Diabetes-specific family conflict and responsibility among emerging adults with type 1 diabetes.

Introduction

Type 1 diabetes (T1D) is one of the most common childhood chronic illnesses, with roughly 1 out of every 500 adolescents living with the disease (Mayer-Davis et al., 2017). Late adolescence and early young adulthood, also known as emerging adulthood, is a particularly risky time for diabetes management (Miller et al., 2015; Peters & Laffel, 2011). Registry studies suggest as few as 14% of youth ages 15–24 meet standard recommendations for glycemic control (e.g. hemoglobin A1c <7.5%), and A1c values peak at age 19 (Miller et al., 2015).

Emerging adulthood is a distinct developmental period marked by increased independence, changing living situations, and instability (Arnett, 2000). Hanna’s framework for youth with type 1 diabetes during the emerging adulthood transition specifies a number of personal and environmental characteristics that exert influence in the context of these transitional events, and impact longitudinal health, developmental, and behavioral outcomes (Hanna, 2012). Given the lack of research with emerging adults with diabetes, many of the constructs were included based on evidence from studies with younger adolescents. However, there is a significant need to better understand how key personal and environmental characteristics impact health outcomes, and identify potential interrelations among these characteristics. The current study focuses on family functioning in the context of emerging adulthood, including adolescent and parent reports on diabetes-specific family conflict and responsibility for diabetes care.

During adolescence, family functioning is an important component in the successful management of T1D, as supportive parent involvement in diabetes care can buffer against poor adherence and risky behaviors (King, Berg, Butner, Butler, & Wiebe, 2014; Wysocki et al., 2009). However, it is also common to see a shift in responsibility for diabetes management tasks (e.g. checking blood glucose levels, calculating and administering insulin doses) during adolescence, and parent involvement in these daily activities typically decreases as youth age (Anderson, Ho, Brackett, Finkelstein, & Laffel, 1997; Peters & Laffel, 2011). Although the initial shift in responsibility may occur before or during high school, there is an additional shift post-high school when emerging adults experience transitional events such as moving out of the family home or transferring to adult medical care (Hanna, 2012). The type of parental support also changes, often shifting from tangible support during adolescence (e.g. refilling supplies, making appointments) to more emotional support (e.g. encouragement, reminders) during emerging adulthood (Hanna, Weaver, Stump, Guthrie, & Oruche, 2014; Markowitz, Garvey, & Laffel, 2015; Wysocki & Gavin, 2004). Some families may navigate this quite well, but over- or under-involved parenting may interfere with the successful mastery of adult roles, specifically in terms of daily diabetes tasks, and places emerging adults at increased risk for poor diabetes outcomes.

As emerging adults take on increasing amounts of responsibility for their diabetes care, the frequency of conflictual interactions around diabetes care or sources of diabetes-specific conflict also may change. Research has consistently demonstrated a strong link between diabetes-specific family conflict and poor diabetes outcomes in adolescence (Hanna et al., 2014; Hilliard et al., 2013; Ingerski, Anderson, Dolan, & Hood, 2010; Moore, Hackworth, Hamilton, Northam, & Cameron, 2013). However, few researchers have examined diabetes-specific family conflict during the unique transitional period of emerging adulthood when these aspects of family functioning are likely shifting. Further, although direct diabetes management tasks such as checking blood glucose levels are endorsed as the most conflictual during adolescence (Gray, Dolan, & Hood, 2013; Hood, Butler, Anderson, & Laffel, 2007), less is known about what aspects of diabetes care remain conflictual as parents become less involved during the transition to adulthood.

