| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLES |
From the Dana Farber Cancer Institute and Harvard School of Public Health, Boston, Massachusetts (J.S.d.M.); Children's Hospital Boston and Harvard Medical School, Boston, Massachusetts (C.A.d.M.); The University of Texas M. D. Anderson Cancer Center, Houston, Texas (K.B.-E., M.W.B., L.C.); and Regional Medical and Research Specialists, Pfizer Oncology, New York, New York (A.K.).
Address correspondence and reprint requests to Janet S. de Moor, MPH, PhD, Dana Farber Cancer Institute, Center for Community Based Research, 44 Binney Street, Smith 342, Boston, MA 02115. E-mail: janet_demoor{at}dfci.harvard.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Methods: Ninety women with epithelial ovarian cancer were assessed at the start and end of chemotherapy. Optimism, distress, and HQoL were measured by self-report; CA 125 levels were gathered from patients' medical charts.
Results: Both measures of optimism were inversely associated with baseline anxiety, perceived stress, and depression. In addition, situational optimism was positively associated with baseline social and physical well-being, and dispositional optimism was positively associated with baseline social and functional well-being. However, neither measure of optimism predicted domains of distress or HQoL at the follow-up assessment after controlling for baseline levels. Dispositional optimism predicted CA 125 at the end of treatment after controlling for baseline levels. However, neither situational nor dispositional optimism predicted CA 125 falling to normal levels (
35 U/mL).
Conclusion: Consistent with prior research, optimism was inversely associated with distress and positively associated with HQoL in patients with ovarian cancer undergoing chemotherapy. Higher levels of dispositional optimism at the start of chemotherapy were associated with a greater decline in patients' CA 125 during treatment.
Key Words: optimism • anxiety • depression • perceived stress • health-related quality of life • CA 125
Abbreviations: CA = cancer antigen; HQoL = health-related quality of life; NK = natural killer; VEGF = vascular endothelial growth factor; PSS = Perceived Stress Scale; CES-D = Center for Epidemiologic Studies Depression Scale; FACT-O = Functional Assessment of Cancer Therapy-Ovarian; LOT-R = Life Orientation Test-Revised; TSOS = Treatment-Specific Optimism Scale; SES = socioeconomic status; SWB = social well-being; PWB = physical well-being; FWB = functional well-being; 
= change; SD = standard deviation; SE = standard error; 
= sample mean.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
One factor that is consistently associated with aspects of mental and physical health is optimism. Optimism has been shown to shape one's appraisal of a stressor (6–8) as well as coping resources (9) and subsequent coping efforts (6–15). Dispositional optimism, which is defined as generalized expectations that good things will happen (11), and situational optimism, which can be thought of as expectations that good things will happen in a given situation, have both been found to be positively associated with mental and physical health.
With respect to mental health, dispositional optimism was negatively associated with anxiety (7,16), depression (7,17,18), and distress (15,19–21) and positively associated with psychological adjustment during stressful life events (22–24). Studies that have focused on situation-specific optimism have also found an inverse association with distress (8) and level of depressive symptoms (25). The consistent positive association between optimism and mental health has been found for healthy (9,17,19,22), chronically ill (8,22,23), and cancer populations (7,15,20,21,25). Collectively, these findings support the importance of exploring whether optimism is protective against distress for patients with ovarian cancer during chemotherapy.
Fewer studies have examined the association between optimism and physical health. However, existing research suggests that dispositional optimism is positively associated with quality of life (13,24,26), indices of recovery (13), and physical functioning (27,28) and inversely associated with subjective symptom reports (29–31). A positive association between optimism and physical health has been found for healthy (27,29,31,32), acutely/chronically ill (28,30,33), and cancer (24,26) populations. Although few studies have explored the relationship between optimism and physical health outcomes in patients with cancer, the current literature strongly suggests that optimism may be salutary for patients with ovarian cancer undergoing chemotherapy.
