Psychological Functioning of Adults With Cystic Fibrosis

Psychological Functioning of Adults With Cystic Fibrosis

Psychological Functioning of Adults With Cystic Fibrosis* Deborah L. Anderson, PhD; Patrick A. Flume, MD, FCCP; and Kristina K. Hardy, PhD Study obje...

110KB Sizes 0 Downloads 130 Views

Psychological Functioning of Adults With Cystic Fibrosis* Deborah L. Anderson, PhD; Patrick A. Flume, MD, FCCP; and Kristina K. Hardy, PhD

Study objectives: The purpose of this study is to assess the psychological profiles of adult patients with cystic fibrosis (CF) and to investigate predictors of patients’ psychological status. Patients and methods: Thirty-four adults with CF completed a battery of psychological testing including the Minnesota Multiphasic Personality Inventory-2, Beck Depression Inventory, and State-Trait Anxiety Inventory. These were compared to health status data, including pulmonary function testing and nutritional status measures. Results: As a group, adults with CF did not demonstrate significant levels of depression, anxiety, or other psychopathology. Results were not affected by age, sex, or severity of disease. Male gender predicted higher scores for depression and anxiety, and better lung functioning predicted less anxiety. Having a higher level of psychosocial support emerged as a strong predictor of better psychological functioning. Conclusions: Overall, adults with CF report relatively healthy psychological functioning. Better lung function and a strong social support system predicted better psychological functioning, which may have implications for clinical intervention. (CHEST 2001; 119:1079 –1084) Key words: adults; cystic fibrosis; Minnesota Multiphasic Personality Inventory; psychological tests Abbreviations: BDI ⫽ Beck Depression Inventory; CF ⫽ cystic fibrosis; IBW ⫽ ideal body weight; MANOVA ⫽ multivariate analysis of variance; MHLOC ⫽ multidimensional health locus of control; MMPI ⫽ Minnesota Multiphasic Personality Inventory; NIH ⫽ National Institutes of Health; STAI ⫽ Spielberger State-Trait Anxiety Inventory

is a high prevalence of psychiatric disorders T here and distress in individuals with chronic physical illness.1 A direct relationship between the severity of chronic conditions and psychosocial adjustment has not been found.2 Instead, there is a broad range of responses to the stress of chronic medical illnesses. Some individuals seem to cope better with the problems related to their illness than others. Poor psychosocial functioning in such patients may adversely affect their physical health. Toward this end, there has been increased attention directed toward investigating psychosocial functioning in adults with chronic medical conditions. Cystic fibrosis (CF) is an example of a chronic illness with an improving survival rate. CF is a multisystem disease characterized by chronic airways obstruction and subsequent infection of the bron*From the Departments of Pediatrics (Dr. Anderson and Ms. Hardy) and Medicine (Dr. Flume), Medical University of South Carolina, Charleston, SC. Manuscript received June 2, 2000; revision accepted November 11, 2000. Correspondence to: Patrick A. Flume, MD, FCCP, Medical University of South Carolina, 812-CSB, 96 Jonathan Lucas St, Charleston, SC 29425; e-mail: [email protected]

chial airways and maldigestion because of pancreatic dysfunction. Before 1940, 85% of CF patients died before the age of 2 years.3 With improved medical management, the number of patients with CF reaching adulthood is steadily increasing. The present median age of survival is 32.3 years (Cystic Fibrosis Foundation registry data; 1998). While the physical impact of CF on young adults has been well studied, less is known about the psychological distress experienced by CF adults. Psychological functioning has been assessed in both children and adults with CF, but the results have been variable. While some investigators have reported relatively normal adjustment in older adolescent and adult CF patients,4 –9 other studies have suggested elevated levels of psychosocial impairment, including anxiety, depression, and eating disorders.10 –13 Several studies have reported that emotional disturbance increases significantly in adult CF patients as they become older.4,5,9,12 Because older patients typically have more advanced disease than younger ones, the higher prevalence of psychopathology may be related to disease severity rather than chronologic age. The mixed results of previous studies may be CHEST / 119 / 4 / APRIL, 2001


