Psychological moderator variables and metabolic control in recent onset type 1 diabetic patients—a two year longitudinal study

Journal

of Psychosomo~ic

Research, Vol. 38, No. 3. pp. 249-258, 1994 Copyright 0 1994 Elsevier Science Ltd Printed in Great Britain. All rights resewed 0022-3999194 $6.00 + .OO

PSYCHOLOGICAL MODERATOR VARIABLES AND METABOLIC CONTROL IN RECENT ONSET TYPE 1 DIABETIC PATIENTS-A TWO YEAR LONGITUDINAL STUDY K. SPIES&*

G. SAcHs,t G. MOSER~, P. PIETSCHMANN~, G. SCHERNTHANER~and R. PRAGER$. (Received 22 June 1993; accepted in revisedform

14 September

1993)

Abstract-The relationships between psychosocial adjustment and subsequent glycaemic control were prospectively examined in forty-three adult patients during the first 2 yr after onset of type 1 diabetes mellitus. Decreasing depression was the single psychosocial parameter that changed over time. No correlations were found between the decrease in HbAlc levels and psychological variables at 8- and 16month follow-ups. Global and specific coping features such as high control attitude, low coping anxiety and low emotional attribution correlated significantly with the decrease in HbAlc levels at the 2-yr followup, whereas stressful life events, depression, stat&trait anxiety did not correlate. In a regression analysis coping explained 22% variance of the 2 yr decrease in HbAlc levels. We conclude that coping is a better predictor for metabolic control than emotional adaptation and life events. Metabolic control might deteriorate with prolonged stage of the disease being a first sign for psychophysiological coping exhaustion. INTRODUCTION METABOLIC control and various psychiatric conditions as well as psychological variables correlated only to a moderate degree in retrospective cross-sectional studies [ 141. Some studies found no correlation between psychological factors and metabolic control [7-lo]. Retrospective evaluations revealed significant relationships between stressful life events and poor metabolic control [ 1, 10-l 21, but failed to demonstrate a definite causal connection. Studies assessing the stability of psychosocial as well as metabolic parameters over time, which would enhance the strength of evidence have not very often been performed [4, 131. The association between anxiety and current HbAlc levels persisted after statistical control for potentially confounding variables, including the previous value of HbAlc which had been taken into account. Despite the stability of HbAlc values over time, anxiety scores did not significantly correlate with follow-up HbAlc [4]. Depressive disorder in diabetic patients and the correlation of depression with poor metabolic control proved to be stable over a follow-up period of 5 yr [ 131. Thus in general, the central issue of psychosomatic research in diabetes mellitus is one of methodological problems. A major shortcoming of many studies is the fact that only the bivariate relationships between HbAlc and each of the different *Institute of Medical Psychology, Vienna University. TDepartment of Psychiatry, Vienna University. IMedical Departments, Vienna University. Address correspondence to: Klaus Spiess, MD, Institute of Medical Psychology, University of Vienna, Severingasse 9, A-1090 Vienna, Austria. This study was supported by Research Grant No P-6190, &terreichischer Fonds zur Fiirderung der Wissenschaften.

