Low parental stress and positive well-being in Finnish children and adolescents with type 1 diabetes

WHAT IS ALREADY KNOWN ON THIS TOPIC International research has shown that family dynamics significantly influence metabolic control in pediatric type 1 diabetes (T1D), but little is known about these factors in the Finnish context, highlighting the need for this study.WHAT THIS STUDY ADDS Children with more stable glycemic control had better well-being scores, and parents of young children with T1D reported higher stress levels.HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE, OR POLICY This study underscores the need to integrate routine psychosocial assessments into pediatric diabetes care to better identify and support families at risk. Its findings may facilitate future research and policy to promote family-centered management strategies.Introduction Type 1 diabetes mellitus (T1D) is an autoimmune condition characterized by the inability of pancreatic B cells to produce sufficient insulin. This leads to hyperglycemia, which, if inadequately treated, can cause long-term damage to the eyes, peripheral nerves, kidneys, and cardiovascular system. T1D is a lifelong condition without a cure or prevention and is one of the most common pediatric chronic illnesses. 1 There is a high rate of T1D incidence among children and adolescents aged <15 years in most European countries. For many decades, the highest incidence has been observed in Finland (from 2013 to 2022: 56.42/100 000/year; 95% CI 47.80 to 65.04).2While much research has focused on advancements in the clinical management of diabetes, psychosocial factors affecting clinical care have rarely been addressed. The International Society for Pediatric and Adolescent Diabetes (ISPAD) guidelines emphasize the importance of psychological well-being and quality of life in children and adolescents with T1D. These should be routinely assessed during outpatient follow-ups in an age-appropriate manner, and attention should be paid to the well-being of caregivers.3Over the decades, the treatment of T1D has significantly improved, largely because of technological advancements.4 However, most children and adolescents fail to reach treatment targets,5 6 possibly owing to a combination of physiological, psychological, and behavioral factors, and the burden of diabetes care can be overwhelming. Psychological disorders, such as depression, are common, and the incidence of suicide among adolescents with T1D has increased.7 Pediatric T1D is often characterized as a ‘family disease’ due to the crucial role of family interactions and parental support.8 9 Additionally, anxiety, burnout, and depression are common among the parents of a child with T1D.10–12According to Sepa et al,13 familial psychosocial strain, such as high parenting stress, is involved in the onset or progression of diabetes-related autoimmunity during the first year of life. Furthermore, adverse life events experienced in childhood, such as death or illness in the family, family conflicts, and unemployment, are associated with a higher risk of future T1D diagnosis, regardless of genetic predisposition.14 Psychosocial stress within families can negatively impact children by affecting hormonal levels and nervous signals, which influence insulin sensitivity and the immune system.13 Additionally, stressful familial life events are linked to worsening diabetes management in children.15The importance of family dynamics in the metabolic control of pediatric T1D has been recognized in international studies.16–18 Positive parental emotional support, effective family communication, and adequate parental guidance in diabetes-related care have been associated with better metabolic control, while high levels of family conflict as well as negative and unsupportive parental behavior are linked to poorer metabolic control and adherence.17 However, the family dynamics and well-being of pediatric patients with T1D and their families in Finland remain underexplored.19 20We, therefore, aimed to determine (1) the levels of stress in parents of children with T1D, (2) whether the experienced stress was associated with their child’s glycemic control, (3) the well-being of Finnish children with T1D, and (4) whether children’s well-being was linked to metabolic control.Materials and methods Participants and protocol The children participated in a larger cross-sectional study that assessed sociodemographic and psychosocial factors affecting glycemic control in children with T1D. More detailed sociodemographic characteristics of the participants and detailed data collection methods have been reported previously. 