Stress Buffering and Glycemic Control - Diabetes Care

7 downloads 0 Views 677KB Size Report
RESEARCH DESIGN AND METHODS— Subjects consisted of 105 insulin- treated adults from the Diabetes Division of Henry Ford Hospital who filled out.
mmmrn

I

N A L

A R T I C L E

Stress Buffering and Glycemic Control The Role of Coping Styles MARK F. PEYROT, PHD JAMES F. MCMURRY, JR, MD

OBJECTIVE — To test the hypotheses that chronic psychosocial stress is associated with worse glycemic control and that coping moderates (buffers) this effect. RESEARCH DESIGN AND METHODS— Subjects consisted of 105 insulintreated adults from the Diabetes Division of Henry Ford Hospital who filled out questionnaires on stress and coping and received an HbAi test at a clinic appointment. Six coping styles were examined, including both emotion- and problemfocused styles. Two standardized stress inventories were administered. Ineffective coping was defined as scoring below the median for stress-dampening coping styles and above the median for stress-exacerbating styles. RESULTS— Stress was significantly (P < 0.05) correlated with higher HbAi in all but one ineffective coping subgroup. Conversely, none of 12 correlations between stress and glycemic control was significant in the effective coping subgroups. CONCLUSIONS — Chronic psychosocial stress is associated with worse glycemic control among those who do not cope effectively with stress. Effective coping can protect individuals from the deleterious effects of stress.

C

hronic psychosocial stress is related to deranged metabolic control in diabetes among adults (1) and children (2). Two alternative explanations have been proposed for this relationship (3). The first hypothesis states that stress disrupts behavioral routines in which the disease is managed (medication, eating, and exercise), resulting in derangements

in metabolic control (4). The second hypothesis states that stress initiates a psychophysiological process of arousal in which counterregulatory hormones are secreted, which in turn increase blood levels of physiological fuels such as glucose and free fatty acids (5,6). Several studies have attempted to assess the viability of one or both of the

FROM THE LOYOLA COLLEGE CENTER FOR SOCIAL AND COMMUNITY RESEARCH, BALTIMORE; AND THE JOHNS HOPKINS UNIVERSITY SCHOOL OF MEDICINE, BALTIMORE, MARYLAND; AND THE GEORGETOWN UNIVERSITY MEDICAL CENTER, WASHINGTON, DC. ADDRESS CORRESPONDENCE AND REPRINT REQUESTS TO MARK PEYROT, PHD, LOYOLA COLLEGE, 4501 NORTH CHARLES STREET, BALTIMORE, MD 2 1 2 1 0 - 2 6 9 9 . RECEIVED FOR PUBLICATION 18 JULY 1991 AND ACCEPTED IN REVISED FORM 24 JANUARY 1992.

842

hypothesized pathways for the effect of stress on metabolic control. Nonexperimental studies have controlled statistically for self-care behavior to determine whether such controls eliminate the relationship between stress and glycemic control, thereby indicating that self-care behavior acts as the intervening variable. Several of these studies demonstrate the plausibility of the psychophysiological hypothesis (1,2,7). Experimental studies have examined the psychophysiological hypothesis directly by exposing subjects to stressors and observing transient biochemical alterations. Some studies have found the hypothesized effect (8), whereas others have not (9,10). These differences in results have produced confusion regarding the psychophysiological hypothesis. Although the inconsistencies in research results are troubling, there may be a simple explanation for the inconsistent findings. For example, there is variation among individuals in their stress responsiveness. Some of the reported studies may include more stress-responsive subjects, whereas other studies may include fewer such subjects. In one study of adults, the effects of chronic stress differed from individual to individual, with some individuals responding to stress with elevated blood glucose, whereas others responded with decreased blood glucose or no change (11). Another study found differences in adults' stress-response profile to be stable over time, indicating that the differences in stress responsiveness were the result of stable individual traits (12). The authors suggested that coping styles may account for the differences in stress responsiveness. The stress literature has suggested that the deleterious effects of stress can be minimized through the use of stress-management techniques, such as social support, coping behavior, and the relaxation response. Research has shown that glycemic control can be improved by relaxation training (5) and interventions to increase social support