Hanna and colleagues (2014) examined the parent-youth relationship, including the nature of parent involvement in diabetes care, during the year post-high school and found that emerging adults perceived less parental involvement and less family conflict across this time period. However, only youth-reported outcomes were presented. Additionally, while parent involvement in diabetes care typically decreases during late adolescence and emerging adulthood, there is evidence to suggest that parents are not completely removed from management tasks, providing continued opportunity for family conflict to occur. For example, parent knowledge about diabetes management remains an important protective factor among emerging adults (Berg et al., 2017), suggesting that better diabetes management occurs in the context of continued parental support. Further, Goethals and colleagues (2017) demonstrated the positive role of parents in diabetes management during emerging adulthood, yet benefits only emerged when parents did not place blame on the emerging adult (i.e., low psychological control) and when parents made emerging adults feel better after discussing diabetes worries (i.e., high responsiveness). Thus, it is possible that when parents remain involved in diabetes management and are experiencing conflict during this transition, emerging adults may experience detrimental outcomes. Conversely, poor glycemic control (e.g., higher HbA1c) during emerging adulthood may trigger increases in parental involvement and related diabetes-specific conflict.

Parents and emerging adults also may have differing views of diabetes responsibility and diabetes-specific family conflict; however, this has not been accounted for in the literature to date. A mismatch in perceptions of responsibility allocation may contribute to family conflict and subsequent diabetes outcomes (Miller & Drotar, 2003; Savin et al., 2018). As parents are still involved in diabetes care in emerging adulthood, it is important to obtain both reports to better understand their unique contributions to glycemic control (Goethals et al., 2017). Further, much of the research on family conflict in adolescents and emerging adults has been cross-sectional, and therefore unable to describe how it impacts diabetes outcomes over time.

The current study aimed to evaluate diabetes-specific family conflict and perceived responsibility for diabetes care over a one year period close to the time of high school graduation. We assessed diabetes-specific family conflict and responsibility in a sample of emerging adults with T1D and a primary caregiver at baseline (during senior year of high school), including demographic and clinical characteristics associated with these indicators of family functioning, changes over the one year period, and predictive associations with glycemic control. As diabetes-specific family conflict has been less studied in emerging adult populations, we also examined specific diabetes behaviors that were most frequently rated as conflictual. We expected diabetes-specific family conflict and parental involvement in diabetes care to decrease during the year post-high school, and hypothesized that these changes would predict better glycemic control. We also expected that higher levels of diabetes-specific family conflict and parental responsibility would interact to predict worse glycemic control. Specifically, we expected that among families in which the emerging adult held more responsibility for diabetes care tasks, the association between family conflict and glycemic control would be reduced. As a secondary analysis, we also examined predictive associations of glycemic control on diabetes-specific family conflict and family responsibility. Findings from this study can add to the existing literature by describing the nature of family conflict in emerging adults with T1D and its impact on diabetes outcomes during a challenging developmental period.

Methods

Participants were part of a larger longitudinal study examining executive functioning, adherence, and other predictors of T1D-related health outcomes in emerging adults with T1D. At study enrollment, all participants were high school seniors in the second half of their senior year with a diagnosis of T1D for at least 6 months and no other major chronic illnesses, psychiatric disorders, developmental disorders, or significant cognitive limitations. All participants were fluent in reading and writing English. Potential participants were recruited from a large pediatric diabetes clinic in the Mid-Atlantic. A total of 207 recruitment letters were sent to eligible participants. Of these, 48 potential participants were unable to be reached, and 4 participants cancelled their upcoming clinic visit after the recruitment letter was sent and were subsequently not contacted by the research team. Of the resulting 155 participants who were contacted, 17 were ineligible and 50 were not interested in participating, due primarily to time constraints and/or emerging adult disinterest. Eighty-eight participants expressed initial interest in participating; nine did not complete consent and/or baseline questionnaires, resulting in a final sample of 79 emerging adults and 79 corresponding primary caregivers (hereafter referred to as “parents”). At study completion, all participants had graduated from high school, with 65% of our sample living independently from parents at least part of the year (e.g. living on campus at college/university) and 90% enrolled in post-secondary education. Table 1 provides more information about the convenience sample’s demographic characteristics.