249

250

K. SPESSet al.

psychosocial variables were explored. These do not take into account the effects of potentially confounding variables such as duration of the disease, adherence to treatment and previous values of HbAIc. As was recently pointed out, these studies do not deal with the buffering function of coping [14]. It is necessary to use multidimensional psychological measures, since a good outcome of coping measures does not necessarily correlate with good psychological adaptation in diabetics [15]. The usual cross-sectional and correlational design of such studies does not assist in clearly determining the direction of effects or their mechanisms. Obtaining longitudinal data covering more variables is therefore, one of the major goals in demonstrating the direction of temporal relationships. Another shortcoming is that all of these studies included patients with a large range of disease duration; furthermore disease duration also varies considerably within the individual studies. Relationships between psychological disease impact and metabolic control may become too entangled during that long period to permit any causal suggestions. This is why many authors recommend that future analyses examine the relationship between coping soon after onset of the disease. This point in time is considered to be of particular importance for future disease adaptation and adjustment to subsequent diabetes-related behaviours [4,5, 161. In children it has been observed that negative defence mechanisms with a high rate of denial reactions interacting with regimen adherence may be adopted during the initial period of the disease [15, 171.Studies in diabetic children, however, either failed to predict psychosocial adjustment and metabolic control later in life from psychological criteria at onset [ 18, 191or could predict psychological adjustment only but not metabolic control [20]. To the best of our knowledge, no study has been published to date, which evaluates the effect of the initial psychological adjustment phase in adult diabetic patients on subsequent metabolic control. In this study we examined prospectively, with multidimensional psychological measures, the relationship between psychosocial adjustment and glycaemic control during the first 2 yr immediately after onset of type 1 diabetes mellitus in adults. Thus the major aim of the present study was to assess whether psychosocial variables identified in retrospective studies as influencing metabolic control at time of onset can predict future metabolic adaptation.

METHOD Subject recruitmem None of the patients was aware of a history of diabetes before the following symptoms occured: six had ketoacidosis, twenty-seven showed other major symptoms of diabetes such as polyuria and weight loss, whereas in ten patients hyperglycemia was found to be associated with mild symptoms such as tiredness. The mean duration of diabetes-associated symptoms was not longer than 4.3 (4.2 SD) weeks. Immediately after the detection of hyperglycaemia by their community physicians, patients were admitted to the Department of Endocrinology and Diabetes where the definite diagnosis of IDD, (based on the definition of the National Diabetes Data Group [21] requiring initiation of insulin treatment was made. Forty-three consecutively diagnosed patients (twenty-eight men and fifteen women) were included in the study. Patients were hospitalized for a mean duration of 10 days to observe hypoglycaemic reactions during the first week of insulin treatment and for diabetes education. Insulin doses were adjusted during hospitalization and this was continued at the follow-up consultations in order to achieve optimal metabolic control. All patients had the same long-term treatment consisting of an insulin basis bolus and additional injections depending on the actual values recorded by home blood-glucose monitoring. None of the patients had medical problems other than diabetes which might be important for the evaluation in this study.

Coping and metabolic control in diabetes

251

The department of Endocrinology and Diabetes is one of three major diabetes centres in Vienna. There is no reason to assume that the characteristics of our patients are different from those seen at the other ccntres.

Study design Baseline data for each patient were collected during bedside interviews at the start of the study (within the third to fifth day of hospitalization). At this time all patients were aware of the diagnosis of a severe chronic disease; insulin therapy had been initiated. All patients were interviewed for the first time before they entered the 1 week educational programme on diabetes self-management. All patients gave their informed consent before participating in the study; none of them refused. In all patients measures were taken at the onset of the disease and at S-monthly follow-up intervals during the 2-year study programme. Total insulin dose (U/kg bodyweight) was documented in all patients in order to detect a decrease of insulin requirements during a possible ‘honeymoon’ period. Measures Diabetic control was determined by glycosylated haemoglobin (HbAlc) measured by high pressure liquid chromatography (HPLC method, Biorad Inc). HbAlc levels up to 7.0% without severe hypoglycaemic reactions were accepted as satisfactory metabolic control.