DIABETES CARE, VOLUME 15, NUMBER 7, JULY 1992

Peyrot and McMurry

(13). However, there have been few studies demonstrating that coping styles protect individuals from the effects of stress (14,15), especially among adults. Most of the research on the effect of personality or behavioral style on glycemic control has examined the possibility of simple direct effects (3,16,17). Selected styles have been hypothesized to be associated with better or worse glycemic control. The assumption of this main effect model is that a coping style increases or decreases psychosocial stress, which in turn elevates or depresses glycemic levels. In contrast, this study investigated the stress-buffering hypothesis (18), examining whether certain coping styles are protective against stress or make individuals more vulnerable to stress. If a particular type of coping behavior serves a stress-buffering function, people performing that style of coping at different levels of frequency or intensity exhibit different levels of stress responsiveness. For example, the type A behavior style has been hypothesized to increase stress responsiveness; type A people are hypothesized to be more responsive to stress than type B people. Some research has supported the hypothesis that the type A style affects stress responsiveness (19) but other studies have not (1,20). The two major categories of coping style are emotion-focused and problem-focused coping (21,22). Problemfocused coping seeks to reduce the threat associated with some environmental condition by changing the condition, thereby eliminating the threat. Emotionfocused coping addresses the threat without changing the environment, e.g., by reinterpreting the condition so that it no longer seems to pose a threat or calming oneself in the face of the threat. Both types of coping styles have been hypothesized to ameliorate the effects of stress, although both may be ineffective or even exacerbate the effects of stress. This study examines the role of both coping styles in buffering the effect of psychosocial stress on glycemic control.

DIABETES CARE, VOLUME 15,

NUMBER 7, JULY

1992

This study examines the hypothesis that stress responsiveness conditions the relationship between stress and metabolic control. Specifically, we hypothesize that the relationship between stress and metabolic control will depend on individuals' coping styles. For a coping style that exacerbates stress, metabolic control will be more strongly related to stress among those who use that style more compared with those who use it less. For a coping style that dampens stress, metabolic control will be more strongly related to stress among those who use that style less compared with those who use it more. That is, for people who typically use stress-dampening coping styles or infrequently use stressexacerbating coping styles, glycemic control will be less responsive to stress. We further hypothesize that differences in emotion-oriented coping produce greater differences in stress responsiveness than different levels of problemfocused coping. That is, people lacking emotion-focused coping skills or exhibiting emotional arousal coping styles will be more stress responsive than those lacking problem-focused coping skills. Research suggests that emotion-focused coping is more effective than denial or problem-focused coping when dealing with a chronic stressor that cannot be easily changed (23). RESEARCH DESIGN AND METHODS— Subjects were adult patients from the Diabetes Division of Henry Ford Hospital in Detroit, Michigan. Only insulin-treated patients were recruited for this study. During a clinic visit, patients were asked to participate by filling out a self-administered questionnaire packet containing several measures. Of those patients contacted, most (n = 189) signed a consent form and filled out the questionnaire. Patients who had an HbAx test performed at that visit (n = 105) were included in our analysis. Subjects were equally divided between early (age at onset 30 yr) onset. The

mean ± SD duration of diabetes was 12.5 ± 9.9 yr, and mean age was 45.1 ± 15.7 yr. There were more women than men (58 vs. 42%). Sixtytwo percent of the subjects were nonHispanic white, 69% were married, and 44% had attended college.

Dependent variable Total-blood HbAt measured by affinitive chromatography, was used to assess metabolic control. The published range of normal values for this assay was 4.08.8%, and the scores in this study ranged from 6.1 to 16.0% (mean ± SD 11.2 ± 2.4%).