5 The children and their parents were invited to participate in this study and complete questionnaires during regular outpatient control visits at the pediatric diabetes clinic of Tampere University Hospital (TAYS) between November 2020 and October 2021.The inclusion criteria were as follows: age between 1 and 16 years, diagnosis of T1D at least 1 year prior to recruitment, and follow-up at the TAYS outpatient pediatric diabetes clinic. If multiple siblings in a family had T1D, only the oldest sibling was selected for participation. The exclusion criterion was the inability to speak or understand Finnish.At the time of the study, 429 children under the age of 16 years were followed up at the TAYS pediatric diabetes clinic. Of the 334 eligible children, 60% (n=199) participated in this study. The most common reason for exclusion was the short duration of diabetes; only approximately 1% of families were non-Finnish speaking.Informed consent was obtained from participants aged 6–16 years in an age-appropriate manner and from all caregivers. Given that the study involved children and adolescents, participation was entirely voluntary. The study posed no costs or risks to participants. All information collected about the child and family was handled confidentially and coded to ensure that individual participants could not be identified from the study results.Measurements Demographic and diabetes measures The demographic data in table 1 (sex and age) and diabetes-related information (T1D duration, most recent glycated hemoglobin (HbA1c) record, time in range (TIR) from the previous 14-day period if available, and type of insulin treatment) were collected from the patient’s medical records.Table 1Participant characteristics in the study (N=199)Whole study (N=199)WHO-5 Well-Being Index analyses (N=180)PSI-SF total stress analyses (N=133)Dropout (n=66) analysis p valueMale gender, n (%)104(52)96(53)69(52)0.878Age, years, median (IQR, range)11(9–13, 2–15)11(9–13, 6–15)10(7–11, 2–13)<0.001Family characteristics, n (%)0.353 Nuclear family135(68)124(69)96(72) Other family arrangement56(28)53(29)36(27) Not known8(4)3(2)1(1)Time from type 1 diabetes diagnosis, years, median (IQR, range)4(2–7, 1–13)5(2–7, 1–13)4(2–7, 1–11)0.045Glycated hemoglobin, %,(mmol/mol), mean (SD, mmol/mol)7.7 (60.9)(12.5)7.8 (61.4)(12.6)7.5 (58.3)(8.8)0.002Time in range, %, mean (SD)55.4(17.1)54.4(17.1)58.0(15.4)0.013Treatment type, n (%)0.883  Insulin pump107(54)97(54)72(54)  Insulin injections92(46)83(46)61(46)Dropout analysis compared caregivers who had PSI-SF total stress data with those without.T1D=Type 1 diabetes; HbA1c=Glycosylated hemoglobin; IQR=Interquartile range; SD= Standard deviation; N/n=number; PSI-SF=Parenting Stress Index Short FormPSI-SF, Parenting Stress Index Short Form.Background information questionnaire A questionnaire was developed for the purposes of the study to collect psychosocial and other demographic information about the family. The questions included the caregiver’s education level, employment status, current smoking status (yes/no), diagnosis of T1D (yes/no), family structure (nuclear family or other), presence of a support network outside the family (yes/no), and number of siblings. Furthermore, information on the hobbies of the child (yes/no), special educational arrangements at school (yes/no), and any other medical conditions (no/yes; specify) was collected.Parenting Stress Index Short Form (PSI-SF) questionnaire Caregivers completed the PSI-SF, fourth edition (PSI-4-SF, hereafter PSI-SF), an instrument designed to assess general parenting stress in parents of children aged 0–12 years. 21 The PSI-SF consists of 36 items divided into three subscales: (1) the Parental Distress (PD) subscale (12 items), which measures the perceived stress related to parenting; (2) the Parent-Child Dysfunctional Interaction (P-CDI) subscale (12 items), which assesses the parent’s perception on whether the child meets their expectations and whether interactions with the child are rewarding for the parent; and (3) the Difficult Child (DC) subscale (12 items), which assesses the extent to which children have behavioral characteristics that make them easy or difficult to manage.21The PSI-SF is rated on a five-point Likert scale ranging from ‘strongly agree’ to ‘strongly disagree’. The subscale scores range from 12 to 60. The sum of the PD, P-CDI, and DC scores provides a total stress (PSI-TS) score ranging from 36 to 180, with higher scores indicating higher levels of parental stress. Raw scores were converted to percentiles according to the manual, with scores in the 85th to 89th percentiles considered high and those in the 90th percentile or higher considered ‘clinically significant’.21The coefficient alphas for the PSI-SF subscales were 0.90 for PD, 0.89 for P-CDI, 0.88 for DC, and 0.95 for the PSI-TS scale.21 A translated Finnish version of the PSI-SF was obtained from Psychological Assessment Resources, the publisher of psychological assessment materials.WHO-5 Well-Being Index (WHO-5) questionnaire Patients aged 6–16 years were asked to complete the WHO-5 questionnaire, developed to assess emotional well-being over a 14-day period in various age groups. The WHO-5 consists of five statements and uses a six-point Likert scale ranging from 0 (not present) to 5 (constantly present). The total raw score, ranging from 0 to 25, is multiplied by four to obtain the final score, ranging from 0 to 100, with higher scores representing better well-being. A score ≤50 indicates depressive affect. Cronbach’s alpha for this five-item scale is 0.82. The WHO-5 has been translated into more than 30 languages, including Finnish, and has been used in diabetes studies worldwide. 