Emerging adult participants provided written assent or consent, and participating parents provided written consent. Following informed consent, all participants completed questionnaires at 3 time points (baseline, Time 2, and Time 3). Each data collection time point was approximately 5 to 6 months apart (M months between baseline and Time 2 = 5.16 months; M months between Time 2 to Time 3 = 5.18 months). Questionnaires included demographic information and psychosocial and health outcome measures, and emerging adults reported current living situations at all 3 time points. Diabetes-specific measures were chosen as they have been used in other work with this population to provide information about family functioning specifically within the context of managing T1D (Hilliard et al., 2013; Ingerski et al., 2010; Savin et al., 2018). Glycosylated hemoglobin (HbA1c) was extracted from the medical records at each time point. Study visits took place during regularly scheduled clinic visits with the medical team; a minority of participants took questionnaires home after clinic visits or completed questionnaires by mail as needed. In these cases, HbA1c values closest to the time when questionnaires were completed were used. Emerging adults received a gift card for completing questionnaires at each time point ($30 at baseline; $35 at Time 2; $40 at Time 3). This study was approved by the Institutional Review Board (IRB) at Children’s National Health System (Protocol #Pro00001399; Predictors of Health Outcomes in Emerging Adults with T1 Diabetes).

Diabetes-specific family conflict was assessed using the Revised Diabetes Family Conflict Scale (DFCS-R) (Hood et al., 2007), a 19-item questionnaire that measures the frequency of family conflict around direct and indirect diabetes management tasks over the preceding month. Direct management tasks include items such as checking blood glucose levels and administering insulin, while indirect management tasks include items such as making appointments and monitoring diabetes supplies. Participants rated each item on a 3-point Likert scale (1 = almost never argue, 2 = sometimes argue, and 3 = almost always argue) with scores ranging from 19–57. Higher scores indicate higher levels of family conflict. Parent and emerging adult participants rated each item separately. Scores on this measure have been strongly correlated with diabetes health outcomes (Gray et al., 2013; Herzer, Vesco, Ingerski, Dolan, & Hood, 2011). Internal consistencies for this sample were good and comparable to previously reported statistics (emerging adult α = .87; parent α = .87).

To measure changes in parent involvement post-high school, participants reported on who was primarily responsible for various diabetes management tasks using The Diabetes Family Responsibility Questionnaire (DFRQ) (Anderson, Auslander, Jung, Miller, & Santiago, 1990). The DFRQ is a 17-item questionnaire that assesses shared responsibility and decision-making about diabetes management tasks; a specific time frame was not specified. Participants responded to questions on a 3-point Likert scale (1 = emerging adult takes or initiates responsibility for this most or all of the time, 2 = emerging adult and parent share responsibility for this about equally, and 3 = parent takes or initiates responsibility for this most or all of the time). Both parent and emerging adult participants completed this questionnaire separately, reporting their perceptions of division of responsibility. Higher scores indicated more parent responsibility. The internal reliability of scores in this sample were adequate (emerging adult α = .74; parent α = .70).

HbA1c was obtained from the medical record at the medical visits closest to baseline, Time 2, and Time 3 as available. These data were obtained using the DCA Vantage Analyzer for point-of-care testing, a reliable method for obtaining HbA1c values (Szymezak, Leroy, Lavalard, & Gillery, 2008). HbA1c reflects the average glucose level over a 2–3 month period, with higher levels indicating poorer glycemic control.

Associations among diabetes-specific family conflict, responsibility, and glycemic control were assessed using SPSS version 23 (IBM) and SAS 9.2 (SAS Institute, Inc.). Mean scores at all 3 time points were used to characterize change in diabetes-specific family conflict and diabetes responsibility over the first year post-high school. Item-wise descriptions of the DFCS-R measure were used to describe the most commonly reported topics associated with family conflict during the post-high school period, and to highlight areas of diabetes management that may require more attention from parents and emerging adults as they navigate this transition. Generalized Estimating Equations (GEE) models were conducted to evaluate changes in parent- and emerging adult-reported diabetes-specific family conflict and diabetes responsibility over time and relevant demographic and clinical factors that may be related to diabetes-related family conflict and responsibility. Linear regression models were conducted to evaluate if baseline diabetes-related family conflict and responsibility for diabetes care and their interaction had a significant effect on A1c at Time 3. Because zero is not a meaningful value in the diabetes-related family conflict and responsibility scores, in order to make the main effects of the two variables interpretable when their interaction is included in the model, the two measures were recoded as deviations from their grand mean scores, respectively (Hox, 1995). Relevant covariates were controlled for in each regression model, including baseline A1c, sex, ethnicity, income, duration of diabetes, insulin regimen, and living away from home; covariates were selected based on common non-modifiable factors influencing diabetes care and potential relations with the outcome variable of interest (Clements et al., 2016; Pierce & Wysocki, 2015). To better inform the directionality of these relations, the linear regression models were repeated with the same relevant covariates and baseline diabetes-specific family conflict or responsibility to evaluate whether glycemic control predicted diabetes-specific family conflict or responsibility at Time 3.