Psychological

assessment

Life events were evaluated for a better baseline comparability by the Junk and Junk questionnaire [22]. This self-rating scale was designed for the evaluation of quantity and upset of life events. It is an extended version of the method for exploring life events developed by Paykel[23]. The scale assesses five categories of seventy stressful events (entrance events, exit/loss events, conflict/undesirable events, changes in social field events and others) the items being quantified by ten degrees, with 0 indicating the lowest and 10 the highest upset intensity caused by the event. The quality of coping with the rated events was then scored by a IOZitem self-rating scale. The items were scored by patients’ self-rating in five steps and summed up in four internal categories: fatalism, anxious coping behaviour, degree of internal and external control, and fatalistic control by the subject; and by two external factors assessing social integration and familial adherence as a coping support. These six categories were summarized to a global coping index that indicates the degree of coping quality, with higher sum scores signifying poorer coping. The reliability and validity of this instrument have been verified [22]. Junk and Junk have shown good correlation (0.80) between the internal control scale of their instrument and the internal control scale of Levenson’s IPC scale [24]; whereas the external control scale of the instrument shows a good correlation with the external control (powerless) scale, and the fatalism scale with the C scale (external fatalistic control, chance = 0.92). Physical and mental symptoms were evaluated using the Zerssen symptoms check list [25]. Depression was rated by the Beck depression inventory (BDI, [26]) and by DSM III R Interview [27]. Anxiety was determined by the forty-item Spielberger state and trait anxiety inventory (STAI, [28]). Attributional beliefs were measured by the factor analysed scale of Ahrens and Elsner [29] with thirty questions rated in five steps. This scale is similar to Lazarus’ two different coping strategies (problemoriented and emotion-oriented, [30]). The Ahrens scale was specifically adapted to diabetes treatment. Thus a high score of internal attribution indicates an emotional view of origin, whereas a high score of external atribution indicates outside responsibilities and physician and treatment-oriented actions. Denial was measured by an interview which was derived from the Hackett Denial Scale [31] and specifically adapted to diabetes. Each item was read out and discussed with the patient, then scaled in &3 steps. According to Hackett’s scale the first area of the scale assesses diabetes-specific denial, the second estimates whether patients tend to neglect danger, to shift the threat to other objects or to project their fear. The third area estimates the amount of denial constituting a major personality pattern. When used in a preliminary investigation in diabetics this adapted scale showed satisfactory external validity and test-retest reliability over a l-week period (0.85) (after eliminating items with an internal consistency of less than 0.50 from the original scale, so that finally Cronbachs’ alpha could be set at 0.78). Cut-off points were t&10 for minor deniers, II-25 for partial deniers, and 2540 for major deniers.

Data analysis Data were screened was used to determine

with the Kolmogorov-Smirnov test for normal distribution. Spearman correlation the relationship between the psychosocial variables and metabolic control (HbAlc).

252

K.

SPIESS

et

d.

Multiple regression analyses were performed to investigate whether the psychosocial data at disease onset can predict subsequent metabolic control. Analysis of variance (matched samples) was performed to evaluate the interaction between change of questionnaire data (factor time) and good vs poor metabolic control (factor group). Statistical consultation was provided by the Institute of Statistics and Documentation of the Medical School of the University of Vienna. A SPSS programme packet was used. Data are given as mean with standard deviation.

RESULTS

Metabolic

and psychosocial

characteristics

of the sample

At the start of the study our patients were in a state of poor metabolic control, as evidenced by glycosylated haemoglobin values (x = 10.12; SD = 2.65). Mean daily blood glucose levels after 1 week of insulin treatment were 141.43 mg% (SD = 49.12). The age range of the patients was 18-31 yr with a mean of 25.3 (6.3 SD). The population was predominantly urban (N = 34); three of the patients were unemployed and twelve attended school. The social ranking of the study group was lower/ working class (N = 14) middle class (N = 17) and upper middle class (N = 12). None of the patients met the criteria for the presence of a major affective depressive disorder according to DSM-III-R. Beck depression scales showed a depressive range (x = 3 1.60; SD = 3.49). Spielberger scales indicated values similar to healthy controls. Also the mean number of life events, coping strategies and support scores were not significantly different from their respective healthy standardization samples. Metabolic

control over time

All patients achieved good metabolic control within 8 months (HbAlc x= 7.04; 1.24) and this was maintained during the 2 years of out-patient follow up (I= 7.03; SD = 1.37). Seventeen patients showed partial remission with a mean duration of 10.45 weeks (SD = 8.15). None of the patients was still in remission at the 8-month follow-up.