Independent variables Psychosocial stress was measured by two instruments: the short form of the Perceived Stress Scale (24) and the 10 most commonly chosen items from the Hassles Scale (25). Response options for the instruments were those used in the original publications, with higher scores indicating greater stress. Subjects reported stress levels over the last 1-2 mo. The Perceived Stress Scale has an internal consistency of 0.72 and a 2-mo testretest correlation of 0.55. The Hassles inventory is not assessed for internal reliability because it is not intended to measure a single underlying construct but a collection of disparate events; the average monthly test-retest correlation for the entire scale is 0.48 (no data is available for the 10 most commonly chosen items).

Conditioning variables Coping styles were measured with our own previously published instrument (3). This instrument yields measures of six global (nondiabetes-specific) coping styles: one problem-focused coping style (pragmatism), one avoidance coping style (denial), one emotional control coping style (stoicism), and three emotional arousal coping styles (anxiety, impatience, anger). Higher scores indicate that the coping style is more like me. Internal consistency for these coping

843

Stress buffering

measures ranged from 0.59 to 0.64, indicating moderate item agreement. Each of these coping styles, except pragmatism, was associated with glycemic control in the original study. Each coping style score was dicotomized at the sample median to create a low and high group for each coping style. Ineffective coping subgroups are those high in stress-exacerbating coping styles (anger, impatience, anxiety) and low in stress-dampening coping styles (stoicism, denial, pragmatism). For convenience, people low in stress-exacerbating styles and high in stress-dampening styles were referred to as effective coping subgroups (although strictly speaking, only the latter are effective copers, those low in stress-exacerbating styles are more effective or less ineffective than their counterparts in the ineffective coping subgroup).

Analysis The relationship between stress and glycemic control was measured with zeroorder Pearson correlations. The conditioning (stress-buffering) effect was assessed by identifying high and low scorers for each coping style and obtaining separate correlations between stress and HbAx in each group. One-tailed significance levels were obtained for each correlation between stress and glycemia; this procedure was used because all correlations were hypothesized to be positive. The differences in the correlations for the high and low groups were analyzed by transforming the correlations into 2 scores and computing a 2 score for the difference between high and low groups (26). One-tailed significance levels were used because we tested the directional hypothesis that the correlations between stress and glycemic control were larger in the ineffective coping subgroup than in the effective coping subgroup. RESULTS— Before conducting the main analyses, we examined the correla-

844

Table 1—Correlations glycemic control

of

stress

with

COPING STYLE

Z

STRESS MEASURE

Low

HASSLES

0.37* 0.23t

HIGH

SCORE

0.00 -0.01

1.91T 1.20

STOICISM

PSS

DENIAL HASSLES

PSS

0.28t 0.20T

0.13 0.03

0.74 0.82

PRAGMATISM HASSLES

PSS

0.15 0.16

0.08 0.01

0.35 0.75

0.37* 0.24T

1.82t 1.37*

ANGER HASSLES

PSS

0.02 -0.03

IMPATIENCE HASSLES

PSS

0.17 -0.02

0.26T 0.28t

0.47 1.53*

0.25t 0.27t

0.47 1.48*

ANXIETY HASSLES

PSS

0.16 -0.02

Squares, ineffective coping subgroups. PSS, perceived stress scale. *P < 0.01 (1 tail). tP < 0.05 (1 tail). * P < 0 . 1 0 (1 tail).

tions among our coping style measures. The three emotional arousal styles (anger, impatience, anxiety) were all positively correlated with each other (r = 0.35-0.44, P < 0.001). Pragmatism was negatively correlated with denial, anxiety, and anger (r = —0.19 to - 0 . 2 8 , P < 0.025). No other coping style measures were significantly correlated (P > 0.05). Table 1 presents the correlations between glycemic control and the two measures of psychosocial stress. All cor-