Research conducted in women with ovarian cancer suggests that psychosocial factors may influence tumor-related markers. For example, Lutgendorf et al. found that aspects of social well-being, specifically support from family and friends and less distance from friends, was associated with lower levels of vascular endothelial growth factor (VEGF) (34), a cytokine that stimulates angiogenesis and acts on the vascular endothelium to enhance vascular permeability (35). They also found that hopelessness and worthlessness were related to higher levels of VEGF (34). Although Lutgendorf et al. did not examine optimism per se, these findings suggest that psychosocial factors are associated with disease-related biologic markers in women with ovarian cancer. In addition, establishing an association between psychosocial variables and VEGF is significant because VEGF levels are higher in patients with ovarian cancer who have metastases, and they are negatively associated with survival (36). Recently, it has been proposed that by increasing vascular permeability, VEGF enables the release of cancer antigen (CA) 125 into circulation (35). Therefore, examining whether optimism is associated with CA 125 is warranted.
The research summarized here supports the importance of exploring whether optimism is salutary for women with ovarian cancer who are on chemotherapy. However, to date, no studies have examined whether optimism is protective against distress and diminished health-related quality of life (HQoL) in this population. Moreover, no studies have explored the association between optimism and disease-related markers in patients with ovarian cancer. We hypothesized that: a) situational and dispositional optimism would be negatively associated with ovarian cancer patients' level of distress; b) situational and dispositional optimism would be positively associated with HQoL; c) situational and dispositional optimism at baseline would predict a decline in tumor burden, as measured by CA 125 levels, from the beginning to the end of treatment; and d) situational and dispositional optimism would predict a decline in CA 125 levels to 35 U/mL, the clinical cut point for normal CA 125. We also explored whether situational and dispositional optimism buffered the distress associated with having elevated CA 125 levels at the end of treatment. Although the current literature suggests that both situational and dispositional optimism are associated with mental and physical health outcomes, these variables are rarely measured in the context of a single study, thus prohibiting conclusions about the relative importance of each to mental and physical health. Although we expect situational and dispositional optimism to be associated with the dependent variables, an important aim of this research is to compare the relative strength of the association between each type of optimism and distress, HQoL, and CA 125. To test these hypotheses, the present study used data collected as part of an observational study of adjustment in patients with ovarian cancer.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Procedure
The timeframe of interest for the present study was a full course of chemotherapy that typically comprised six individual cycles spaced 3 weeks apart. Women who provided written informed consent were asked to complete a baseline set of questionnaires at the beginning of treatment and a follow-up set of questionnaires before their sixth (e.g., last) cycle of chemotherapy. Women who changed treatment drugs before cycle six also completed the follow-up questionnaires after six total cycles of treatment so the time between our two assessment points would be the same for all participants. Participants completed a battery of psychosocial measures, and in the present study, we examined the measures of distress, HQoL, and optimism.
Measures
Distress
Distress was measured along three dimensions, including perceived stress, anxiety, and symptoms of depression. Perceived stress was assessed with the Perceived Stress Scale (PSS), a 14-item instrument that measures respondents' stressful appraisal of their lives over the previous month (37). Internal consistency scores for the current sample were 0.83 at baseline and 0.85 at follow up. Anxiety was measured using the state anxiety subscale (STATE) of the Spielberger State/Trait Anxiety Inventory (38). The STATE is a 20-item questionnaire that measures the respondent's current level of anxiety (38). Internal consistency scores for the current sample were 0.95 at baseline and 0.94 at follow-up. Finally, symptoms of depression were measured using the Center for Epidemiologic Studies Depression Scale (CES-D), a 20-item scale that measures respondents' depressive symptomatology over the previous week (39). Internal consistency for the present sample was 0.62 at baseline and 0.71 at follow-up. Each dimension of distress was measured at baseline and follow-up.
Health-Related Quality of Life
HQoL was measured using the Functional Assessment of Cancer Therapy for patients with ovarian cancer (FACT-O) (40). The FACT-O is a 39-item scale that measure different aspects of quality of life and cancer-related concerns. In the present study, HQoL was measured with the Social Well-being, Physical Well-being, and Functional Well-being subscales to avoid overlap with our measures of distress. Internal consistency scores for the present sample ranged from 0.65 to 0.87 at baseline and 0.73 to 0.89 at follow-up. HQoL was measured at baseline and follow-up.