SD =

Correlations

between psychological

data at onset andfollow-ups

A correlation matrix of psychosocial measures at the time of onset and at the follow-ups indicated that our variables correlate well according to the theoretical constructs of coping and adaptation (an abbreviated matrix at time of onset is given in Table I). Change of psychological

parameters

over time

No difference in STAI anxiety, stress upset intensity, global coping and coping features, attributional belief or denial was found between onset and after the followups at 8, 16 and 24 months. Depression (SDS) was the single psychosocial variable that decreased over time with a maximum decrease between onset and the 8-month follow-up (selected scales are given in Table II). Interaction

of psychological

data and metabolic

control

Cross-sectional analyses of variance showed no significant interactions between psychosocial parameters and metabolic control at onset and at the 8,16 and 24 month follow-ups (Table II).

253

Coping and metabolic control in diabetes TABLEI.-CORRELATIONSBETWEEN PSYCHOLOGICAL VARIABLES AT DIABETES ONSET 1

2

3

4

5

6

8

7

1 Trait anxiety

1.00 -

2 Depression

0.77 ***

1.00

3 Global coping

0.78 ***

0.43 *

1.00

4 Anxiety coping

0.64 ***

0.44 *

0.78 ***

1.oo

5 Internal control

0.63 ***

0.37 *

0.87 ***

0.53 **

1.oo

6 External control

0.73 ***

0.59 **

0.85 ***

0.79 ***

0.69 ***

1.oo

7 Family support

0.40 *

0.31

0.43 **

0.30 *

0.24

0.42 *

- 0.42 *

- 0.24

- 0.30 *

- 0.26

- 0.30 *

- 0.42 *

1.00 *

0.78 l **

0.75 ***

0.65 ***

0.47 **

0.62 ***

0.51 **

- 0.25

8 Denial 9 Internal attribution

-044 ** 0.64 ***

9

1.oo

1.00

**p < 0.001; *p < 0.05.

***, < I-J)oOl;

TABLEII.-HBAIc Onset Mean (SD)

AND

PSYCHOSOCIAL VARIABLES OVERTIME

8 months Mean (SD)

16 months Mean (SD)

HbAlc

10.12

(2.65)

7.04

(1.24)

6.43

(1.20)

Trait anxiety Depression

34.57 31.60

(10.78) (5.49)

34.95 25.82

(12.24) (4.72)

33.08 23.95

(10.54) (3.64)

160.57 29.05 44.17 32.27 12.41

(29.90) (8.41) (18.68) (7.98) (4.98)

154.76 22.16 36.77 25.58 9.80

(29.35) (12.28) (17.30) (11.87) (5.49)

155.42 23.54 37.13 24.27 9.45

(28.00) (10.28) (16.97) (12.35) ‘(6.67-3

Denial

16.31

(7.32)

15.92

(7.13)

16.12

Internal attribution

26.36

(8.31)

24.28

(8.45)

26.50

Global coping Anxiety Internal control External control Family support

Repeated measures analysis of variance, g = group (HbAlc

2 years Mean (SD)

PPP tg g x r

(1.37)

*

33.50 (11.28) 24.27 (4.44)

*

7.03

158.69 23.86 38.08 27.24 9.86

(28.55) (11.21) (16.86) (12.59) (6.62)

(6.98)

15.50

(6.67)

(6.73)

28.12

(8.17)

< > 7, 5), r = time, g x f group x time,

*p < 0.05.