because of other factors as well. Many of the older studies focused on a sample of CF patients who are significantly younger than the individuals who are currently surviving into their 30s and beyond. Most studies used checklists and brief questionnaires rather than more time-consuming interviews that provide richer sources of information. Only one study used the more comprehensive Minnesota Multiphasic Personality Inventory (MMPI).5 The goal of our study was to assess the psychological profiles of adult patients with CF using a battery of well-standardized tests. We sought to extend the results of Strauss and Wellisch5 by utilizing the revised version of the MMPI (ie, MMPI-2) and including a larger sample size with a more extensive age range. We further sought to explore whether psychological status varies as a function of age, gender, or disease severity. In addition, we wanted to evaluate the effects of psychosocial support and locus of control on psychological functioning. Materials and Methods We offered multiple tests of psychosocial functioning to all patients attending our adult CF center. Patients who had undergone lung transplantation were excluded from the study. We evaluate patients in the clinic an average of four times annually. The multidisciplinary team includes a pulmonologist, nurse, dietician, physiotherapist, respiratory therapist, and psychologist. Patients received the tests either during a regular clinic visit or at the end of a hospital stay. The staff psychologist instructed the patients on the completion of the forms. In the majority of cases, forms were filled out at home and were returned to the adult CF center by mail. Patients were assured that future care in the CF clinic was in no way dependent on participation in the study and that the results and interpretation of testing would be available through the clinic psychologist. The psychologist was available by telephone to answer any additional questions about the study, completion of questionnaires, or interpretation of results. The study was approved by the Institutional Review Board for Human Research. MMPI-2 The MMPI-214 is a 567-item, true/false, self-report questionnaire designed to assess a number of the major patterns of personality and emotional disorders. The test yields t scores with a mean of 50 and an SD of 10; scores of ⬎ 70 are considered to be clinically abnormal. There are 10 clinical scales plus 3 validity scales that serve as internal checks to ensure satisfactory cooperation and commitment in test completion. Beck Depression Inventory The Beck Depression Inventory (BDI)15 is a 21-item selfreport questionnaire that assesses level of depression. Each item consists of four statements of that the client is to select one. Scores range from 0 to 63; a higher score is associated with a greater degree of depression. A score of ⬍ 10 is considered to be normal and relatively nondepressed, 10 to 20 is considered to be marginally depressed, 20 to 30 is considered to be moderately 1080

depressed, 30 to 40 is considered to be moderately severely depressed, and ⬎ 40 reflects severe depression. Spielberger State-Trait Anxiety Inventory The Spielberger State-Trait Anxiety Inventory (STAI)16 is a 40-item self-report questionnaire that yields scores for stable long-term susceptibility to anxiety (trait anxiety) as well as situational anxiety (state anxiety). For each item, subjects respond “almost never,” “sometimes,” “often,” or “almost always.” Scores range from 20 to 80 for each scale, with higher scores reflecting a greater degree of anxiety. There was a high correlation between state and trait subscales in our population (r ⫽ 0.84; p ⬍ 0.001), so the two were averaged to create a composite anxiety score. Psychosocial Variables Questionnaire Social support was assessed with a 15-item self-report questionnaire.17 For each item, subjects responded “never,” “sometimes,” “usually,” “more often than not,” or “always.” Scores range from 15 to 75 for each subject, with higher scores reflecting greater levels of perceived social support. Multidimensional Health Locus of Control Scales The multidimensional health locus of control (MHLOC) scale18 is an 18-item Likert scale designed to assess beliefs underlying the source of reinforcements for health-related behaviors. Three dimensions of locus of control have been described pertaining to the belief that one becomes sick or healthy as a result of: (1) his or her own behavior (internal locus of control); (2) the influence of powerful others (external locus of control); or (3) by chance (external locus of control). For each item, subjects’ responses range from “strongly disagree” to “strongly agree,” and scores range from 6 to 36 for each of the three subscales. Physical Parameters Measures of health status were taken from the clinical record and included body weight (as a percentage of ideal body weight [IBW]) and pulmonary function testing. An aggregate assessment of each participant’s health was calculated by using the modified National Institutes of Health (NIH) score 19 by the attending physician. Statistical Methods Correlations between similar variables (eg, NIH score, FEV1 percent predicted values, and percentage of IBW) were first examined in order to determine the extent of their association. NIH score was noted to be significantly correlated with both FEV1 percent predicted) (r ⫽ 0.82; p ⬍ 0.001) and percentage of IBW (r ⫽ 0.43; p ⬍ 0.001), so this factor was excluded from multivariate analyses. Comparisons were made based on age, gender, severity of lung disease, and nutritional status. Subjects were classified into two groups based on age (above and below the median age), three groups based on disease severity (FEV1, ⬍ 40% predicted, 40 to 60% predicted, and ⬎ 60% predicted), and two groups based on percentage of IBW (⬎ 90% or ⬍ 90% IBW). Student’s unpaired t tests were conducted in order to examine group mean differences according to age, gender, and IBW. A one-way analysis of variance was run to compare differences between levels of FEV1 percent predicted. In order to determine whether psychosocial support or locus of control might moderate the psychological adjustment of CF patients, multiple regression analyses were used. We conducted a Clinical Investigations