Optimism
Dispositional optimism was measured with the Life Orientation Test-Revised (LOT-R) (41). The LOT-R is a 10-item scale, with four filler items, that measures respondents' generalized expectations of good or bad outcomes (41). Internal consistency in the current sample was 0.71. Situational optimism was measured with the Treatment-Specific Optimism Scale (TSOS), which was developed to measure positive outcome expectations, optimistic bias, and confident emotions as they pertain to cancer treatment (25). Internal consistency in the current sample was 0.78. The LOT-R and the TSOS were correlated at r = 0.44. Each measure of optimism was measured at baseline. Although we measured both situational and dispositional optimism, they will be analyzed separately to assess their respective contribution to distress, HQoL, and change in CA 125.
Disease Progression Marker
CA 125 levels together with clinical evaluation have been shown to detect disease progression with a high degree of accuracy (42). CA 125 alone is only one measure of disease progression; however, we chose to use it because it is objective and was available for the entire sample. Because CA 125 levels in the present study ranged from 7 to 9380 U/mL at baseline and 7 to 26,660 U/mL at follow up, and the distribution of these levels was highly skewed, the logarithm base 10 of CA 125 was used in all linear regression analyses. For the logistic regression analyses, CA 125 was dichotomized according to the standard clinical cut point of 35 U/mL and >35 U/mL. CA 125 was measured at baseline and the follow-up assessments.
Sociodemographics
Age, ethnicity, and socioeconomic status (SES) were measured with a brief survey developed for this project. Fifty-three percent of the sample was retired, so education rather than annual income was chosen as the measure of SES.
Clinical Variables
Time since diagnosis, age at diagnosis, stage of disease at diagnosis, and recurrence status at the time of the survey were extracted from the patients' medical records.
Statistical Analyses
Age at diagnosis and stage of disease are prognostic factors for ovarian cancer (43,44), so we controlled for them in all multivariate models. Patients with ovarian cancer with recurrent or persistent disease after a platinum-based chemotherapy typically die from their cancer (45), so we also controlled for recurrence status in our models. To account for the length of time that study participants would have been living with their cancer, we controlled for time since diagnosis. We also controlled for patients' ethnicity and education because of literature documenting racial and class disparities in aspects of mental and physical health (46,47). Finally, for the longitudinal analyses, we also controlled for the baseline value of the dependent variable. Recurrence status (newly diagnosed, number of recurrences, or persistent disease or early recurrence), ethnicity (white or nonwhite), and education (high school or less, some college, college graduate, and graduate training) were modeled using dummy variables. Newly diagnosed recurrence status, nonwhite ethnicity, and educational attainment of high school or less served as the reference categories.
Because the sample was heterogeneous with respect to clinical and prognostic factors, which could be associated with the variables under study, a series of one-way analyses of variance were used to test whether there were any statistically significant differences in the independent and dependent variables between subgroups of patients: a) stage I and II versus stage III and IV; b) newly diagnosed versus one or more recurrences or persistent/early recurrent disease; and c) diagnosed less than 1 year before the study versus diagnosed 2 to 5 years or 5 or more years before the study.
Bivariate correlations were used to explore the univariate association among optimism, the dependent variables, and the prognostic, clinical, and sociodemographic covariates at baseline and follow up. Hierarchical linear regression was used to test whether situational and dispositional optimism were associated with patients' distress at baseline and predicted change in patients' distress during treatment. The same procedure was used to test whether optimism was associated with HQoL at baseline and test whether optimism predicted change in HQoL during treatment. Hierarchical linear regression was also used to test whether situational and dispositional optimism predicted CA 125 at the end of treatment after controlling for baseline levels, and hierarchical logistic regression was used to assess whether situational and dispositional optimism predicted a decrease in CA 125 levels to 35 U/mL. In addition, hierarchical linear regression was used to test whether situational and dispositional optimism buffered the impact of elevated CA 125 on anxiety and depression at the end of treatment.
