Routine Behavioral and Mental Health Screening in Young Children with Type 1 Diabetes Mellitus

Abstract

Background

The American Diabetes Association and International Society for Pediatric and Adolescent Diabetes recommend that providers of diabetes care receive training in the recognition of psychosocial problems related to diabetes.

Objective

To report the results of routine behavioral/mental health screening for children with type 1 diabetes mellitus (T1D) seen in a multidisciplinary pediatric diabetes program.

Subjects and Methods

This was a cross-sectional study of children with T1D ages 4 – 11 years, who underwent behavioral/mental health screening as part of their diabetes care. Screening utilized the Strengths and Difficulties Questionnaire (SDQ) Parent Proxy Version, and scores were reviewed by a social worker. SDQ scale and total difficulties scores were compared by gender, visit type, age, T1D duration and HbA1c. Scores were also compared to age-appropriate normative data for children in the United States (US).

Results

SDQ Parent Proxy Version total difficulties and scale scores did not differ by patient or visit characteristics. Compared with normative data for US children, a greater proportion of children with T1D ages 4 – 7 and 8 – 10 years had borderline/abnormal scores on the emotional symptoms scale (p = 0.01 and p = 0.03, respectively), suggesting risk for psychological disorders, such as anxiety and depression.

Conclusions

Our findings suggest that children less than 11 years old with T1D may have greater emotional symptoms as compared to their age-matched healthy peers. Pediatric diabetes care providers, with access to mental health services, should consider incorporating routine behavioral/mental health screening for children less than 12 years old in their practice.

INTRODUCTION

The American Diabetes Association and International Society for Pediatric and Adolescent Diabetes recommend that the diabetes care team receive training in the recognition, identification, and provision of information and counseling on psychosocial problems related to diabetes^1,2. Moreover, the identification of psychosocial adjustment problems, depression, eating disorders, and other psychiatric disorders should be conducted at planned intervals by mental health professionals². Assessment for these difficulties at diagnosis has been shown to identify patients at risk for early emerging complications and nonadherence³ Although important for all youth with type 1 diabetes mellitus (T1D), ongoing assessments are particularly important in young people not achieving treatment goals or who have chronically poor metabolic control (high hemoglobin A1c [A1c] levels, recurrent diabetic ketoacidosis)^1,2, as the presence of psychological disorders in adolescents with T1D has been linked to poor diabetes self-care^3-9.

In October 2009, the Diabetes Program at Boston Children’s Hospital (BCH) undertook a quality improvement project to evaluate the feasibility of implementing routine behavioral and mental health screening during visits for adolescents with T1D. In February 2011, screening was expanded to include children ages 4 – 11 years old using the Strengths and Difficulties Questionnaire (SDQ) Parent Proxy Version¹⁰. Although studies have examined the prevalence of behavioral and emotional difficulties in adolescents with T1D^9,11-14, there have been few reports about potential mental health concerns of children less than 12 years old⁷. This report of routine behavioral and mental health screening of children 4 – 11 years old with T1D adds to the limited data for this age group.

METHODS

This was a cross-sectional study of children with T1D ages 4 – 11 years old receiving care in the Diabetes Program at the Longwood Medical Campus of BCH between February 2011 and October 2012. Screening occurred in the following settings: (i) social work (SW) visits for children with newly diagnosed T1D (ii) individual or group SW visits per protocol for children with established T1D transitioning to continuous subcutaneous insulin infusion (CSII) therapy; (iii) SW visits by referral for psychosocial concerns; and (iv) during a routine diabetes care visit with a medical provider. Prior to each visit, a parent or adult caregiver completed the SDQ Parent Proxy Version, which was available in English or Spanish¹⁰. The social worker reviewed the questionnaire for all screening settings.

The SDQ Parent Proxy Version is a brief, 25 item, behavioral screening questionnaire that assesses positive and negative attributes across 5 scales: 1) emotional symptoms (5 items); 2) conduct problems (5 items); 3) hyperactivity-inattention (5 items); 4) peer relationship problems (5 items); and 5) pro-social behavior (5 items)¹⁰. A total difficulties score is calculated by summing the scores of the emotional symptoms, conduct problems, hyperactivity-inattention, and the peer problems scales¹⁰. Scales are scored from 0 to 10, and total difficulties score ranges from 0 to 40. For all scales, except prosocial behavior, higher scores are associated with worsening symptoms. Scale scores could be prorated if at least 3 items were completed. Computer-based, automated scoring of the SDQ was performed by the social worker in the Diabetes Program using an online tool¹⁵. The BCH Institutional Review Board approved this study.