As expected, there was participation attrition from baseline to Time 2 and Time 3; 87.3% of the total sample (n = 69) had data at Time 2 and 84.8% of the total sample (n = 67) had data at Time 3. We applied multiple imputation (MI) to handle missing data due to participant attrition; MI utilized a Markov Chain Monte Carlo (MCMC) technique. MCMC assumes missing at random (MAR), which allows missingness to be dependent on observed variables (e.g., baseline value; Arbuckle, 1996; Little & Rubin, 2002). To capture the uncertainty in missing value imputations, MI estimates the values multiple times (10 times in this study). Our model was implemented separately on each data set, parameter estimates were averaged over the set of analyses, and standard errors were computed using the average of the standard errors over the set of analyses and the between analysis parameter estimate variation (Rubin, 1987). MI has been shown to produce adequate results in the presence of missing data; its performance is robust to small sample size and departure from normality assumptions (Graham, Hofer, Donaldson, MacKinnon, & Schafer, 1997; Graham & Schafer, 1999; Rubin, 1987; Schafer & Graham, 2002).

Results

Results indicate that diabetes-specific family conflict was relatively low and did not significantly change across the year post-high school (see Table 2). Parent- and emerging adult-reported conflict was not significantly different from each other at each time point (ps>.05). GEE models indicated parent- and emerging adult-reported mean family conflict scores did not differ across time (p>.05). Male emerging adults reported lower diabetes-specific family conflict than females (β = −1.46, p<.05). Further, parent-reported family conflict was lower in families where the emerging adult used an insulin pump (β = −4.05, p<.05). See Table 3.

Consistent with previous findings, the most frequently endorsed conflict items were around direct diabetes management tasks, such as remembering to check blood glucose levels or arguing over the results of blood glucose checks. Results indicate that parents and emerging adults mostly agreed that these tasks were the most common conflict items, but there were inconsistencies between parent and emerging adult perceptions of conflict around food-related items (e.g. meals and snacks, what to eat when away from home), such that parents perceived more conflict around these items than emerging adults. Table 4 presents the percentage of parents and emerging adults who reported at least some conflict (i.e., DFCS-R score >1) around the most commonly endorsed conflict items at baseline and Time 3.

Parent responsibility in diabetes care decreased over the year post-high school per both emerging adult and parent report (see Table 2); however, at each time point, parents perceived that they were more responsible for diabetes care than emerging adults perceived parents to be involved (ps<.01). GEE models found a significant decrease in parental involvement in diabetes care at both follow-up periods as reported by parents and emerging adults. Controlling for key demographic variables (sex, ethnicity, household income, duration of diabetes, insulin regimen, and living situation), parent-reported parental involvement in diabetes care decreased from baseline to Time 2 (β = −2.56, p<.001) and Time 3 (−3.54, p<.001) follow-up periods. Similarly, emerging adult-reported parental involvement in diabetes care decreased at Time 2 (β = −1.33, p<.01) and Time 3 (β = −2.72, p <.001) follow up periods. Ethnicity emerged as a significant contributor to parent responsibility, with Caucasian participants reporting lower parental involvement. See Table 3.