relations in the ineffective coping subgroups were in the predicted direction; higher stress was associated with higher HbA: levels (worse glycemic control). In the ineffective coping subgroups, 10 of 12 correlation correlations were significant (r > 0.20, P < 0.05); only for low pragmatism did the correlations between stress and glycemic control fail to reach significance. The largest correlations (r > 0.35) occurred for ineffective emotion-oriented coping, i.e., people high in anger and low in stoicism. In the effective coping subgroups, none of the 12 correlations were significant. Table 1 also presents the tests of differences in the magnitude of the correlations among people high and low in each of the coping styles. Only 2 of 12 comparisons were significant but, as predicted, both were for coping styles that involved emotions—anger and stoicism. Although only 2 of the tests were statistically significant, all 12 differences were in the predicted direction, with stress more strongly related to glycemic control for the ineffective coping group. The probability of all 12 differences being in the predicted direction due to chance is very low (P < 0.001 by the sign test). CONCLUSIONS— This study supports the hypothesis that coping styles buffer the impact of chronic stress on glycemic control. Caution should be exercised in interpreting these results because effect sizes are not large, and potentially confounding demographic and clinical factors have not been considered. However, in this analysis, higher levels of stress result in higher glycemic levels among people who do not cope effectively with stress, whereas the glycemic control of those who cope more effectively with stress is not affected. Two emotion-oriented coping styles (stoicism and anger) have the strongest effect on stress responsiveness. Stoicism is an emotionally controlled style reflecting a tendency not to respond emotionally to stressful situations. The glycemic control of people who are stoic is not responsive

DIABETES CARE, VOLUME 15,

NUMBER 7, JULY

1992

Peyrot and McMurry

to stress, but for those low in stoicism, glycemic control is significantly related to stress. Anger is another behavioral style that influences stress responsiveness. People who typically respond to stressful events and situations with anger exhibit a moderately strong, positive relationship between stress and glycemic control. This finding fits well with recent research that implicates anger/hostility as a pathogenetic factor in cardiovascular disease (27). People high in anger/hostility exhibit greater cardiovascular reactivity to stress (28). That research has suggested that the type A behavior pattern is a risk factor only to the degree that it is confounded with anger/hostility (29). This hypothesis may explain the failure of some studies to find either an additive or interactive relationship between type A behavior pattern and glycemic control (1,21). Perhaps future research should focus on the anger/hostility component of the type A behavioral cluster. However, there is some evidence that the impatient coping style, another component of type A behavior, also conditions the effects of stress. Among those who respond more with impatience, stress is significantly related to glycemic control. Some cardiovascular research has indicated that impatience is an important pathogenetic factor (30). Impatience may influence stress responsiveness by increasing emotional arousal. People who are impatient may respond to stressful situations with frustration, which may in turn lead to anger arousal (31; H.P. Machado, Loyola College, Baltimore, MD, unpublished observations). For two of the three remaining coping styles (anxiety, denial), stress was associated with glycemic control in the ineffective coping group. People who respond with anxiety are more sensitive to stress, as are people who do not use denial. The finding for anxiety is intuitively understandable, although that for denial may be counterintuitive. Denial is generally regarded as an ineffective problem-oriented coping style for dealing

DIABETES CARE, VOLUME 15,

NUMBER 7, JULY

with chronic stressors (23); one must first acknowledge a problem before it can be dealt with. However, research has suggested that denial can be an effective emotional control technique for dealing with stress, especially where there is no behavioral remedy for the stressful situation (32). Although this study suggests that denial can be effective for stress buffering, people who engage in denial do not necessarily have better glycemic control. In our earlier study of main effects for coping styles (3), denial was associated with worse glycemic control and this effect might offset any stress-buffering benefits. Of the six coping styles, pragmatism showed the smallest degree of stress buffering. The relationship between stress and control in the ineffective coping group was the smallest of all coping styles and the difference in the stressglycemia relationship between effective and ineffective coping groups was also the smallest. Despite the finding that this problem-oriented coping style may not serve a stress-buffering function, we note that it may serve to reduce stress by remedying the problem causing the stress. A pragmatic coping style may represent a methodical approach that can be sustained over time. In this respect, it may be the opposite of denial; its efficacy may be in dealing with chronic stressors that respond to a behavioral strategy, e.g., meeting the demands of an intensive treatment regimen. Although this is a plausible hypothesis (and pragmatism is negatively correlated with denial), this study has not addressed this issue. This study provides evidence that several behavioral dispositions (coping styles) are protective against chronic stress (or alternatively their absence makes people more vulnerable to stress). The results suggest which types of coping styles are most promising candidates for future research—those that increase or decrease emotional arousal. More generally, this study suggests an alternative approach to studying the effects of psychological traits on glycemic control.