Patient characteristics included age, gender, primary language, duration of T1D, and glycosylated hemoglobin A1c (A1c), either from the date of screening or if not performed concurrently, the most proximal value prior to screening. A1c values were obtained either by point of care testing [DCA Vantage Analyzer (Siemens Healthcare Diagnostics)] or by whole blood laboratory testing (Turbidimetric Inhibition Amino Assay). Visit type was characterized as (i) SW visit for newly diagnosed T1D; (ii) individual or group SW visit for CSII therapy transition; (iii) SW visit by referral for psychosocial concerns; or (iv) routine visit with medical provider with SW review. The outcome measures were SDQ Parent Proxy Version scale scores and total difficulties score. Due to small cell sizes, the borderline and abnormal results were collapsed into a single category.

Data analyses were performed with SAS version 9.3 (SAS Institute, Inc, Cary, NC). Descriptive statistics were presented as proportions, medians and interquartile ranges, or means and SD, as appropriate. For bivariate analyses, chi-square test and Spearman rank-order correlation were used to assess the relationship between SDQ Parent Proxy Version scores and patient and visit characteristics. The SDQ Parent Proxy Version Scores were also compared to age-appropriate normative data for children in the US¹⁶. To perform these analyses, patient age was categorized according to the age categories in the normative sample (4 – 7, 8 – 10, and 11 -14 years old). Although not reported here, youth in our sample who were 12 years old or greater were screened using a different instrument. The proportion of children in our clinical sample with borderline/abnormal scores on each scale was compared to the proportion of children with borderline/abnormal scores in the normative sample using the binomial test. Statistical significance was defined as a P-value less than 0.05.

RESULTS

From February 2011 through October 2012, 130 children were eligible for screening. Descriptive statistics are presented in Table 1. Only 9 (7%) of patients did not complete screening, and no parent or adult caregiver refused screening. Mean age was 8.2±2.2 years, and median duration of diabetes was less than 1 year [interquartile range (IQR) 0, 1]. More than half the patients were female (57%). Half of the children (50%) were newly diagnosed with T1D. HbA1c values were unavailable in three children. Mean HbA1c was 8.6±1.7% (70.5±18.6 mmol/mol).

Table 1.

Characteristics of patients eligible for screening (N=130)

Variable	N (%)
Age (years)a	8.2 (2.2)
Duration of DM (years)b	0 (0, 1)
Sex (female)	75 (57)
Hemoglobin A1c (%) [(mmol/mol)]a	8.6 (1.7) [70.5 (18.6)]
Primary Language
English	125 (96)
Spanish	3 (2)
Other	2 (2)
Visit Type
SW* visit for newly diagnosed type 1 DM	65 (50)
Individual or group SW visit for CSII† therapy   transition	34 (26)
SW visit for clinical concerns	18 (14)
Routine visit with medical provider with SW   review	13 (10)
Completed Screening	121 (93)
Reason not completed (N=9)
Refused	0 (0)
Comprehension/developmental issue	1 (1)
Language issue	3 (2)
Not given form by receptionist	2 (2)
Other reason	1 (1)
Unknown	2 (2)

^aMean (SD)

^bMedian (25^th, 75^th percentile interquartile range)

^*social work

^†continuous subcutaneous insulin infusion therapy

In one child, only the SDQ Parent Proxy Version total difficulties score was available. For the sample, mean (SD) SDQ Parent Proxy scores were: total difficulties, 6.8 (5.5); emotional symptoms, 2.3 (2.1); conduct problems, 1.0 (1.5); hyperactivity-inattention symptoms, 3.0 (2.3); peer relationship problems, 1.4 (2.5); and prosocial behaviors, 8.8 (2.3). Aggregate SDQ Parent Proxy Version total difficulties and scale scores were not correlated with HbA1c or duration of T1D and did not differ by gender, visit type, or age category (data not shown). There were also no differences in scores between those being evaluated for a new diagnosis of T1D vs. other reasons. Compared with normative data for US children, a greater proportion of children with T1D ages 4 – 7 and 8 – 10 years old had borderline or abnormal scores on the emotional symptoms scale (p = 0.01 and p = 0.03, respectively) (Table 2). The distribution of scores of the children with T1D did not differ from the normative sample in terms of total difficulties, conduct problems, hyperactivity-inattention symptoms, peer relationship problems, or prosocial behaviors.

Table 2.