Higher parent-reported diabetes-specific family conflict was positively correlated with parent-reported parent responsibility at baseline (r = 0.46, p<.01), Time 2 (r = 0.55, p<.001), and Time 3 (r = 0.46, p<.01). A similar relationship was found between emerging adult-reported conflict and parental responsibility at baseline and Time 3 (baseline r = 0.32, p<.01; 6 month r = 0.20, p =.10; Time 3 r = 0.26, p<.05).

Emerging adult reported diabetes-specific family conflict was positively correlated with HbA1c at baseline (r = 0.47, p<.01) and Time 3 (r = 0.41, p<.01), indicating that more diabetes-specific family conflict is associated with worse glycemic control. Similarly, emerging adult reported parental involvement in diabetes care was positively associated with HbA1c at Time 3 (r = 0.25, p<.01), indicating that more parent responsibility in diabetes care was associated with worse glycemic control. After controlling for gender, ethnicity, income, duration of diabetes, insulin regimen, living situation, and baseline HbA1c, emerging adult-reported diabetes family conflict and diabetes family responsibility at baseline was not significantly associated with HbA1c at Time 3.

Parent-reported diabetes-specific family conflict was positively correlated with HbA1c at all 3 time points: baseline (r = 0.56, p<.01), Time 2 (r = 0.56, p<.01), and Time 3 (r = 0.35, p<.05), indicating that more diabetes-specific family conflict is associated with worse glycemic control. Parent-reported parental involvement in diabetes care was positively associated with HbA1c at all 3 time points: baseline (r = 0.29, p<.05), Time 2 (r = 0.39, p<.01), and Time 3 (r = 0.35, p<.01), indicating more parent responsibility in diabetes care is associated with worse glycemic control. After controlling for gender, ethnicity, income, duration of diabetes, insulin regimen, living situation, and baseline HbA1c, higher parent-reported conflict at baseline predicted worse HbA1c at Time 3 (β = 0.11, p<.01). The interaction between parent-reported diabetes-specific family conflict and responsibility for diabetes care was also significant, indicating that when parent responsibility for diabetes care increases by 1 unit, the direct effect of conflict on HbA1c is reduced (β = −0.03, p<.01). See Table 5.

To inform the directionality of the relationship among diabetes-specific conflict, responsibility, and glycemic control, exploratory linear regression analyses were conducted to examine whether glycemic control at baseline predicted diabetes-specific family conflict or responsibility at Time 3. After controlling for gender, ethnicity, income, duration of diabetes, insulin regimen, living situation, and baseline diabetes-specific family conflict, baseline HbA1c did not predict diabetes-specific family conflict at Time 3 by emerging adult report (β = −0.28, p=.30) or parent report (β = −0.56, p=.15). Similarly, after controlling for gender, ethnicity, income, duration of diabetes, insulin regimen, living situation, and baseline diabetes responsibility, baseline HbA1c did not predict diabetes responsibility at Time 3 by emerging adult report (β = 0.07, p=.78) or parent report, although parent-reported responsibility trended toward significance (β = 0.70, p=.06).

Discussion

Diabetes-specific family conflict remained present over the year post-high school regardless of living situation, suggesting that families with existing conflict around diabetes management tasks may continue to have conflict even after emerging adults move out of the family home. Although our sample reported relatively low levels of family conflict, this aspect of parent-reported family functioning was a significant predictor of diabetes outcomes one year post-high school. As expected, greater parent responsibility predicted worse glycemic control. Although greater parent responsibility is consistently linked with better diabetes outcomes among early adolescents, our findings suggest that during a later developmental period, high levels of parental responsibility for care may actually be a marker of problematic health outcomes.

Parent-reported diabetes-specific family conflict and family responsibility during high school interacted to predict glycemic control at Time 3, suggesting these two aspects of family functioning may be working together to impact diabetes management during this transitional period. Interestingly, the interaction effect was in the opposite direction than expected. When parents reported greater than average responsibility, the negative effects of family conflict on HbA1c were reduced. It is possible that greater parental involvement during high school may be beneficial only for families experiencing conflict around diabetes management. Secondary analyses supported the directionality of these relations, as higher HbA1c at baseline did not predict increased family conflict or responsibility during the year post-high school. Together, these findings offer an important motive for families to reduce diabetes-specific conflict during the adolescent years, and indicates the opportunity for health care providers and psychologists in identifying families that may need additional support from health care professionals to reduce conflict while maintaining appropriate levels of involvement during adolescence. For example, routine psychosocial screening in diabetes clinic visits can assist health care providers in identifying families who may benefit from interventions to improve family communication and reduce conflict around diabetes management (Hilliard et al., 2018).