22 23Statistical analysis For analyses, some characteristics were recategorized. Diseases occurring in children other than T1D were classified into four groups. The atopy and/or asthma group included children with atopic eczema (n=9), allergies (n=9), and/or asthma (n=10). Autoimmune diseases included hypothyroidism (n=12), juvenile rheumatoid arthritis (n=1), and celiac disease (n=19). The neuropsychiatric/neurological condition group included neuropsychiatric (n=10) and/or neurological (n=1) conditions reported by parents. Other conditions included lactose intolerance (n=2), hearing defects (n=2), eye disease (n=2), scoliosis (n=1), kidney disease (n=2), and cardiac disease (n=1). Information on other diseases besides T1D was missing for eight children. Social support was a combination of support from friends (n=41), neighbors (n=9), relatives (n=122), social work (n=3), and/or other support outside the family (n=4).Children and caregiver characteristics were described by population means with SDs, median (Md) with IQRs and ranges, or number of cases with percentages and tested using the Mann-Whitney or Pearson’s χ2 test, as appropriate. Owing to the skewness of distributions of scores or subscales, pairwise interactions of HbA1c or TIR with other factors were analyzed using a bivariate Spearman’s rank correlation coefficient (rho) with 95% CIs. PSI-SF stress scores were measured only from one of the child’s caregivers, never from both. Combined information referred to data provided by the parents who completed the survey.The sample size in the WHO-5 analyses varied from 121 to 180 due to missing data on TIR (n=156), PSI-TS (n=121), and matters related to children. The correlations between WHO-5 and glycemic control (HbA1c and TIR) were analyzed using Pearson’s correlation, showing results by coefficient r with 95% CI. The increase in WHO-5 scores was modeled using univariable linear regression. Results are shown as beta (β) values with 95% CIs. The explanatory factors were modeled as either continuous, normally distributed, or dichotomous. Skewed distributions were normalized using logarithmic transformation (ln). Dropout analysis was performed on the children’s characteristics with missing information on the PSI-TS.Statistical analyses were performed using SPSS V.29.0.1.0 (SPSS, IBM Corp, Armonk, New York, USA). A two-sided p value <0.05 was considered statistically significant.Results Sociodemographic and diabetes-related data Specific sociodemographic and diabetes-related data are presented in table 1. Based on the dropout analysis, PSI-TS information was obtained from caregivers of younger children with better metabolic balance and shorter duration of T1D.Parental stress A total of 133 caregivers (67%) had their PSI-TS levels measured. Mothers (Caregiver 1, n=104) PSI-TS Md was 29 (IQR 12–49); for fathers (Caregiver 2, n=30), it was 50 (IQR 15–59). When the answers were combined, the caregiver PSI-TS Md was 31 (IQR 12–55, range 1–92). One mother and one father had clinically significantly high PSI-TS percentile scores ≥90th percentile (92nd and 90th percentiles, range 1–92 and 3–90, respectively). Weak correlations were observed between PSI-TS or subscale scores and the children’s glycemic control, as estimated by HbA1c or TIR (Spearman’s rho varied between −0.21 and 0.22, all p>0.10).In the age group analyses, combined PSI-TS information derived from either parent correlated significantly only with glycemic control in young children, as estimated by HbA1c and TIR. Higher total parental stress was associated with better glycemic control in young children. Among mothers of children aged <7 years, higher PSI-TS, P-CDI, and DC scores were associated with better HbA1c control, and P-CDI scores were associated with higher TIR (online supplemental Table 2).SP110.1136/bmjdrc-2025-005248.supp1Supplementary dataThe PSI-TS scores were higher in the insulin pump treatment (IPT) than in the insulin injection treatment (IIT) groups. A significant difference was found in the PSI-TS whole data score (Md 37, IQR 16–61 for those treated with IPT and Md 25, IQR 8–47 for those treated with IIT, p=0.017). The duration of diabetes did not correlate significantly with PSI-TS (rho 0.11, 95% CI −0.06 to 0.28, n=131, p=0.202) or with the subscales.Well-being of children A total of 180 children completed the WHO-5 questionnaire, with a mean score of 70 (SD=15, IQR 64–80, range 8–100). Especially, higher parental stress, measured by the PSI-TS score, was strongly related to poorer well-being in children (Pearson’s correlation coefficient r=−0.29, 95% CI −0.45 to −0.12, p≤0.001, n=124). In 14 participants (7.8%), the WHO-5 score was ≤50, indicating depressive symptoms. Those children were aged 10–14 years (girls 57%, n=8). WHO-5 scores did not correlate significantly with HbA1c levels (r=−0.075, 95% CI −0.219 to 0.072, p=0.318, n=180). The mean HbA1c level was higher (8.1% (65 mmol/mol), SD=16) in children (n=27) with missing TIR data. However, a positive correlation was found between the WHO-5 scores and metabolic control of the children, as evaluated by TIR (r=0.166, 95% CI 0.009 to 0.315, p=0.038, n=156) and according to unadjusted linear regression ( online supplemental Table 3). No significant interactions with HbA1c or TIR were