PEDIATRICS Vol. 115 No. 5 May 2005, pp. 1315-1319 (doi:10.1542/peds.2004-1717)
Depressive Symptoms Predict Hospitalization for Adolescents With Type 1 Diabetes Mellitus
* Department of Psychiatry
Department of Pediatrics, University of Texas Southwestern Medical Center, Dallas, Texas
Department of Social Work, Children's Medical Center, Dallas, Texas
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Objective. To examine the role of self-reported depressive symptoms in predicting hospitalization for complications of diabetes mellitus over a period of up to 2 years.
Study Design. Two hundred thirty-one adolescent outpatients (age range: 11–18 years) with type 1 diabetes completed the Center for Epidemiological Studies Depression Scale, a self-report measure of depressive symptoms. Glycosylated hemoglobin levels were also assessed, to account for this known predictor of hospitalization. With survival analysis methods, hospitalizations for medical complications that occurred up to 2 years after this assessment were recorded.
Results. After controlling for age, gender, socioeconomic status, and glycosylated hemoglobin levels at baseline, the odds ratio for prediction offered by Center for Epidemiological Studies Depression Scale scores above the cutoff point (12 for boys and 22 for girls) was 2.58 (95% confidence interval: 1.12–5.98).
Conclusions. Young people with type 1 diabetes who show high levels of depressive symptoms are at increased risk for hospitalization for disease complications. Interventions aimed at improving their depressive symptoms may result in positive health outcomes, as well as improved quality of life.
Key Words: depression • type 1 diabetes mellitus • hospitalization • adolescent
Abbreviations: MDD, major depressive disorder • CES-D, Center for Epidemiological Studies Depression Scale • HbA1c, hemoglobin A1c
Depressive disorders occur at high rates among individuals with diabetes mellitus,1 with controlled studies reporting that 9 to 27% of diabetic patients suffer from major depressive disorder (MDD) at any single point in time.2,3 Aside from the negative effects on the quality of life of these patients, depression affects the management and complications of the disease. Adult depressed patients reported more symptoms of diabetes4,5 and showed worse metabolic control6 than did diabetic patients who were not depressed. The severity of depressive symptoms was found to be correlated with glycosylated hemoglobin levels.6 Among adults, depressive symptoms were independent predictors of heart disease.7 The presence of depression was one of the best predictors of hospitalization among adult patients with diabetes.8
Most studies of the link between depression and health outcomes among patients with diabetes have been conducted with adults. There is some evidence, however, that similar patterns are present among children and adolescents with type 1 diabetes. A recent review concluded that base rates of depression are twofold higher among children and threefold higher among adolescents with diabetes, compared with the corresponding age groups in the general population.9 During the first 10 years after the diagnosis of diabetes and by the average age of 20 years, 48% of youths monitored prospectively had developed a psychiatric disorder and 28% had experienced at least 1 episode of MDD.10 Kovacs and colleagues11–13 showed associations between behavioral disturbance and health outcomes among adolescents and young adults. By 20 years of age, retinopathy was predicted by the time spent in depressive episodes,11 independent of the duration of illness and glycemic control during the course of the illness. A longitudinal study that monitored newly diagnosed youths for up to 14 years found that behavioral problems were risk factors for rehospitalization within 2 years after diagnosis, as well as multiple rehospitalizations up to 14 years later.12,13 It is not known whether depressive symptoms predict future hospitalizations among youths who have had a diagnosis of type 1 diabetes for some time.
Therefore, diagnosable psychiatric illness appears to be associated with health outcomes among both adults and youths. However, although full syndrome manifestations of depressive disorders increase in adolescence, they are still relatively rare; 
3% of adolescents meet criteria for the diagnosis of MDD at any one time.14 There is evidence that "subthreshold" manifestations among young people may be just as pernicious and disabling as syndromic depression.15 Such individuals may, for example, endorse numerous core symptoms (such as persistent feelings of sadness or irritability and loss of interest in previously pleasurable activities) and secondary symptoms (such as pessimism about the future, sleep disturbances, and poor concentration) of depression. However, the combination or range of symptoms experienced may not be broad or persistent enough to meet the criteria for this diagnosis. These subthreshold manifestations of depressive symptoms are more common than the full syndrome, but to date there are no published studies on relationships between high levels of depressive symptoms (regardless of categorical diagnosis) and hospitalization among youths with diabetes.