While parent-reported family conflict significantly predicted glycemic control at the 12 month follow-up, emerging adult reports did not. This is somewhat consistent with the literature, though findings in this area have been mixed. For example, one study found only parent-reported family conflict to be significantly associated with health outcomes in younger adolescents (Anderson et al., 2009), whereas another study found emerging adult report of family conflict to be significantly associated with HbA1c, although parent-reported measures were not used (Hanna et al., 2014). Additionally, Lancaster and colleagues (2015) found that both parent- and child-reported conflict were significantly associated with HbA1c in a cross-sectional study with youth ages 8 to 18 years old. Nonetheless, our findings are consistent with the developmental framework of emerging adulthood, when emerging adults are obtaining greater autonomy and their differing perception of family dynamics may be evidence of their increasing autonomy. While there was no statistical difference in parent versus emerging adult reports of conflict, their longitudinal relationship with glycemic control did differ. This may reflect the continued role of the parents during the year post-high school, which is consistent with findings regarding discrepancies in parent-versus adolescent-reported adherence (Berg et al., 2016).

Given the lack of prior research on family functioning in emerging adulthood, this study examined demographic and clinical characteristics associated with diabetes-specific conflict and responsibility. Results indicated that males and parents of emerging adults on insulin pumps reported less diabetes-specific family conflict. These findings are somewhat aligned with what has been found in younger samples of adolescents. For example, Hilliard and colleagues (2013) found intensive insulin regimens were associated with less family conflict. Further, the finding that Caucasian parents report less involvement in diabetes care may reflect cultural differences in expectations for adulthood (Arnett, 2000). For example, some research has found that interdependency is more common in Latino and Asian cultures than Caucasian cultures (Campos & Kim, 2017). However, given the relatively small sample size, it is important that these findings are replicated in larger and more diverse samples to better understand non-modifiable predictors of these indicators of family functioning.

Remembering to check blood glucose levels, results of blood glucose monitoring, and discussion about meals and snacks were the most commonly endorsed conflict items, which is consistent with prior research (Gray et al., 2013; Hood et al., 2007; Savin et al., 2018). These items are considered direct management tasks, and conflict around these tasks are consistently better predictors of higher HbA1c levels than conflict around indirect management tasks (Hood et al., 2007). Remembering to check blood glucose levels was the most commonly endorsed conflict item by both parents and emerging adults. An essential part of successfully managing T1D, this task is commonly used as an indicator of adherence to the medical regimen and consistently has a significant association with glycemic control (Miller et al., 2013). Further, conflict around the results of blood glucose checks was the second highest rated conflict item, suggesting that even when emerging adults are checking blood glucose levels, parents may become upset about the results and engage in argumentative discussions with the emerging adult. Emerging adults may adopt maladaptive behaviors such as avoiding blood glucose checks altogether, leading to even more conflict around this management task.

Results from the current study have implications for interventions for this developmental period. Although family conflict was relatively low in our sample, findings suggest that when present, it remains a significant predictor of glycemic control. Further, patterns of family responsibility reflected increasing independence on the part of the emerging adult and parents did not increase involvement across emerging adulthood when glycemic control was poor, as poor glycemic control at baseline did not predict increases in parental responsibility or diabetes-specific family conflict during the year post-high school. Existing behavioral interventions that target diabetes-related family conflict and family communication around diabetes management could be adapted to specifically address the unique challenges of this age group, as most of these interventions have only been examined in younger adolescent samples. For example, a Family Teamwork program designed to promote parent-child collaboration and goal-setting for diabetes care prevented the decline in glycemic control typically see in early adolescence (Holmes, Chen, Mackey, Grey, & Streisand, 2014). Additionally, Thompson and colleagues (2016) adapted a Family Teamwork program for web-based delivery to increase reach and potential impact. While the focus on parent-emerging adult collaboration may require a different developmental approach, strategies from interventions such as Family Teamwork could be modified to promote the shift in parental support from tangible to emotional support and increasing self-efficacy for emerging adults with type 1 diabetes (Monaghan, Helgeson, & Wiebe, 2015).