Interaction

between psychological

data and decrease of HbAIc

over time

Spearman correlations between psychological variables and changes in metabolic control over time showed no correlation between any of the variables and metabolic control (HbAlc at onset minus follow-up HbAlc were calculated) at the 8- and 16month follow-up respectively. At the 24-month follow-up, however, a stronger

254

IL SPrEXiet al. TABLE III.-PSYCHOSXIAL VARIABLESAT ONSET VERSUS2 YEAM DECREASEOF HBA~c

Spearman State anxiety Trait anxiety

- 0.40 -0.34

Depression

- 0.03

Coping Fatalism Anxiety ht. control Ext. control Social support Family support

-0.64 -0.08 -0.51 -0.64 -0.51 - 0.26 -0.37

Life events Upset intensity

0.06 0.17

Denial

- 0.20

Ext. attribution Int. attribution

-0.12 - 0.45

*p < 0.05;

**p

<

r

p

** ** ** **

*

0.01.

decrease of HbAlc over time was found in patients who had better global coping, higher coping control, lower non-control, lower coping anxiety and emotion-oriented attributional beliefs (Table III). To predict the course of metabolic control, the following variables were analysed by a series of multiple regression analyses: age, sex, remission, STAI, Beck score, life events, coping scales, denial scale, attributional belief scale. Metabolic control (HbAlc) was calculated as the dependent variable. HbAlc decrease over time was not predictable at 8 and 16 month follow-ups using any of the psychosocial parameters. Poor global coping quality at onset was correlated with poor metabolic control (t = 2,50, p < 0.04) at the 2-yr follow-up. Coping explained 22% of variance in the decrease of HbAlc. No other psychological variable was found to predict a decrease of metabolic control at the 2-yr follow-up (Table IV).

DISCUSSION

The most important question of our prospective long-term study was whether early psychological adaptation and moderator variables can predict later metabolic control. Better global coping and lower coping anxiety, and better internal and external controlability, as well as a lower degree of internal health beliefs were associated (Spearman correlations) with a more pronounced decrease of HbAlc between onset and 24-month follow-up but not 8- and 16-month follow-ups on one hand. Anxiety, depression, life events and support, as well as denial on the other hand did not correlate with the decrease in HbAlc over time at any of the follow-ups. If the predictive value of the psychological parameters is calculated by regression analysis, overall coping persisted with a relatively low (22%) variance in predicting HbAlc levels at 2-yr follow-up.

Coping

and metabolic

control

255

in diabetes

The results of our longitudinal study correspond with certain aspects of previous cross-sectional studies, which also found that coping and controllability is associated with good adaptation and good metabolic control [5, 10, 14, 32, 331. Our data corroborate the previous finding that internal health beliefs are correlated with poor metabolic control [34]. However, in our investigation no relationship was found between HbAlc decrease and anxiety, depression, or life events which are frequently associated with poor metabolic control in retrospective cross-sectional studies [l-12]. Our prospective results thus agree with the recently formulated buffer hypotheses [ 141, which suggest, on the basis of retrospective studies, more subtle relationships between the stress caused by anxiety and depression, life events and metabolic control. In these studies stress was correlated with poor metabolic control only when it was associated with ineffective coping [14]. From our finding that coping is the best predictor of later metabolic control, we conclude that coping which moderates stress, life events, depression and anxiety is a better predictor of long-term metabolic control than the latter variables. Since controlability, low coping anxiety and overall coping (the sum of which is to a large extent composed of controlability factors) in our sample were the best predictors for an improved metabolic situation, we assume that, in particular, reframing the disease by directly addressing and confronting behaviour renders metabolic control more independent of possible direct influences of upset emotions. Furthermore psychophysiological studies have shown that individuals react to stress with increase, decrease or unchanged blood glucose levels [35] depending on different hormone moderating coping styles [36]. In particular, reframing as a cognitive style has been shown to act as a direct buffer between cortisol response and anxiety [37], and strong psychophysiological reactions develop when a poor degree of control leads to further attempts at control [38]. This suggests controlability influences metabolic control directly in our sample, especially since in our patients the parameters indicating treatment adherence gave no indication of changes in metabolic control due to behavioural factors. However, it should be mentioned that some of our findings do not support the buffering hypotheses. No relationship was detected in our study between the decrease TABLEW-REGRESSION ANALYSIS OF HBA~c DECREASE OVER 2 YEARS(DEPENDENT VARIABLE) AGAINST THEDIABETES ONSET PSYCHOLOGICAL VARIABLES Parameter estimate