The goal of this study was to examine the prospective prediction offered by baseline depressive symptoms for future hospitalizations during a period of up to 24 months. We hypothesized that high levels of self-reported symptoms of depression would predict hospitalization among adolescents with diabetes.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Subjects
These data were obtained in the course of an ongoing longitudinal study on treatment adherence among adolescents with diabetes in the Children's Endocrinology Center at Children's Medical Center of Dallas. Inclusion criteria for the current analyses were age of 11 through 18 years, diagnosis of type 1 diabetes, and availability of an English-speaking primary caretaker (usually the mother). Exclusion criteria were coexisting primary medical diseases, eg, chronic active hepatitis or severe cardiac, renal, or hematologic disease. Two hundred thirty-one adolescents participated in this study (see Table 1 for demographic information).
|
Recruitment was conducted during clinic visits in a period of 18 months. The participation rate among those who met the criteria was 90%. The primary reason stated for nonparticipation was inconvenience because of the time required to administer the study measures; no other characteristics of refusers could be recorded without informed consent. Participation was voluntary, and confidentiality of information (even from the treatment team) was ensured. Written informed consent was obtained from the parent and assent was obtained from the adolescent. This study was approved by the institutional review board of the University of Texas Southwestern Medical Center at Dallas.
Demographic Measures
The participant's age at diagnosis was obtained from the medical records. Parents' reports of paternal occupation were analyzed as an approximate measure of socioeconomic status, according to the following categories: unemployed, blue-collar worker (unskilled or manual labor, eg, construction, domestic help, or landscape crew), white-collar worker (eg, clerical staff or sales), and professional (eg, physician, university professor, or lawyer).
Depressive Symptoms
Participants completed the Center for Epidemiological Studies Depression Scale (CES-D),16 a 20-item, self-report, rating scale developed to measure current levels of depressive symptoms in community samples. Participants received the self-administered questionnaire from a research assistant during a clinic visit. The research assistant remained available to answer questions and collect the forms after completion.
Each participant was asked to indicate on a Likert scale (0, rarely to none of the time, to 3, most or all of the time) the extent to which each item was true of him or her in the past week. Examples of items are "I felt sad," "I had trouble keeping my mind on what I was doing," and "I thought my life had been a failure." Summed scores on the 20 items ranged from 0 to 60, with higher scores indicating more symptoms. The CES-D was originally developed for adults, but it has been validated with children and younger and older adolescents.17,18 In the validating study, which included youths 
11 years of age,18 the self-report scale was followed by a diagnostic interview, to obtain information about the relationship between the scale scores and diagnoses of MDD. Optimal cutoff points, ie, scores at or above which the likelihood of false-negative and false-positive results is lowest, were reported as 12 for boys and 22 for girls. With these cutoff points, sensitivity and specificity from receiver operating characteristic curves were 0.85 and 0.49 for boys and 0.83 and 0.77 for girls, respectively. For the present study, the scores were grouped, relative to these cutoff points, as low (ie, <12 for boys and <22 for girls) or high (ie, 12 for boys and 22 for girls). Although some of the youths who scored above the cutoff point for the scale would be expected to meet criteria for a diagnosis of MDD, the limited specificity of the CES-D (particularly among boys) indicates that many subjects with high scores would fall into the subthreshold category.
Glycemic Control
Hemoglobin A1c (HbA1c) (the main form of glycohemoglobin) levels are routinely measured at clinic visits. The level measured at the visit at which participants were enrolled in the study and completed the CES-D was obtained from the patients' medical records and used as a control variable for predicting hospitalization.