series of three multiple regressions with patient psychopathology measures (BDI, STAI [state and trait scales], and MMPI-2 depression [scale 2] and psychasthenia [scale 7] scores) as the dependent variables and patient support (psychosocial variables questionnaire) and locus of control as independent variables. In order to test the hypothesis that health status variables predict psychological adjustment, a 2 ⫻ 2 ⫻ 3 multivariate analysis of variance (MANOVA) was conducted with gender, percentage of IBW, and FEV1 percent predicted as the independent variables and MMPI-2 depression and anxiety scales, BDI, and STAI (state and trait scales) as the dependent variables. The threshold of statistical significance was set at p ⫽ 0.05.

completed in an acceptable manner, as indicated by the scores on the validity scales. Mean MMPI-2, BDI, and STAI scores were within normal limits on all scales, although subclinical elevations were noted for scale 1 (hypochondriasis) and scale 3 (hysteria) (Fig 1). With regard to individual scores, the percentage of individuals scoring at clinical levels for depression and anxiety was similar to the prevalence

Results Of the 43 patients recruited for the study, 34 (20 men and 14 women) agreed to participate and provided complete data. Except for one AfricanAmerican man, all subjects were white. The age of participants ranged from 18 to 49 years with a mean age of 28.5 years. There was also a wide range of disease severity, as measured by lung function (FEV1 range, 14 to 96% predicted). Only one subject was receiving medication for the treatment of anxiety, and no patients were receiving medication to treat depression. Demographic information is presented in Table 1. The primary reason given for not participating in the study was time involved in completing the questionnaires. Demographic and health status data were available for nonparticipating patients. Participants were significantly less healthy than nonparticipants when health status was based on FEV1 percent predicted (p ⬍ 0.02), NIH score (p ⬍ 0.01), or percentage of IBW (p ⫽ 0.08). Group Results Psychological functioning was investigated first for the group as a whole. All MMPI-2 profiles were Table 1—Characteristics of the Study Population* Characteristics Subjects Total Male Female Age, yr Range Median FVC, % predicted FEV1 % predicted ⬍ 40% predicted 40–60% predicted ⬎60% predicted NIH score Weight, % IBW



34 20 14 28.5 ⫾ 8.0 18 –49 27.5 66.8 ⫾ 19.6

9 5 4 23.1 ⫾ 5.7 17 –34 23.0 89.7 ⫾ 16.4

47.9 ⫾ 20.2 13 13 8 61.6 ⫾ 13.8 91.1 ⫾ 12.2

66.1 ⫾ 15.8 0 3 6 76.4 ⫾ 10.6 98.8 ⫾ 7.1

*Values given as mean ⫾/SD or No. of patients, unless otherwise indicated.

Figure 1. MMPI-2 profile for adults with CF based on gender (top, A), age (above and below median) (middle, B), and severity of lung disease (bottom, C). L ⫽ validity scale; F ⫽ validity scale; K ⫽ validity scale; HS ⫽ hypochondriasis; D ⫽ depression; HY ⫽ hysteria; PD ⫽ psychopathic deviate; MF ⫽ masculinity/femininity; PA ⫽ paranoia; PT ⫽ psychasthenia; SC ⫽ schizophrenia; MA ⫽ mania; SI ⫽ social introversion. CHEST / 119 / 4 / APRIL, 2001