observed in further analyses. The means of the WHO-5 scores were at the same level in the groups of factors to be modeled. The characteristics of the children and families that were associated with WHO-5 scores are presented in online supplemental Table 3.Discussion In this study, parental stress levels assessed using the PSI-SF questionnaire were low. Existing literature presents conflicting evidence regarding the relationship between parental stress and metabolic control in children and adolescents. Ferrito et al reported poor metabolic control as a significant source of parental stress,24 whereas another study reached the opposite conclusion.25 For the overall data of our study, no significant correlation was found between metabolic control in children and parental stress. However, despite the limited sample size (n=9), our findings suggest that parents of children aged <7 years who completed the PSI-SF questionnaire experienced higher stress levels than did parents of older children with T1D, and higher parental stress levels correlated with lower HbA1c and higher TIR in young children. This finding supports the results of a previous meta-analysis reporting that for younger children (aged 0–11 years), parenting stress was either unrelated or negatively associated with HbA1c levels.26 This may be because parents of older children often share responsibilities with their child, whereas parents of younger children bear full responsibility for managing their child’s T1D. Another explanation for the higher stress levels observed among parents of young children with T1D might be the fear of hypoglycemia, given that young children are at increased risk for hypoglycemic events.27 28 Also, according to ISPAD 2022 guidelines, younger age and low HbA1c were earlier indeed risk factors for severe hypoglycemia, but low HbA1c is no longer a strong predictor of severe hypoglycemia in pediatric T1D cohorts on contemporary therapy.28 Unfortunately, our data collection did not include information on time below range values, leaving the underlying causes of elevated stress levels among parents of young children partly unresolved.Our study offers additional insights into paternal stress, with fathers reporting slightly higher stress levels than mothers. A recent meta-analysis found that over half of children with T1D had a parent showing signs of depression. The authors emphasized the need for further research, particularly regarding fathers for whom data were limited.10 Although our study did not specifically aim to investigate parental depression, stress is a well-established risk factor for depression.29Living in a nuclear family appears to be associated with better glycemic control in children.5 30 In nuclear families, diabetes-related supervision by parents may be shared more equitably, and caregivers tend to report a better quality of life.30 In our study, the prevalence of children living in a nuclear family aligns with that of the general Finnish population (76.5%) in 2020.31 This finding may partially explain the moderate levels of perceived stress reported by the parents.A higher TIR was positively related to children’s well-being, as measured using the WHO-5 questionnaire. This result aligns with previous research indicating that better glycemic control, as measured by HbA1c, can enhance the overall quality of life.32 33 It is possible that better mental health affects the child’s motivation for diabetes care and, thus, affects glycemic control. Although our study did not confirm this relationship when using HbA1c levels, TIR correlates better with glycemic variability than does HbA1c,34 suggesting that stable glucose levels are related to overall well-being.The mean WHO-5 score of 70 in our study population was higher than that reported in both a previous diabetes study and the general child and adolescent population in Finland.23 35 However, 7.8% of the children had WHO-5 scores indicating depression. All of them were aged 10 years or older. The percentage is slightly lower than the international data on depression among children with T1D. Hood et al36 reported that the prevalence of depressive symptoms in children with T1D aged 10–18 years was 15.2%, with a higher prevalence observed among females. In healthy children under 10 years of age, the prevalence of depression was estimated to be 1.6%–7.1% depending on the severity of depression.37 Overall, in our study, T1D did not negatively affect the general well-being of the patients. However, the WHO-5 is a measure of well-being rather than a diagnostic tool for depression23; thus, the results should be interpreted with caution.Children’s engagement in hobbies and higher parental education were related to better well-being in the WHO-5 questionnaire, consistent with findings from the BELLA (German: BEfragung zum seeLischen Wohlbefinden und Verhalten, i.e. Survey on Mental Health and Well-Being) study,38 which linked external social support and parental socioeconomic status to improved mental health in children. These results underscore the role of parental resources in supporting diabetes care and well-being, while hobbies may provide stress relief, social interactions, and self-esteem development.Conversely, older