Hospitalization
All inpatient admissions secondary to complications of diabetes were obtained from the hospital database for the study participants for 2 years, from January 2002 through December 2003. International Classification of Diseases, 9th Revision, codes used for the search included 250.03 (poorly controlled insulin-dependent diabetes), 250.11 (diabetic ketoacidosis with unstated control), and 250.13 (diabetic ketoacidosis with poor control). The first subject was enrolled in December 2001 and the last in May 2003. Only admissions that occurred after enrollment were considered, and the time from CES-D completion to first hospitalization was recorded. Therefore, there was variability in the time that each participant was monitored for the study; those enrolled in the early weeks of the study were monitored for a full 2 years, whereas those enrolled at the end of the study were monitored for only 7 months. Statistical strategies specific for such a design were used to ensure that the varying periods from enrollment to the end of the study did not confound the findings (see below).
Analysis
SPSS software (version 11.5; SPSS, Chicago, IL) was used to analyze data. Descriptive information was compiled for demographic measures and the variables in the study. Youths with and without hospitalizations during the study period, as well as those with results below and above the CES-D cutoff points, were compared with respect to these variables with independent-sample t tests.
In earlier prospective studies of youths with diabetes,12,13 age at diagnosis and socioeconomic background predicted rehospitalization. For this reason, these 2 variables were controlled in the regression analyses reported in this article. In addition, because gender differences in diabetes management have been documented,19 gender also served as a covariate in the analysis. Controlling for gender in multivariate analyses served a different purpose than adjusting for gender differences in the cutoff point. The former was included to account for variance in hospitalization outcomes attributable to gender. The latter provided a more-accurate, gender-specific threshold for classification of high versus low levels of symptoms.
Survival analysis procedures (Cox regression analysis) were used to examine the relationship between depressive symptoms and hospitalization, after controlling for gender, age at onset, socioeconomic status, and HbA1c level. These procedures were developed for designs that examine the time between entry into the study and a subsequent event, in this case hospitalization. In time-limited and continuous-enrollment studies, there is variation in the period of time between study entry and the end of the study, and it is possible that additional participants may go on to experience the event (hospitalization) later. More commonly used analytic strategies, such as logistic regression, are not appropriate for such designs, because they cannot adjust for the confounding presented by varying times of entry into the study. In contrast, survival analysis procedures account for the fact that, with longer follow-up periods, there is greater likelihood of rehospitalization.
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Descriptive Data
Table 1 provides information on the demographic characteristics of the study patients. The mean ± SD scores on the CES-D were 14.30 ± 9.57 for girls and 15.05 ± 9.55 for boys. Thirty-three percent of patients had CES-D scores above the cutoff points. This proportion is consistent with results of studies of adult diabetic patients with similar self-report symptom scales.2,3,20 Table 2 presents the demographic characteristics of participants with results above and below the CES-D cutoff points.
|
Survival Analysis
Twenty-six patients (11% of the total group) had at least 1 hospitalization between entry into the study and the end of the study period. Cox regression analysis allowed for inclusion of multiple prognostic factors (Table 3). After controlling for gender, age at onset, socioeconomic status, and HbA1c level, adolescent patients whose scores on a scale of depressive symptoms fell above the CES-D cutoff points at entry into the study (cutoff points of 12 for boys and 22 for girls) were >2.5 times more likely to be hospitalized in the remainder of the study period than were their counterparts whose depressive symptoms fell below the cutoff points. Figure 1 plots the hazard function, showing the cumulative rates of hospitalization for the groups below and above the CES-D cutoff points.
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Depressive symptoms obtained at entry into the study significantly predicted the likelihood of subsequent hospitalization for up to 24 months, after controlling for gender, age, duration of illness, and HbA1c level. These findings are consistent with reports from studies with adults. To our knowledge, however, this is the first study to report the relationship between self-reported symptoms of depression and hospitalization among adolescents with diabetes.