Age Sex

Remission Anxiety Depression Coping Denial Int. attribution Life events Upset intenstiy

0.07 0.70 0.24 - 0.00 -0.21 0.06 0.13 0.19 -0.11 0.02

Standardized estimate 0.42 0.25 - 0.27 - 0.33 -0.71 1.32 0.40 0.65 -0.19 0.17

R’ = 0, 77, t = r-distribution with 6 degrees of freedom. *p < 0.05, Coping accounting for 22% of variance.

t

0.88 0.64 0.37 - 0.09 - 1.89 2.50 1.03 1.53 -0.34 0.34

P

*

256

K. SPIESS et al.

of HbAlc and social support, which other authors have found to act as mediating buffer between stress, depression and adaptation [39,40]. Neither was a relationship between HbAlc decrease and denial, which retrospective cross-sectional studies found to be associated with poor metabolic control in children [41] and adults under stress [5, 141, found. In our sample, depression decreased significantly 8-months after disease onset. Recovering from psychiatric reactions after diabetes onset has been demonstrated elsewhere [18, 42, 431. It was also demonstrated [15, 391 that a connection between psychiatric maladaptation and coping inability (which was not detectable at the onset of the disease) becomes evident only after prolonged disease duration. Subsequently occurring poor metabolic control and late complications correlate positively only with denial processes during the early stage and not with the current psychological adaptation [44]. These findings may explain why in our sample metabolic control was not predictable before the 2-yr follow-up, and they suggest that a relationship between metabolic control and adaptation should predominantly be examined in longitudinal long-term studies. Following the concept of homeostasis and heterostasis [45], individuals try to achieve a balance by adapting the stress resistance regulator to a higher level of defence capacity by using artificial external feeding. In this way, individuals maintain their capacity of coping with stressful demands. As Selye [45] pointed out, arousal, shock and counter-shock phases are followed by a phase of resistance during which the capacity for resistance rises beyond its normal level. During the third stage, the exhaustion phase, adaptation energy may become completely exhausted when the stressor, in our case the diabetic disease, persists. The results of our study may indicate that poor early-coping patterns are balanced during the first 2 yr by increased resistance, but that the first signs of psychophysiological exhaustion occur after 2 yr. It may well be that this results less from direct psychological exhaustion, such as increasing depression, but is effected via moderator variables and psychophysiological processes. Acknowledgements-The authors greatly appreciate the statistical consultation the support by the nursing staff of the Department of Endocrinology and comments of Drs Anton Luger, Irene Virgolini and Christoph Schnack.

provided Diabetes

by V. Scheiber, and the helpful

REFERENCES 1. Cox DJ, TAYLOR AG, NOWACEK G, HOLLEY-WILCOXP, POHL SL, GUTHROW E. The relationship between psychological stress and insulin-dependent diabetic blood glucose control: preliminary investigations. Health Psycho1 1984; 3: 63-75. HANSONCL, HENGGELER SW, BURGHENGA. Social competence and parental support as mediators of the link between stress and metabolic control in adolescents with insulin-dependent diabetes mellitus. .I Consult Clin Psycho1 1987; 55: 529-533. LUSTMANPJ, GRIFFITHLS, CLOUSERE, CRYER PE. Psychiatric illness in diabetes mellitus. Relationship to symptoms and glucose control. J Nerv Menr Dis 1986; 174: 736742. NIEMCRYKSJ. SPEERS MA. TRAVISLB, GARY HE. Psychosocial correlates of hemoglobin Ale in young adults with tipe 1 diabetes. J Psychosom Res 1990; 34: 617427. PEYROTM, MCMURRY JF. Psychosocial factors in diabetes control: Adjustment of insulin-treated adults. Psychosom Med 1985; 47: 542-557. MAYOU R, PEVELERR, DAVIESB, MANN J, FAIRBURNC. Psychiatric morbidity in young adults with insulin-dependent diabetes mellitus. Psycho1 Med 1991; 21: 639-646.

