SEARCH

SEARCH BY CITATION

Keywords:

  • adaptation;
  • advanced cancer;
  • long-term survivorship;
  • psychospiritual well being

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. Conflict of Interest Disclosures
  8. References

BACKGROUND:

With improved treatment, increasing proportions of patients with advanced cancer are surviving longer with their disease: into a second year after diagnosis and beyond. These longer term survivors face continuing challenges in selecting and shifting personal life goals and goals of care over years (rather than months) of life with incurable cancer. Studies are needed to explore adaptation over time in patients who are living longer term with late-stage cancer, including anxiety, depression, and spiritual well being, which are conceptualized as indicators of psychospiritual well being in patients with advanced cancer.

METHODS:

Psychospiritual well being and adaptation were explored in a study of middle-aged and older patients with advanced cancer (N = 142) who survived into a second year after diagnosis and were assessed in interviews across 4 time points. Examining patterns of adaptation over time called for in depth analytical techniques to identify variation in key outcome trajectories. General growth mixture modeling was used to explore heterogeneity in adaptation using a multivariate parallel model of anxiety, depression. and spiritual well being.

RESULTS:

Modeling revealed 3 distinct group trajectories of psychospiritual well being and adaptation (low-worsening, moderate-improving, and high-stable). Age and education were correlated with group membership. Advanced cancer survivors who were older and had more years of education were more likely to be members of the high-stable group in psychospiritual adaptation throughout the study.

CONCLUSIONS:

The current findings suggested that psychospiritual adaptation, as measured in this study, is not uniform but is characterized by heterogeneous trajectories. The results contribute to the development of better hypotheses regarding the processes of adaptation in longer term survivors with advanced cancer and to the identification of potential subgroups at greatest risk for poor outcomes. Cancer 2009;115(18 suppl):4298–310. © 2009 American Cancer Society.

Adaptation in cancer survivorship has received growing attention because of the increasing numbers and proportions of cancer patients surviving long-term after diagnosis and active treatment for their disease.1 Long-term survivorship has been defined as living for ≥5 years after a cancer diagnosis.2 Historically, the great majority of long-term survivors were diagnosed with early stage cancers. A major focus of survivorship research has been latent and long-term effects of the disease and/or its treatment, with special attention to implications for quality of life over time.3-5 With new discoveries and access to treatments, patients with advanced cancer also are living longer.6-8 Survivorship in advanced cancer covers a more compressed period, and we consider longer term survivors as those who live into a second year after the diagnosis of advanced cancer and potentially beyond.

The majority of studies on patient quality of life and adaptation in advanced cancer have focused on the early treatment phase (within the first year after diagnosis) and at the end of life rather than on longer term survivorship. Most observational and intervention studies assess outcomes for the initial months or through the first 6 to 9 months after the diagnosis of advanced cancer (see, eg, Rose et al,9 Brown et al10). However, longer term survivors face continuing challenges in selecting and shifting personal life goals and goals of care over years, rather than months, of extended life with incurable cancer. The few studies that have assessed adaptation into a second year of survival with advanced cancer generally reported on small samples of primarily middle-aged, middle-class, and Caucasian individuals. One such study from Australia examined patterns of adaptation in patients with late-stage melanoma (N = 44) who survived over a 2-year period. Although measures of mood and coping fluctuated, the authors observed little systematic change over time.11 Studies with larger sample sizes are needed to identify potential patterns of adaptation. especially among older longer term survivors with advanced cancer, who constitute the majority.5, 10, 12, 13 Attention should be given to low-income and underserved populations, in which individuals more frequently have advanced disease at diagnosis and for whom personal and resource burdens may be greatest.14-16

Anxiety and depression are key concepts in assessing psychological distress, and clinical studies consistently have reported heightened levels of both among individuals who are diagnosed with cancer.17-19 These findings have informed the National Comprehensive Cancer Network practice guidelines for the management of psychosocial distress, which recommend screening for anxiety and depression in all patients with newly diagnosed cancer.20, 21 Follow-up assessments also have been recommended, because higher levels of anxiety and depression may persist or develop, especially among patients who are diagnosed with late-stage cancer. Most psycho-oncology interventions for patients with advanced cancer attempt to reduce levels of anxiety and/or depression over a period of weeks or months after diagnosis.17 However, studies are needed that assess these outcomes over a longer period to inform the development of interventions for those at greatest risk for poor coping and adaptation in longer term survivorship with late-stage cancer.