These findings have implications for the treatment of youths with diabetes. Given the high frequency of depressive symptoms, routine assessment and interventions to treat dysphoric mood appear particularly important for the population of adolescents with diabetes. High levels of depressive symptoms heralded later syndromic difficulties among young people.15 Depressive episodes lasted longer among youths with diabetes than among psychiatric control subjects who were otherwise medically well.10 Active treatment of early depression may therefore significantly improve the quality of life for these youths.
A second compelling reason to assess and to treat mood difficulties among these youths is emphasized by results from the central analyses of this study. Treatment of mood difficulties among youths with diabetes, even when they do not reach full criteria for MDD, might prevent medical hospitalization. Cognitive-behavioral therapy has been shown to be an effective treatment for depressed adults with diabetes,21 with effects not only on mood but also on metabolic control (possibly through effects on adherence). To date, however, there have been no controlled trials assessing the effects of depression treatment on disease management and hospitalization status among youths with diabetes. In addition to formal assessment and treatment of adolescents who demonstrate subclinical signs of depression, clinicians and parents may benefit from information about strategies to prevent and manage depression on an outpatient basis. Mood among youths with diabetes has been linked to health-related beliefs such as self-efficacy,22,23 ie, the confidence patients have that they can manage their disease adequately. Such beliefs have been found to play an important role in promoting positive health behaviors. Education about strategies to enhance self-efficacy and other cognitions that are protective against depression would allow clinicians to integrate these interventions into the medical care of their adolescent patients. Younger children may benefit from family counseling and behavioral interventions. Parents can play a major role in preventing depressive processes by providing support and guidance and encouraging positive coping strategies among youths with diabetes.
Kovacs and colleagues12,13 monitored 92 children with repeated surveys from diagnosis up to 14 years later and examined predictors for early hospitalization and for readmissions. They assessed internalizing and externalizing symptoms, each based on 12 items rated by a clinician. Externalizing symptoms (such as temper tantrums, stealing, anger, and disobedience) predicted subsequent hospitalization, whereas internalizing symptoms (such as anxiety and depressed mood) did not.
The difference between the results reported by Kovacs and colleagues12,13 and those of the present study might be attributable to differences in methods. Kovacs and colleagues12,13 predicted multiple hospitalizations after diagnosis during a 14-year period, whereas the current study examined much shorter-term outcomes. Although the grouping of internalizing symptoms overlaps with depressive symptoms, the former construct is broader and includes anxiety; this might have decreased the specificity of internalizing symptoms as predictors of hospitalization. Moreover, symptoms in the study by Kovacs and colleagues12,13 were rated by a clinician and not by self-report. Although clinicians' assessments have been considered the standard method for psychopathologic assessments,24 our findings indicate that adolescents' self-report of depressive symptoms is a robust predictor with respect to health outcomes. Indeed, other investigators noted that adolescents may be more accurate in describing internal states, whereas observers may be more reliable recorders of externalizing behaviors.24 The fact that a self-completed form provides adequate information to predict the important health outcome of hospitalization makes such assessments a cost-effective step in determining the need for additional evaluation and treatment.
The mechanisms through which depressive symptoms are related to hospitalization cannot be determined with the methods of our study. Poor adherence to the medical regimen among depressed youths is one obvious explanation. Moreover, psychologic stress, particularly depression, may be accompanied by increased levels of counter-regulatory hormones (cortisol, epinephrine, glucagon, and growth hormone) that antagonize insulin action and thus exacerbate poor metabolic control of diabetes.20 Depression does not appear to act entirely through decreased metabolic control, however, because the relationship between depression and hospitalization persists even after controlling for baseline HbA1c values. Finally, physicians may be more likely to hospitalize adolescents with depression if these patients have obvious difficulties coping with their illness.