Coping and metabolic control in diabetes

257

and metabolic responses to brief stress in non 7. NALIB~FFBD, COHENMJ, SOWERS JD. Psychological insulin-dependent diabetic and control subjects. J Psychosom Res 1985; 29: 367-374. predictors of 8. LANE JD, STABLERB, Ross SL, MORRIS MA, LITTON JC, SURWIT’~ RS. Psychological glucose control in patients with IDDM. Diabetes Care 1988; 11: 798-800. characteristics of 9. JACOBSEN AM, ADLER A, WOLF~DORF JI, ANDERSON B, DERBY L. Psychological adults with IDDM. Diabetes Care 1990; 13: 375-381. 10. BRADLEY C. Life events and the control of diabetes mellitus. J Psychosom Res 1979; 23: 159-162. 11. GRANT I, KYLE CC, TEICHMAN A, MENDELS J. Recent life events and diabetes in adults. Psychosom Med 1974; 36: 121-128. factors and admission for poor glycaemic control: a study of 12. WRIGLEY M, MAYOU R. Psychosocial psychological and social factors in poorly controlled insulin dependent diabetic patients. J Psychosom Res 1991; 35: 335-343. 13. LUSTMANPJ, GRIFFITH LS, CLOUSE RE. Depression in adults with diabetes. Results of a 5-yr follow up study. Diabetes Care 1988; 11: 605412. 14. PEYROT MF, MCMURRY JF. Stress buffering and glycemic control. Diabetes Cure 1992; 15: 842-846. 15. KOVACS M, BRENT D, STEINBERGTF, PAULAUSKA~ S, REID J. Children’s self-reports of psychologic adjustment and coping strategies during first year of insulin-dependent diabetes mellitus. Diubefes Care 1986; 9: 472479. research in diabetes. Diabetes Care 1986; 9: 546548. 16. JACOBSON AM. Current status of psychosocial program in newly diagnosed 17. GALATZER A, AMIR S, GIL R, KARP M, LARZON 2. Crisis intervention diabetic children. Diabetes Cure 1982; 5: 414-419. 18. KOVACS M, REINBERG T, PAVLASKASS, FINKELSTEINR, POLLOCK M, CROUSE-NOVAK M. Initial coping response and psychosocial characteristics of children with insulin-dependent diabetes mellitus. J Pediatrics 1985: 106: 827.-834. and metabolic 19. KOVACS M, Kiss RE, SCHNELL TM, GOLDSTON D, MARSH J. Family functioning control of school-aeed children with IDDM. Diabetes Cure 1989; 12: 409414. functioning of children with insulin20. KOVACS M, IYEN&R S, GOLDSTON D et al. Psychological dependent diabetes mellitus-A longitudinal study. J Pediutr Psycho1 1990; 15: 619632. 21. NATIONAL DIABETES DATA GROUP. Classification and diagnosis of diabetes and other categories of glucose intolerance. Diabetes 1979; 28: 1039-1057. 22. JUNK J, JUNK U. Entwicklung eines MejZnstrumentes zur epidemiologischen Untersuchung derpsychosozialen Belastung durch Lebensereignisse (live events) und des Coping Potentials. Diplomarbeit. Psychologisches Institut Heidelberg, 198 1. of live events research. Adv Psychosom Med 1987; 17: 13-29. 23. PAYKEL ES. Methodology locus of control in psychiatric patients. J Consult Clin Psycho1 1973; 24. LEVEN~CINH. Multidimensional 41: 397403. Manual. Klinische Selbstbeurteilungsskalen 25 v ZERR~EN D, KOELLER DM. Die Beschwerdenliste. (KSbs). Miinchner Psychiatrisches Informationssystem (Psychis Mcnchen). Beltz Test. Weinheim 1973. 26. BECK A, WARD C, MENDEL~QN M, MOCK J, ERBAUGHJ. An inventory for measuring depression. Arch Gen Psychiatry 1961; 4: 561-571. 27. SPITZERRL, ENDICOTTJ, ROBINSE. Research diagnostic criteria: Rationale and reliability. Arch Gen Psychiatry 1978; 35: 773-782. 28. SPIELBERGER CD, G~RSUCH RL, LUSHENERE. STAI. Manual for the State-trait Anxiety Inventory. Palo Alto: Consulting Psychologist’s Press, 1970. 29. AHRENSS, ELSNERH. Empirische Untersuchungen zum Krankheitskonzept neurotischer, psychosomat&her und somatisch kranker Patienten. Medizinische Psychologie 1981; 7: 95-109. 30. LAZARUSRS, AVERILLJR, OPTON EM. The psychology of coping: Issues of research and assessment. In Coping and Adaptation (Edited by COELHOGV et al.). New York: Academic Press, 1974. 31. HACKETT TB, CA~.$EMNH. Development of a quantitative scale to assess denial. J Psychosom Res 1974; 18: 413420. 32. HAMBURGB, INHOFFG. Relationships between behavioral factors and diabetic control in children and adolescents: a camp study. Psychosom Med 1982; 44: 321-329. 33. RUDOLFMC, AHERN JA, GENELM et al. Optimal insulin delivery in adolescents with diabetes: Impact of intensive treatment on psychosocial adjustment. Diabetes Care 1982; 5: 53-57. 34. SENSKYT, Psm R. Type A behaviour pattern and glycemic control in type 1 diabetes. Psychother Psychosom 1989; 52: 96-100. 35. HALFORDWK, CUDDIHYS, MORTIMER RH. Psychological stress and blood glucose regulation in type 1 diabetic patients. Health Psycho1 1990; 9: 516528. 36. MANSON JW. Organization of psychoendocrine mechanisms. A review and reconsideration of research. In Handbook of Psychophysiology. New York: Holt Rinehart, 1972. 37. BOHNENN, NICOL~~NN, SULONJ, JOLLES J. Coping style, trait anxiety and cortisol reactivity during mental stress. J Psychosom Res 1991; 35: 141-147.

