferes with other important responsibilities. Likewise, if an individual pursues play when his/her partner is in need of support, the behavior is likely to have negative consequences because it is insensitive and unresponsive to the partner’s needs (Feeney & Collins, 2014; Reis, Clark & Holmes, 2004). Moreover, if during a competitive play activity one’s primary goal of having fun becomes replaced by winning, the activity would no longer be considered play and could potentially have negative consequences. Much more work is needed to better understand circumstances under which play may have negative effects.

As a whole, this work suggests that daily play may improve the well-being of individuals living with type 1 diabetes without distracting from their self-care responsibilities. On days patients reported doing something fun with their partner, their mood was better, they told their partner more about their diabetes, they felt their partner was more supportive, and they felt they handled the day’s stressors better than on days when they did not. Findings from this and future work may be used as the basis of interventions in which couples are encouraged to make time for play, with the goal of improving well-being and bringing the partners closer together. From this solid relational base, couples may be better equipped to successfully work as a team in diabetes management.