Spiritual well being is an additional key factor in effective coping and adaptation among patients with advanced cancer from the early treatment phase through longer-term survivorship and end of life.10, 22-24 The National Consensus Project for Quality Palliative Care describes spiritual well being as 1 of the 8 evidence-based clinical practice domains in guidelines to ensure the quality and consistency of palliative care.25, 26 A growing number of psycho-oncology interventions designed to reduce distress also address patients' spiritual needs, primarily in the early treatment phase and at end of life (see, eg, Jacobson et al,17 Balboni et al23). An integrative review of the literature conducted by Lin and Bauer-Wu identified “psychospiritual well being” as a unique focus area of research on coping and adaptive processes in patients with advanced cancer.27 In the current study, we assessed psychospiritual well being in measures of anxiety, depression, and spiritual well being and explored patterns of psychospiritual adaptation from the early treatment phase into a second year, or longer term survivorship, in patients with incurable, late-stage cancer.

Previous literature and research have linked several sociodemographic variables to the psychosocial and spiritual well being of patients with advanced cancer, including patient age, race, and education. Generally, older patients10, 13, 28 and more highly educated patients14 report lower levels of anxiety and depression, and African Americans give greater importance to spirituality in coping with advanced cancer and end of life.23, 29 Consequently, it is important to assess associations between these characteristics and patterns of psychospiritual adaptation in patients who survive longer term with advanced cancer.

In the current study, we examined psychospiritual well being as parallel processes in levels of anxiety, depression, and spiritual well being assessed across 4 time points over a 12-month period. Age, race, and education were analyzed in association with patterns of psychospiritual adaptation in patients who were living longer term with advanced cancer from the early treatment phase into a second year after an advanced cancer diagnosis. This research is especially timely given recent Institute of Medicine and National Cancer Institute reports highlighting the importance of meeting patients' needs for psychosocial care30 and health communication31 from cancer diagnosis through long-term survivorship.30-32

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. Conflict of Interest Disclosures
  8. References

Participants

The sample consisted of 142 patients diagnosed with cancer who survived into a second year after diagnosis of late-stage cancer. All patients were enrolled in a randomized controlled trial to test the effects of a coping and communication support intervention for patients with advanced cancer,16, 33 and they received their care at the MetroHealth Medical Center (MHMC) or the Veterans Affairs Medical Center (VAMC) in Cleveland, Ohio. Inclusion criteria were age ≥40 years, recent diagnosis of advanced-stage cancer, not yet referred to hospice, cognitively intact, and able to speak English. On the basis of disease stage (ie, stage IV cancers; stage III lung, pancreatic, or liver cancers) or an oncologist's prognostic estimate equivalent to stage IV for hematologic/lymphatic cancers, all patients had a median life expectancy of ≤1 year. Approximately 71% of eligible patients were enrolled in the study (N = 559), and recruitment occurred on average 9 weeks to 10 weeks after the diagnosis of late-stage cancer. There were multiple types of cancer represented, including lung cancer (37%), gastrointestinal cancer (23%), genitourinary cancer (10%), head and neck cancer (10%), breast cancer (7%), reproductive cancer (5%) hematologic cancer (3%), sarcomas (3%), and unknown cancers (2%). To date, 45% of enrolled patients have survived through 4 waves of data collection (baseline, 3 months, 6 months, and 12 months postenrollment). In the current report, we describe these patients as longer term survivors with advanced cancer. All such survivors in this study completed the 12-month telephone interview, which was conducted approximately 15 months after diagnosis.