As emerging adults typically meet with a multidisciplinary team every 3–4 months for routine follow-up visits, health care providers are in a unique position to guide emerging adults as they navigate this transition period. Health care providers should encourage discussion about how families will manage the year post-high school, with a particular focus on communication around and expectations for diabetes management tasks, regardless if an emerging adult has plans to move out of the family home. For example, health care providers can engage in conversation about anticipating glucose variability during times of transition, clarifying who is responsible for what, and problem-solving around diabetes tasks (i.e. treating a high/low, compromising on meals/diets). Further, dieticians can work with families to brainstorm healthy meal and snack options in order to decrease conflict around these items. Assessing readiness for adult diabetes care is another component of recommended care for this age group, and psychologists play an important role in the evaluation of family functioning in preparation for this transfer, including families who may benefit from strategies to reduce conflict and support emerging adult autonomy in preparation for adult diabetes care. Psychologists can promote cohesive family communication, assist with screening initiatives, and provide counseling services when needed.

Findings should be considered within the context of some limitations. First, this sample reflects a relatively high-functioning sample of emerging adults. Recruitment rates were somewhat lower than expected and most emerging adults enrolled in higher education and were living away from home one year post-high school. The changing living and educational experiences that occur over emerging adulthood may have precluded some participants from enrolling in the study, and those with more uncertain plans may have declined participation. The fairly homogenous sample also may have contributed to the relatively low levels of reported diabetes-specific family conflict, as emerging adults in our sample may have been more mature and motivated than a higher-risk sample of youth. Previous research had found higher rates of conflict for emerging adults living independently of parents (Hanna et al., 2014). However, in our sample, living away from home was associated with lower HbA1c levels, supporting the idea that living away from home could be a marker of greater self-management capabilities. In order to better understand challenges inherent in this age group, future research should include a more diverse sample in terms of post-high school plans and family functioning, including those reporting higher family conflict. Family culture may also influence how families maintain involvement in diabetes care, and a closer look at culture and responsibility is warranted in future research (Campos & Kim, 2017). It is also recommended that future studies restrict enrollment to youth diagnosed with T1D for at least one year to minimize the impact of the honeymoon period and allow diabetes routines to be fully established. As diabetes-specific family conflict and family responsibility were assessed using self-report measures only, observational methods in addition to self-report measures may provide a more accurate picture of the nature of family conflict in this population (Jaser & Grey, 2009) and should be considered in future research. Lastly, although secondary analyses did not support the notion that poor glycemic control in late adolescence predicts increased diabetes-specific conflict and parent responsibility, it is possible that poorer HbA1c in early adolescence may be related to changes in these aspects of family functioning prior to the emerging adulthood period. Analyses that examine the bidirectional nature of these family processes across adolescence and into emerging adulthood could be important contributions of future work.

During the year post-high school, diabetes-specific family conflict remains an important predictor of diabetes outcomes and should continue to be discussed with emerging adult patients and families. Further, family conflict may interact with family responsibility around diabetes management and should be considered within the context of how the family is transitioning responsibility to the emerging adult and preparing for adult medical care. Psychologists and other health care providers play an important role in identifying families experiencing conflict around diabetes tasks or struggling with the shift in responsibility during this transitional period, and can support the shift in parental responsibility in care while promoting self-efficacy and appropriate autonomy in emerging adults. Helping families incorporate strategies to promote developmentally-appropriate engagement in routine diabetes care tasks can lead to optimal diabetes health outcomes during the vulnerable period of emerging adulthood.