1992

Most previous research has only considered simple main effects, that is, whether those who have more or less of a trait exhibit better or worse glycemic control. This study has shown that personalitylike traits can affect glycemic control in a different way, by making individuals more or less responsive to stress. Future research on personality factors in diabetes should investigate this possibility. Future research on the role of personality in diabetic control should also address the issue of how these factors affect the stress-glycemia relationship. Stress may act either through a psychophysiological pathway (e.g., by stimulating the release of counterregulatory hormones) or a behavioral pathway (e.g., by disrupting self-care regimens). Personality or coping styles that buffer the effects of stress may produce their effects through either mechanism (3). The psychophysiological mechanism might be investigated with a controlled stress test in a laboratory experiment and observing whether subjects with different personality traits exhibit different physiological responses to acute stress. The behavioral pathway can be investigated by examining the effect of chronic stress on the diabetes self-care behavior of people with different personality traits. Such research would not only contribute to our understanding of the role of personality traits but also to our understanding of how stress affects different types of individuals.

Acknowledgments—We thank Davida F. Kruger for assistance as project coordinator for data collection and Fred W. Whitehouse for allowing us to conduct this study. References 1. Cox DJ, Taylor AG, Nowacek G, HolleyWilcox P, Pohl SL, Guthrow E: The relationship between psychological stress and insulin-dependent diabetic blood glucose control: preliminary investigations. Health Psychol 3:63-75, 1984 2. Hanson CL, HengglerSW, Burghen GW:

845

Stress buffering

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

846

Model of associations between psychosocial variables and health-outcome measures of adolescents with IDDM. Diabetes Care 10:752-58, 1987 Peyrot M, McMurry JF: Psychosocial factors in diabetes control: adjustment of insulin-treated adults. Psychosom Med 47: 542-57, 1985 Wilson W, Ary DV, Biglan A, Glasgow RE, Toobert DJ, Campbell DR: Psychosocial predictors of self-care behavior (compliance) and glycemic control in non-insulin-dependent diabetes mellitus. Diabetes Care 9:614-22, 1986 Surwit RS, Feinglos MN: The effects of relaxation on glucose tolerance in noninsulin-dependent diabetes. Diabetes Care 6:176-79, 1983 Unger RH, Orci L: Stress and diabetes. In Critical Issues in Behavioral Medicine. West LS, Stein M, Eds. Philadelphia, PA, Lippincott, 1982 Cohen AS, Vance VK, Runyan JW, Hurwitz D: Diabetic acidosis: an evaluation of the cause, course, and therapy of 73 cases. Ann Intern Med 52:55-86, 1960 Baker L, Minuchin S, Milman L, Liebman R, Todd T: Psychosomatic aspects of juvenile diabetes mellitus: a progress report. Mod Prob Paediatr 12:332-43, 1975 Kemmer FW, Bisping R, Steingruber HJ, Baar H, Hardtmann F, Schlaghecke R, Berger M: Psychological stress and metabolic control in patients with Type I diabetes. N Engl J Med 314:1078-84, 1986 Edwards C, Yates AJ: The effects of cognitive task demand on subjective stress and blood glucose levels in diabetics and nondiabetics. J Psychosom Res 29:59-69, 1985 Halford WK, Cuddihy S, Mortimer RH: Psychological stress and blood glucose regulation in type 1 diabetic patients. Health Psychol 9:516-28, 1990 Gonder-Frederick LA, Carter WR, Cox DJ, Clarke WL: Environmental stress and

13.