age, due to puberty, increased responsibilities and self-management demands, as well as neuropsychiatric or neurological disorders and special school arrangements, was related to poorer well-being. Children with these conditions may face additional challenges in the management of T1D. Mental health also appears to influence treatment outcomes, as psychiatric comorbidities in pediatric T1D are linked to poor metabolic control.39In this study, higher parental stress, measured by the PSI-TS score, was strongly related to poorer well-being in children. One possible explanation is that higher parental stress may also lead to increased risk of diabetes-related conflicts with their children, which in turn could negatively affect the children’s well-being. This relationship highlights the potential impact of family dynamics and parental mental health on children’s well-being, as suggested by the BELLA study group.38 Additionally, parental unemployment had a negative impact on children’s well-being in our study, indicating the influence of socioeconomic instability on children’s well-being as demonstrated previously.38In the WHO-5 analyses, IPT was related to a near-statistically significant reduction in children’s well-being scores. IPT was also related to increased parental stress levels compared with IIT. Previous literature offers mixed findings regarding IPT and parental stress.40–42 However, recent articles suggest that the use of modern hybrid closed-loop insulin pumps is related to improved well-being and lower parental stress in families of children with T1D.33 43 During the data collection for our study, hybrid closed-loop pumps were yet to be used in our clinic. As the use of hybrid closed-loop pumps is rapidly increasing, it may positively affect the well-being of patients and their families.This study had some limitations. Neither the PSI-SF nor the WHO-5 questionnaires are diabetes-specific instruments. However, they are easy to use, validated, widely used in diabetes research, and have been translated into Finnish. When examining the PSI-SF results, it must be noted that, based on the dropout analysis, information was more likely to be missing from the parents of children with poorer glycemic control and older age, which may bias the results. Additionally, due to the small sample size of parents with children aged under 7 years (n=9) in the PSI-TS analyses, these results are approximate and should be interpreted with caution. Finally, as our study relied on voluntary participation and had a relatively low participation rate (60%), it is possible that families facing more significant challenges chose not to participate, which could introduce bias in the results. The low participation rate may be attributable to the sensitive nature of the questions included in the questionnaire. The reasons for declining participation in the study were not asked or investigated, and glycemic control data for individuals who refused to participate were not collected. However, the participating cohort was broadly representative of the clinic’s entire patient population with respect to glycemic control. The Md HbA1c in the study group was approximately the same as the mean HbA1c of all patients treated at our clinic prior to the study period in April 2020 (HbA1c 59.4 mmol/mol). Although glycemic control among children from non-participating families did not markedly differ from the clinic’s overall average, the possibility remains that, for example, some familial psychosocial factors affected participation and may have introduced bias into the results. In the future, as the use of hybrid closed-loop insulin pumps becomes more widespread, it would be valuable to investigate their potential effects on parental stress levels and child well-being within the Finnish population. Some previous studies have not demonstrated statistically significant differences in quality of life between families using hybrid closed-loop systems and those using sensor-augmented pumps.44 However, de Beaufort et al reported significant improvements in parental quality of life and reduced fear of hypoglycemia among parents of very young children using hybrid closed-loop systems.45 Furthermore, given the increasing number of immigrants with T1D in Finland, future research should also explore the well-being of these families. To gain a more comprehensive understanding of the quality of life in families of children with T1D, incorporating annual assessments into routine follow-ups for patients and their parents would be beneficial.In conclusion, although the average glycemic control in our study population was suboptimal, the children’s well-being was good, and parents’ stress levels were low. In particular, more stable glycemic control, as measured by TIR, appears to be related to the well-being of children and adolescents with T1D. Our findings highlight the important roles of family resources and dynamics, socioeconomic stability, and well-being in managing pediatric diabetes. Regular assessments, as an integral part of follow-up visits, are essential to identify patients and their parents whose well-being may be compromised.SP210.1136/bmjdrc-2025-005248.supp2Supplementary data