258

K. SPIESS et al.

38. FRANKENHAEUSER M, LUNDBERG V. Psychoneuroendocrine aspects of effort and distress as modified by personal control. In Mental Load and Stress In Activity. European Approaches (Edited by BACHMANN W, UDRIS I),pp. 97-103. Amsterdam: North Holland Publishing Co, 1982.

39. LITTLEFIELD CH, RODINGM, MURRAYMA, CRAVENJL. Influence of functional impairment and social support on depressive symptoms in persons with diabetes. Health PsyChoI 1990; 9: 737-749. 40. GRIFFITHLS, FIELD BJ, LUSTMANPJ. Life stress and social support in diabetes: association with glycemic control. Inr J Psychiar Med 1990; 20: 365-372. 41. TIETZW, VIDMARJT. The impact of coping styles on the control of juvenile diabetes. Psychiat A4ed 1972; 3: 67-74. 42. AHNSJ~S, HUMBLEK, LARS~QNY, SETIXRGREN-CARLSON G, STERKYG. Personality changes and social adjustment during the first three years of diabetes in children. Acta Paediat &and 1981; 70: 321-327. 43. CA~.SILETH BR, LUSK EJ, STROUSETB et al. Psychosocial status in six diagnostic groups. N Engl J Med

1984; 311: 506511. 44. SINZATO R, FUKINO 0, TAMAI H, ISIZU H, NAKAGAWA T, IKEMI Y. Coping behaviors diabetics. Psychother Psychosom 1985; 43: 219-226. 45. SELYE H. The Stress ofLife. New York: McGraw-Hill,

1956.

of severe