14.

15.

16.

17.

18.

19.

20.

21.

blood glucose change in insulin dependent diabetes mellitus. Health Psychol 9:503-15, 1990 Ametz BB: The potential role of psychosocial stress on levels of hemoglobin A lc (HbAlc) and fasting plasma glucose in elderly people. J Gerontol 39:424-29, 1984 Brand AH, Johnson JH, Johnson SB: Life stress and diabetic control in children and adolescents with insulin-dependent diabetes. J Pediatr Psychol 11:481-95, 1986 Hanson CL, Henggeler SW, Burghen GA: Social competence and parental support as mediators of the link between stress and metabolic control in adolescents with insulin-dependent diabetes mellitus. J Consult Clin Psychol 55:529-33, 1987 Jacobson AM, Adler AG, Wolfsdorf JI, Anderson BJ, Derby L: Psychological characteristics of adults with IDDM: comparison of patients in poor and good glycemic control. Diabetes Care 13:37581, 1990 Lane JD, Stabler B, Ross SL, Morris MA, Litton JC, Surwit RS: Psychological predictors of glucose control in patients with IDDM. Diabetes Care 11:798-800, 1988 Cohen S, Wills TA: Stress, social support, and the buffering hypothesis. Psychol Bull 98:310-57, 1985 Stabler B, Surwit RS, Lane JD, Morris MA, Litton JC, Feinglos MN: Type A behavior pattern and blood glucose in diabetic children. Psychosom Med 49: 313-17, 1987 Stabler B, Lane JD, Ross SL, Morris MA, Litton JC, Surwit RS: Type A behavior pattern and chronic glycemic control in individuals with IDDM. Diabetes Care 11:361-62, 1988 Billings AG, Moos RH: The role of coping responses and social resources in attenuating the stress of life events. J Behav

Med 4:139-57, 1981 22. Folkman S, Lazarus RS: An analysis of coping in a middle-aged community sample. J Health Soc Behav 21:219-39, 1980 23. Collins DL, Baum A, Singer JS: Coping with chronic stress at Three Mile Island: psychological and biochemical evidence. Health Psychol 2:149-66, 1983 24. Cohen S, Kamarak T, Mermelstein R: A global measure of perceived stress. J Health Soc Behav 24:385-96, 1983 25. Kanner AD, Coyne JC, Schafer C, Lazarus RS: Comparison of two modes of stress measurement: daily hassles and uplifts versus major life events. J Behav Med 4:1-39, 1981 26. Rosenthal R, Rosnow RL: Essentials of Behavioral Research: Methods and Data Anal-

ysis. New York, McGraw Hill, 1984 27. Williams RB, Haney TL, Lee KL, Kong H, Blumenthal JA, Whalen RE: Type A behavior, hostility, and coronary atherosclerosis. Psychosom Med 42:539-49, 1980 28. Harbin TJ: The relationship between the Type A behavior pattern and physiological responsivity: a quantitative review. Psychophysiology 26:110-19, 1989 29. Dembrowski TM, MacDougall JM, Williams RB, Haney TL, Blumenthal JA: Components of Type A, hostility, and anger-in: relationship to angiographic findings. Psychosom Med 47:219-33, 1985 30. Booth-Kewley S, Friedman HS: Psychological predictors of heart disease: a quantitative review. Psychol Bull 101: 343-62, 1987 31. Ohman A, Nordby H, Svebak S: Components of type A behavior and taskinduced cardiovascular activation. Psychophysiolo^ 26:81-88, 1989 32. Levinson JL, Mishra A, Hamer RM, Hastillo A: Denail and medical outcome in unstable angina. Psychosom Med 51: 27-35, 1989

DIABETES CARE, VOLUME 15,

NUMBER 7, JULY

1992