SDQ Parent Proxy Version Scores: Comparison to US Normative Data

Age Category	SDQ Scale Category	Clinical Sample		US Population Norm	P value
		N	%	%
Age 4 – 7 years	Total difficulties	38
	Normal (0 – 13)		94.7	87.7	0.28
	Borderline &  Abnormal (≥ 14)		5.3	12.3
	Emotional Symptoms	38
	Normal ( 0 – 3)		73.7	89.0	0.01
	Borderline &  abnormal (≥ 4)		26.3	11.0
Age 8 – 10 years	Total Difficulties	63
	Normal (0 – 13)		85.7	87.1	0.85
	Borderline &  Abnormal (≥ 14)		14.3	12.9
	Emotional Symptoms	62
	Normal ( 0 – 3)		75.8	86.7	0.03
	Borderline &  Abnormal (≥ 4)		24.2	13.3
Age 11 years*	Total Difficulties	20
	Normal (0 – 13)		90.0	85.7	0.88
	Borderline &  Abnormal (≥ 14)		10.0	14.3
	Emotional Symptoms	20
	Normal ( 0 – 3)		80.0	83.9	0.81
	Borderline &  Abnormal (≥ 4)		20.0	16.1

^*The normative US sample included children 11 – 14 years of age.

DISCUSSION

This is a report of the SDQ Parent Proxy Version scores for children ages 4 – 11 years old with T1D, which to our knowledge is the first of its kind in the literature. The aggregated SDQ Parent Proxy Version scores did not differ by age, gender, or visit type and were not correlated with HbA1c or T1D duration. Although there were no detectible differences in the study sample for most SDQ scales as compared to the population norms, a greater proportion of children ages 4 – 7 and 8 – 10 years old had higher scores on the emotional symptoms scale, which screens for symptoms of emotional disorders, such as anxiety and depression. Even though clinically the manifestation of emotional symptoms might be seen as an acute reaction to diagnosis of T1D, in this study, emotional symptoms scores did not differ by duration of diagnosis or visit type. These finding suggests that significant emotional difficulties in youth with T1D may begin to manifest in early childhood even outside the context of a new diagnosis of diabetes. Given the challenges associated with psychological disorders and the management of T1D in adolescence, it is possible that earlier detection and treatment of such psychosocial problems could prevent emotional difficulties in adolescence and, ultimately result in improved glycemic control.

To our knowledge, there are no published studies directly comparing youth with T1D to youth with other chronic diseases; however, similar studies using the SDQ Parent Proxy Version, have shown that, elevated behavioral and emotional disorders in other chronic diseases including chronic kidney disease¹⁷, asthma, neurological disorders, and other chronic illnesses¹⁸, had more emotional and behavioral problems than children without chronic illness¹⁸. However, differences between the studies, including country of origin and subject ages, make comparison across studies difficult.

There are several limitations that should be considered when interpreting the results of this study. First, the results reported were from a screening instrument, and we do not have follow-up data on the children to know whether any with positive screens have been formally diagnosed with an emotional disorder. Second, we do not yet have longitudinal data to determine whether the emotional symptoms persist. Third, insufficient power from a small sample, due to the use of a different screening instrument for youth ages 12 and older, may have limited our ability to detect a difference in the emotional symptoms scores for the older children as compared to the normative sample. Fourth, distinguishing emotional symptoms from hyperglycemic symptoms (e.g., fatigue, concentration difficulties, restlessness, changes in appetite, and irritability) may be difficult for some parents and speaks to the need for ongoing screening, and formal evaluation when indicated. Lastly, in this study, the majority of screening occurred during visits with a social worker. Given the often limited availability of mental health services in the US health care system,^19,20 this may limit the generalizability of the findings.

The study demonstrates the feasibility of incorporating behavioral and mental health screening for children less than 12 years of age as a component of diabetes care in a multidisciplinary pediatric diabetes program. Behavioral and mental health screening in this age group was well accepted by parents or adult caregivers. Although the sample sizes were relatively small, these findings suggest that children less than 12 years old might have more emotional symptoms as compared to the normal population. Further studies, with larger sample sizes, should be performed to validate our findings. Additional, longitudinal studies should also be conducted to determine whether emotional symptoms manifesting in childhood persist into adolescence and adulthood. If symptoms do persist, identifying and treating emotional difficulties in early childhood could potentially result in decreased morbidity in adolescence and adulthood and improve diabetes outcomes. Based on these results, diabetes care providers, who have access to mental health services, should consider incorporating routine behavioral and mental health screening for children less than 12 years old in their practice.

Abbreviations

T1D

type 1 diabetes mellitus

US

United States

SDQ

Strengths and Difficulties Questionnaire

A1c

hemoglobin A1c

BCH

Boston Children’s Hospital

SW

social work

CSII

continuous subcutaneous insulin infusion

IQR

interquartile range