Our study had several limitations. All patients were recruited from a single center. Although the number of refusals was proportionately quite small, we cannot rule out a bias; those who refused might have been in a more distressed and/or poorly compliant group that would avoid a study on mood and treatment adherence. The follow-up period was relatively short. Measures of depressive symptoms were based on self-report alone, which might be less reliable for younger adolescents. The age range of 11 to 18 years was quite wide, and developmental differences might have been present with respect to the variables of interest. Most patients have ongoing relationships with the health care staff at our institution, and it is rare that hospitalizations take place outside the system; nevertheless, it is possible that some participants sought care at other hospitals and so were misclassified for the analyses presented here. We did not gather systematic information about previous and current psychiatric treatment of participants. Future studies should examine potential mediators from depressive symptoms to hospitalization, to elucidate the mechanisms through which depressive symptoms and health outcomes are associated.
|
ACKNOWLEDGMENTS |
|---|
This study was funded in part by a grant to S.M.S. from the Timberlawn Psychiatric Research Foundation.
|
FOOTNOTES |
|---|
Accepted Sep 28, 2004.
Reprint requests to (S.M.S.) Department of Psychiatry, University of Texas Southwestern Medical Center, 5323 Harry Hines Blvd, Dallas, TX 75390-8589. E-mail: sunita.stewart{at}utsouthwestern.edu
No conflict of interest declared.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
- Dantzer C, Swendsen J, Maurice-Tison S, Salamon R. Anxiety and depression in juvenile diabetes: a critical review. Clin Psychol Rev. 2003;23 :787 –800[CrossRef][Web of Science][Medline]
- Gavard JA, Lustman PJ, Clouse RE. Prevalence of depression in adults with diabetes: an epidemiological evaluation. Diabetes Care. 1993;16 :1167 –1178[Abstract]
- Anderson RJ, Freedland KE, Clouse RE, Lustman PJ. The prevalence of comorbid depression in adults with diabetes: a meta-analysis.
Diabetes Care. 2001;24
:1069
–1078
[Abstract/Free Full Text] - Lustman PJ, Clouse RE, Carney RM. Depression and the reporting of diabetes symptoms. Int J Psychiatry Med. 1988;18 :295 –303[Web of Science][Medline]
- Ciechanowski P, Katon W, Russo J. The relationship of depressive symptoms to symptom reporting, self care and glucose control in diabetes.
Diabetes Care. 2002;25
:731
–736
[Abstract/Free Full Text] - Lustman PJ, Anderson RJ, Freedland KE, de Groot M, Carney RM, Clouse RE. Depression and poor glycemic control: a meta-analytic review of the literature. Diabetes Care. 2002;23 :934 –942
- Forrest KYZ, Becker DJ, Kuller LH, Wolfson SK, Orchard TJ. Are predictors of coronary heart disease and lower extremity arterial disease in type 1 diabetes the same? A prospective study. Atherosclerosis. 2000;148 :159 –169[CrossRef][Web of Science][Medline]
- Rosenthal MJ, Fajardo M, Gilmore S, Morley JE, Naliboff BD. Hospitalization and mortality of diabetes in older adults: a 3-year prospective study. Diabetes Care. 1998;21 :231 –235[Abstract]
- Grey M, Whittemore R, Tamborlane W. Depression in type 1 diabetes in children: natural history and correlates. J Psychosomat Res. 2002;53 :907 –911[CrossRef][Web of Science][Medline]
- Kovacs M, Obrosky DS, Goldston D, Bonar LK. Psychiatric disorders in youths with IDDM: rates and risk factors. Diabetes Care. 1997;20 :36 –44[Abstract]
- Kovacs M, Drash A, Mukerji P, Iyengar S. Biomedical and psychiatric risk factors for retinopathy among children with IDDM. Diabetes Care. 1995;18 :1592 –1599[Abstract]