of such individuals in the general population.1 Specifically, only two patients (5.9%) scored outside the normal range on the depression scale for the MMPI-2, and two patients (5.9%) scored outside the normal range on the anxiety scale of the MMPI-2. Categorical Differences Psychological variables (BDI, STAI, and MMPI-2 scores), physical health variables (percentage of IBW and FEV1 percent predicted), gender, and age were evaluated categorically for significant differences. There were no significant differences found between men and women on psychological functioning (Fig 1, top, A). Women tended to have better pulmonary function (t ⫽ ⫺2.03; p ⫽ 0.06), but there was no difference in percentage of IBW between the sexes. With respect to age, scores for younger vs older patients did not differ significantly for either physical or psychological parameters (Table 2; Fig 1, middle, B). There were also no significant differences in psychological functioning when groups were compared according to severity of disease based on lung function (Fig 1, bottom, C) or IBW (⬎ 90% or ⬍ 90% IBW). Predictor Models Although there were no differences in psychopathology when comparing patients based on categorical grouping (ie, age, gender, or severity of disease), we sought to determine whether there were any factors that predicted higher scores on psychological tests. A 2 ⫻ 2 ⫻ 3 MANOVA procedure was performed first on the full set of physical health predictors and mental health outcomes. Specifically, subjects’ gender and levels of percentage of IBW and FEV1 percent predicted served as independent vari-

Psychosocial Support, Locus of Control, and Mental Health

Table 2—Results of Questionnaires* Variables Gender Male Female Age ⬍27 yr ⬎27 yr FEV1, % predicted ⬍40% 40–60% ⬎60% IBW ⬍90% ⬎90%



Psychosocial Support

6.5 ⫾ 3.9 4.1 ⫾ 3.1

37.4 ⫾ 10.7 31.2 ⫾ 6.2

48.6 ⫾ 9.6 55.4 ⫾ 6.1

5.4 ⫾ 3.8 5.7 ⫾ 3.7

33.8 ⫾ 6.9 35.6 ⫾ 11.3

51.3 ⫾ 8.5 51.5 ⫾ 9.4

5.1 ⫾ 3.4 5.7 ⫾ 3.9 6.1 ⫾ 4.1

35.6 ⫾ 12.1 35.8 ⫾ 8.4 31.9 ⫾ 6.7

53.9 ⫾ 9.1 47.8 ⫾ 9.2 53.2 ⫾ 7.0

5.4 ⫾ 3.4 5.7 ⫾ 4.0

32.6 ⫾ 7.0 36.6 ⫾ 11.0

53.4 ⫾ 8.3 49.8 ⫾ 9.3

*Values are given as mean ⫾ SD. 1082

ables, with MMPI-2 depression and anxiety scales, BDI scores, and STAI scores serving as dependent variables. Gender was included as a predictor because of the nearly significant difference in lung functioning between men and women in the sample in order to examine any potential interaction with health status variables. Results from the MANOVA indicated significant main effects on the set of psychological variables for gender (Wilks ␭, 0.66; F ⫽ 2.95; p ⬍ 0.05) and both physical health predictors (percentage of IBW: Wilks ␭, 0.57; F ⫽ 4.34; p ⬍ 0.01; FEV1 percent predicted: Wilks ␭, 0.52; F ⫽ 2.19; p ⬍ 0.05). Specifically, male gender predicted higher scores of depression and anxiety, and having better health status predicted lower scores of psychological distress. No significant interactions were found between predictor variables. In order to determine the effects of physical health status on specific psychological variables, univariate analyses of variance were performed next. With respect to depression scores, gender predicted BDI scores (but not MMPI-2 depression scale scores), such that men reported higher levels of depression than did women (F ⫽ 6.16; p ⬍ 0.05). Physical health indexes failed to predict either measure of depression (BDI or MMPI-2 scores). For anxiety scores, men were found to have significantly higher scores on the MMPI-2 anxiety scale (F ⫽ 6.39; p ⬍ 0.05), and there was a similar trend for the composite STAI score (F ⫽ 3.50; p ⫽ 0.07). In terms of physical health predictors, univariate analyses indicated significant findings for STAI scores only, such that individuals with more impaired pulmonary functioning tended to report increased anxiety (F ⫽ 6.32; p ⬍ 0.01). In contrast, subjects with subnormal body weight reported significantly lower levels of anxiety than those with body weights within the normal range (F ⫽ 17.27; p ⬍ 0.001).