Variables were entered into the regression models in two stages. In the first stage, prognostic (age at diagnosis and stage), clinical (recurrence status and time since diagnosis), and sociodemographic (ethnicity and education) covariates were entered, and in the second stage, a measure of optimism (or an optimism by CA 125 interaction term) was added to assess how much additional variance it explained in the dependent variable. For the longitudinal analyses, the baseline value of the dependent variable was added in the first stage. Regression diagnostics were used to verify concordance with model assumptions as well as identify the presence and influence of outliers. All analyses were conducted using SPSS 11.5 software (SPSS Inc., Chicago, IL).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
|
The women who did not complete the follow-up assessment did not differ in age, education, baseline distress, baseline HQoL, stage, time since diagnosis, percent newly diagnosed, percent with recurrent disease, or baseline CA 125 from the women who did. However, women who did not return the follow-up questionnaires were more likely to have persistent or early recurrent disease than women who did return the follow-up questionnaires (p = .04).
Although the sample was heterogeneous with respect to important prognostic and clinical factors for ovarian cancer, the sample was relatively homogeneous with respect to distress, HQoL, and optimism. Results from a series of one-way analyses of variance suggest that participants' distress, HQoL, and optimism did not differ by stage or time since diagnosis (all p values >.10). One domain of HQoL (social well-being) differed by recurrence status such that women who were newly diagnosed reported statistically significantly higher social well-being (p = .02) and those with two recurrences reported marginally significantly higher social well-being (p = .054) than women with recurrent or persistent disease. Physical well-being, functional well-being, distress, and optimism did not differ by recurrence status. CA 125 differed by stage such that women with stage I or II disease had lower CA 125 than women with stage III or IV disease (p = .01). However, CA 125 did not vary by time since diagnosis or recurrence status. Mean optimism, distress, HQoL, and CA 125 at baseline are summarized in Table 2.
|
In bivariate correlations, situational and dispositional optimism were negatively associated with anxiety, perceived stress, and depression at baseline and follow up (Table 3). In addition, situational and dispositional optimism were positively associated with social, physical, and functional well-being at baseline and dispositional optimism was positively associated with social and functional well-being at follow up. When added to a regression model containing the prognostic, clinical, and sociodemographic covariates, situational (Table 4) and dispositional (Table 5) optimism were negatively associated with each measure of distress at baseline. In addition, situational optimism was positively associated with physical and social well-being at baseline (Table 4). Likewise, dispositional optimism was positively associated with social and functional well-being and marginally associated with physical well-being at baseline (Table 5). Although optimism was consistently associated with distress and HQoL in cross-sectional analyses, neither measure of optimism predicted these variables at the follow-up assessment after controlling for baseline levels.
|
|
|
|
Situational optimism was significantly negatively correlated with CA 125 levels at follow up (r = –0.32 p = .004), and dispositional optimism was marginally negatively correlated with CA 125 levels at the follow up (r = –0.20 p = .07) (Table 3). However, when situational and dispositional optimism were added to a regression model containing the prognostic, clinical, and sociodemographic covariates and baseline CA 125, only dispositional optimism predicted CA 125 levels at follow up (Table 5). In addition, neither variable predicted CA 125 falling below the clinical cut point of 35 U/mL.
To explore whether optimism affected patients' anxiety and depression differently depending on whether they had elevated CA 125 at the end of treatment, a CA 125 (CA 125 35 U/mL vs. CA 125 >35 U/mL) by dispositional or situational optimism interaction term was entered into a hierarchical linear regression model to predict anxiety and depression at the end of treatment. Neither the CA 125 by situational optimism interaction term nor the CA 125 by dispositional optimism interaction term was associated with anxiety or depression at the end of treatment (Tables 4 and 5).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Although we found a cross-sectional association between both measures of optimism and distress at baseline, optimism failed to predict distress at the follow-up assessment after controlling for baseline levels. One possible explanation for the observed lack of association between optimism and follow-up distress was the relative consistency in this variable across the study. The baseline and follow-up values of anxiety, perceived stress, and depression were highly correlated (Table 3); therefore, the majority of variance in follow-up distress was explained by baseline distress.