- Kovacs M, Charron-Prochownik D, Obrosky DS. A longitudinal study of biomedical and psychosocial predictors of multiple hospitalizations among young people with insulin-dependent diabetes mellitus. Diabetes Med. 1995;12 :142 –148[Web of Science][Medline]
- Charron-Prochownik D, Kovacs M, Obrosky DS, Stiffler L. Biomedical and psychosocial predictors of early rehospitalization among children with insulin-dependent diabetes mellitus: a longitudinal study. Diabetes Med. 1994;11 :372 –377[Web of Science][Medline]
- Lewinsohn PM, Hops H, Roberts RE, Seeley J, Andrews J. Adolescent psychopathology, I: prevalence and incidence of depression and other DSM-III-R disorders in high school students. J Abnorm Psychol. 1993;102 :133 –144[CrossRef][Web of Science][Medline]
- Lewinsohn PM, Solomon A, Seeley J, Zeiss A. Clinical implications of "subthreshold" depressive symptoms. J Abnorm Psychol. 2000;109 :345 –351[CrossRef][Web of Science][Medline]
- Radloff LS. The CES-D scale: a self-report depression scale for research in the general population. Appl Psychol Meas. 1977;1 :385 –401[CrossRef]
- Dierker LC, Albano AM, Clarke GJ, et al. Screening for anxiety and depression in early adolescence. J Am Acad Child Adolesc Psychiatry. 2001;40 :929 –936[CrossRef][Web of Science][Medline]
- Garrison CZ, Addy CL, Jackson KL, et al. The CES-D as a screen for depression and other psychiatric disorders in adolescents. J Am Acad Child Adolesc Psychiatry. 1991;30 :646 –691
- La Greca AM, Swales T, Klemp S, Madigan S, Skyler J. Adolescents with diabetes: gender differences in psychosocial functioning and glycemic control. Child Health Care. 1995;24 :61 –78
- Musselman DL, Betan E, Larsen H, Phillips LS. Relationship of depression to diabetes types 1 and 2: epidemiology, biology and treatment. Biol Psychiatry. 2003;54 :317 –329[CrossRef][Web of Science][Medline]
- Lustman PJ, Clouse RE. Treatment of depression in diabetes: impact on mood and medical outcome. J Psychosom Res. 2002;53 :917 –924[CrossRef][Web of Science][Medline]
- Stewart SM, Lee PWH, Low LCK, et al. Pathways from emotional adjustment to glycemic control in youth with diabetes in Hong Kong.
J Pediatr Psychol. 2000;25
:393
–402
[Abstract/Free Full Text] - Stewart SM, Lee PWH, Waller DA, et al. A follow-up study of adherence and glycemic control among Hong Kong youths with diabetes.
J Pediatr Psychol. 2003;28
:67
–79
[Abstract/Free Full Text] - McClellan JM, Werry JS. Research psychiatric diagnostic interviews for children and adolescents. J Am Acad Child Adolesc Psychiatry. 2000;39 :19 –27.[CrossRef][Web of Science][Medline]
PEDIATRICS (ISSN 1098-4275). ©2005 by the American Academy of Pediatrics
CiteULike 
Connotea 
Del.icio.us 
Digg 
Facebook 
Reddit 
Technorati 
Twitter What's this?
This article has been cited by other articles:
|
|
 |
|
  D. Eckshtain, D. A Ellis, K. Kolmodin, and S. Naar-King The Effects of Parental Depression and Parenting Practices on Depressive Symptoms and Metabolic Control in Urban Youth with Insulin Dependent Diabetes J. Pediatr. Psychol., August 26, 2009; (2009) jsp068v1. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. S. Jaser, R. Whittemore, J. M. Ambrosino, E. Lindemann, and M. Grey Mediators of Depressive Symptoms in Children with Type 1 Diabetes and their Mothers J. Pediatr. Psychol., June 1, 2008; 33(5): 509 - 519. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. J. Cameron, E. A. Northam, G. R. Ambler, and D. Daneman Routine Psychological Screening in Youth With Type 1 Diabetes and Their Parents: A notion whose time has come? Diabetes Care, October 1, 2007; 30(10): 2716 - 2724. [Full Text] [PDF]
|
|
|
|
|
|
K. K. Hood, S. Huestis, A. Maher, D. Butler, L. Volkening, and L. M.B. Laffel Depressive Symptoms in Children and Adolescents With Type 1 Diabetes: Association with diabetes-specific characteristics. Diabetes Care, June 1, 2006; 29(6): 1389 - 1389. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