In order to examine the impact of psychosocial support and locus of control on mental health status, several regression analyses were performed. In each case, psychosocial support or the three scales of the MHLOC scales were included as predictors of the four indexes of mental health functioning. Broadly speaking, a higher level of psychosocial support emerged as a strong predictor of better psychological functioning. Specifically, increased psychosocial support scores predicted lower BDI scores (␤ ⫽ ⫺0.47; t ⫽ ⫺3.01; p ⬍ 0.01), MMPI-2 anxiety scale scores (␤ ⫽ ⫺0.41; t ⫽ ⫺2.51; p ⬍ 0.05), and composite STAI scores (␤ ⫽ ⫺0.30; t ⫽ ⫺1.80; p ⫽ 0.08), but not MMPI-2 depression scale scores. Clinical Investigations

We examined the three MHLOC scales in a separate model. No significant findings resulted from these analyses. Discussion Our results show that as a group, adults with CF do not demonstrate significant levels of depression, anxiety, or other psychopathology. The rate of clinical depression in our sample was equivalent to that found in the general population.1 We also found no evidence of psychopathology when our patients were evaluated according to age, gender, and severity of illness. The subclinical elevations on scales 1 (hypochondriasis) and 3 (hysteria) are similar to findings reported in chronic pain populations.20,21 Subclinical elevations of these two scales in a healthy subject would be interpreted as having a tendency to express emotion with somatic complaints, but in the context of a chronic illness such elevations would be expected and not considered to represent psychopathology. It also should be noted that some of our patients completed the study during hospitalization (although at the end of their hospital stays). The questionnaires used are designed to assess stable traits rather than situational states. The stress of illness requiring hospitalization should not have affected the results dramatically. If it could affect the results, one might expect the results to demonstrate greater pathology, which is not what we found. Our findings are consistent with previous studies describing relatively healthy psychological functioning overall in adults with CF.4 –9 Similar to earlier research using the original version of the MMPI to assess psychological functioning,5 we found no evidence of increased anxiety or depression in CF patients using the MMPI-2. Our findings are in contrast, however, to what has been reported previously in children with CF. Although there is evidence that children with more severe CF have more psychological problems than children with milder disease,22 we did not find the same in our adults with CF. There are potential reasons to explain this finding. First, there may have been biases in the population studied such that a larger group would have revealed normal psychological functioning. Second, there may be a difference in the psychological functioning of children who survived into adulthood compared to those who did not. Perhaps the more interesting question is what protective factors allow these patients with CF to function so normally given the physical, social, and emotional stressors that they face. There were some associations between physical parameters and psychological test scores. When the impact of gender on psychological functioning was examined, men were

found to be at higher risk for depression than women, according to scores on the BDI, and were also at higher risk for anxiety, according to scores on both the MMPI-2 and the STAI scale. More severe lung impairment was predictive of higher levels of anxiety, as reflected by the STAI score. Surprisingly, subnormal body weight predicted lower levels of anxiety. Another factor that has been investigated in terms of impacting health status is locus of control, which refers to an individual’s beliefs about what determines his or her health. Previous research concerning adults with CF indicates that measures of locus of control (specifically, external locus of control for powerful others are associated with significantly higher levels of adherence23) were not found to be associated with psychological functioning in our study. Finally, increased level of psychosocial support emerged as a strong predictor of better psychological functioning. This finding is consistent with literature documenting that greater social support reduces the consequences of stress associated with having a chronic medical illness.24 –26 An important point to be made herein is that although the findings of our study demonstrate that, as a group, the adults with CF have normal psychological functioning, this does not mean that every adult with CF is functioning normally. In fact, four of our patients scored outside the normal range on scales for either depression or anxiety. Adults with CF may experience psychological problems that warrant therapy, just as do adults without CF. It is important to note that considering such problems in our patients as psychological dysfunction may adversely affect their health (eg, nonadherence to medical therapy). This may be of particular importance for our patients with lung disease severe enough to consider lung transplantation, as adherence to a strict medical regimen is of critical necessity, especially when donor lungs are scarce. Although such rigorous psychological testing as was performed herein is not necessary for every CF patient, these tests may prove valuable for the evaluation of the patient considering transplantation. In many respects, the results of this study indicate that the psychological functioning of adults with CF resembles that of healthy adults. Despite restrictions in daily living and the likelihood of premature death, adults with CF as a group deny having significant anxiety or depression. In addition, better lung functioning and strong social support were found to predict better psychological functioning. The findings from this study may have implications for the development of interventions to enhance the quality of life for patients with CF. Efforts should be directed at enhancing the social support system of patients with CF, which should act to buffer the CHEST / 119 / 4 / APRIL, 2001