Optimism also failed to predict HQoL at the follow-up assessment after controlling for baseline levels possibly as a result of the relative consistency in HQoL during the study. Baseline and follow-up physical, functional, and social well-being were highly correlated (Table 3); therefore, much of the variance in follow-up HQoL was explained by the baseline HQoL.
Dispositional optimism at the start of treatment predicted CA 125 at the follow-up assessment after controlling for baseline levels such that the higher the level of dispositional optimism the greater the decline in CA 125 during chemotherapy. Although some women knew their approximate level of CA 125 at the time of the follow-up survey, the results would not be influenced by this because optimism was measured at baseline and was not influenced by women's experience with chemotherapy or their CA 125 levels at follow up. Optimism failed to predict a decline in CA 125 to normal levels; however, these findings may still be clinically significant. The rate and magnitude of decline in CA 125 levels during chemotherapy has been used to predict the likelihood of remission and survival in patients with ovarian cancer (48). Research suggests that a CA 125 half-life of less than 25 days was associated with longer survival time (49). In addition, a CA 125 of 65 U/mL or less or a decrease of 50% in CA 125 from prechemotherapy levels after four to six courses of chemotherapy was associated with achieving complete remission, and CA 125 of 65 U/mL or less within 4 weeks after the second course of chemotherapy was associated with the best prognosis (50). We were not able to explore the trajectory of CA 125 during treatment because we only collected data at the beginning and end of treatment. However, additional research on the relationship between optimism and CA 125 is justifiable given our findings and the work of Lutgendorf et al., who found an inverse association between social well-being and VEGF among women with ovarian cancer (34).
As indicated in Tables 4 and 5, both measures of optimism were associated with distress and HQoL at baseline and dispositional optimism predicted CA 125 at the end of treatment. However, neither measure of optimism consistently explained more of the variance in the dependent variables than the other. The greatest differences were seen in the models to predict perceived stress and functional well-being in which dispositional optimism explained more of the variance than situational optimism.
Study limitations should be noted. The greatest limitation of the present study was the heterogeneous sample with respect to stage, recurrence status, and time since diagnosis; however, we controlled for these variables in our analysis. In addition, results from a series of one-way analyses of variance suggested that optimism, distress, HQoL, and CA 125 at baseline were fairly homogeneous across different subgroup of patients. The study was also limited by a small amount of missing data across several of the study variables and 13% missing data for education, which was one of our covariates in the regression analyses. Consequently, fewer women were included in the regression models than actually participated in the study. However, it is important to note that we explored several approaches to impute the missing study variables (e.g., mean substitution and missing data dummy variables). After using these approaches, the results did not differ from the complete case analysis, which suggests that our findings are robust.
This study also had several important strengths. Women with ovarian cancer have not been widely studied, and the present study makes an important contribution to our knowledge of this population of patients with cancer. The design of our study allowed us to test both cross-sectional and longitudinal associations between the independent and dependent variables. Moreover, the breadth of data that we collected made it possible to assess the connection between psychosocial variables and objective measures of disease progression. Finally, this study raises challenging questions about the importance of optimism to patients' mental health, HQoL, and CA 125 during chemotherapy.
|
NOTES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
This research was supported by The Ovarian Cancer Research Program, M. D. Anderson Cancer Center, the Bettyann Asche Murry Fund for Research in Gynecological Medical Oncology, and NCI R25 57730, P.I., Robert M. Chamberlain, PhD.
DOI:10.1097/01.psy.0000222379.71389.91
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  G. Alspach Extending the Tradition of Giving Thanks Recognizing the Health Benefits of Gratitude Crit. Care Nurse, December 1, 2009; 29(6): 12 - 18. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