identified patient from depression and anxiety. This would appear to justify the role of the psychologist and the social worker in the CF clinic. Focusing medical intervention on improving lung function should similarly improve psychological functioning. By assisting the adult CF patient in developing these characteristics associated with better psychological adjustment, we may simultaneously enhance quality of life for these individuals. References 1 Wells K, Golding J, Burnham M. Psychiatric disorder in a sample of the general population with and without chronic medical conditions. Am J Psychiatry 1988; 145:976 –981 2 Wallander J, Thompson RJ. Psychosocial adjustment of children with chronic physical conditions. In: Roberts M, ed. Handbook of pediatric psychology. New York, NY: Guilford Press, 1995; 124 –141 3 Shwachman H, Redmond A, Khaw K. Studies in cystic fibrosis: report of 130 patients diagnosed ⬍ 3 months of age over a 20-year period. Pediatrics 1970; 46:335–343 4 Cowen L, Corey M, Simmons R, et al. Growing older with cystic fibrosis. Psychosom Med 1984; 46:363–376 5 Strauss D, Wellisch D. Psychological assessment of adults with cystic fibrosis. Int J Psychiatry Med 1980; 10:265–272 6 Shepherd S, Hovell M, Harwood I, et al. A comparative study of the psychosocial assets of adults with cystic fibrosis and their healthy peers. Chest 1990; 97:1310 –1316 7 Moise J, Drotar D, Doershuk C, et al. Correlates of psychological adjustment among young adults with cystic fibrosis. J Dev Behav Pediatr 1987; 8:141–148 8 Capelli M, McGrath P, Heick C, et al. Chronic disease and its impact. J Adolesc Health Care 1989; 10:283–288 9 Blair C, Cull A, Freeman C. Psychosocial functioning of young adults with cystic fibrosis and their families. Thorax 1994; 49:798 – 802 10 Boyle I, di Sant’Agnese P, Sack S, et al. Emotional adjustment of adolescents and young adults with cystic fibrosis. J Pediatr 1976; 88:318 –326 11 Lawler R, Nakielny W, Wright N. Psychological implications


of cystic fibrosis. Can Med Assoc J 1966; 94:1043–1146 12 Pearson D, Pumariega A, Seilheimer D. The development of psychiatric symptomology in patients with cystic fibrosis. J Am Acad Child Adolesc Psychiatry 1991; 30:290 –297 13 Pumariega A, Pursell J, Spock A, et al. Eating disorders in adolescents with cystic fibrosis. J Am Acad Child Psychiatry 1986; 25:269 –275 14 Hathaway S, McKinley J. MMPI-2: Minnesota multiphasic personality inventory-2. Minneapolis, MN: University of Minnesota, 1989 15 Beck A. Depression inventory. Philadelphia, PA: Center for Cognitive Therapy, 1978 16 Spielberger C, Gorsuch R, Lushene R. State trait anxiety inventory manual. Palo Alto, CA: Consulting Psychologists Press, 1970 17 Procidano M, Heller K. Modification of perceived social support from friends and from family: three validation studies. Am J Community Psychol 1983; 11:1–24 18 Walston K, Wallston B, DeVellis R. Development of the multidimensional health locus of control (MHLC) scales. Health Educ Monogr 1978; 6:160 –170 19 Sockrider M, Swank P, Seilheimer D, et al. Measuring clinical status in cystic fibrosis: internal validity and reliability of a modified NIH score. Pediatr Pulmonol 1994; 17:86 –96 20 Guck T, Meilman P, Shultery F, et al. Pain patient MMPI subgroups: evaluation of long-term treatment outcomes. J Behav Med 1988; 11:159 –169 21 Bradley L, Prokop C, Margolis R, et al. Multivariate analyses of the MMPI profiles of low back pain patients. J Behav Med 1978; 1:253–257 22 Simmons R, Corey M, Cowen L, et al. Behavioral adjustment of latency age children with cystic fibrosis. Psychosom Med 1987; 49:291–301 23 Myers L, Myers F. The relationship between control beliefs and self-supported adherence in adults with cystic fibrosis. Psychol Health Med 1999; 4:387–391 24 Haggerty R. Life stress, illness, and social supports. Dev Med Child Neurol 1980; 22:391– 400 25 Cohen S, Wills T. Stress, social support, and the buffering hypothesis. Psychol Bull 1985; 98:310 –357 26 Sarason B, Sarason I, Pierce G. Social support: an interactional view. New York, NY: Wiley, 1990; 64 –94

Clinical Investigations