Measurements

In baseline interviews, we collected information on patients' sociodemographic characteristics, including age, sex, race, years of formal education, marital status, religious affiliation, income (both median census tract [ie, “neighborhood”] income and self-reported personal annual income), and insurance type. Key measures of psychospiritual adaptation (anxiety, depression, and spiritual well being) were assessed at baseline and in follow-up telephone interviews at 3 months, 6 months, and 12 months after enrollment. The Profile of Mood States (POMS) Short Form measures of depressed mood and anxiety were used in this study.34 Depressed mood was measured on an 8-item scale, had a score that ranged from 0 to 32, and had an alpha reliability (α) of .93. Anxiety was measured by 6 items, had a score that ranged from 0 to 24, and had an α of .91. The 12-item Functional Assessment of Chronic Illness Therapy-Spiritual Well Being scale (FACIT-Sp)35 was used to measure spiritual well being. It had a score that ranged from 0 to 48 and had an α reliability of .87.

Analytic Approach

The analytic goal of the current analysis was to determine whether a single typology or multiple typologies best explained trajectories for anxiety, depressed mood, and spiritual well being. In this study, we used general growth mixture modeling (GGMM).36, 37 This analytic method uses an individual-centered approach, which focuses on identifying subgroups of individuals by emphasizing individual differences and similarities in change over time.38 Unlike traditional growth modeling, which assumes that all individuals are drawn from a single population with common population parameters, GGMM tests whether the population is composed of discrete typologies (classes) of trajectories by using a categorical latent variable. GGMM estimates the random effects (initial level and rate of change) the same way as the conventional latent growth modeling techniques but allows individuals in different trajectory groups to vary around different mean growth curves. In addition, GGMM allows testing for within-group differences in the growth parameters in cases for which it is not reasonable to assume that all individuals within a specific group follow the same trajectory over time.

We used Mplus version 5.139 (Mplus, Los Angeles, Calif) to examine the presence of heterogeneous trajectories when considering anxiety, depressed mood, and spiritual well being simultaneously. The GGMM approach allows for the estimation of 1) a mixture growth model in which the different trajectories of anxiety, depressed mood, and spiritual well being are represented jointly and simultaneously by class-varying means, and 2) the probability of class membership as a function of covariates. In GGMM, the latent classes explain the relation among the outcomes, similar to a factor analysis. However, in GGMM, instead of the individuals receiving a factor score that best explains their observed outcomes, the outcomes are combined in an analysis that classifies individuals within a categorical latent variable.40, 41

Model Selection

To specify our final model, first, we explored changes in the shape of each psychospiritual process separately (anxiety, depressed mood, and spiritual well being) assuming that all patients exhibited the same trajectory over time. We did this because each process represents different (although closely linked) dimensions of adaptation. After we specified the overall trajectory of each process, we estimated several latent classes of the model in which each class had its own growth factors (intercepts and rate of change). It should be noted that individuals within each latent class shared the same growth factors and had similar growth curve patterns. In addition, because the classes were estimated by the simultaneous combination of anxiety, depressed mood, and spiritual well being, the individuals within each class remained the same across outcomes. After the classes were estimated, based on each individual's classification probability, the average trajectory for each of the 3 groups was described across the outcomes.

By using fit indices and interpretability, we identified the optimal number of distinct trajectories and estimated the final model combining all outcomes and covariates. Figure 1 presents the final version of the model. The extent to which the latent classes are defined clearly can be evaluated by the average posterior probabilities for each individual. With these posterior probabilities, a table can be created that presents the average conditional probabilities of being a member of a class. A probability close to 1 represents good classification. After deciding on the appropriate number of classes based on fit indices, we used multinomial logistic regression to explore demographic characteristics associated with specific trajectory group membership.

thumbnail image

Figure 1. This chart illustrates the multivariate growth mixture model of psychospiritual adaptation. SPWB indicates spiritual well being.

Download figure to PowerPoint

It is important to note that our sample only included advanced cancer survivors who remained enrolled in the study through the 12-month interview (approximately 15 months after diagnosis). The number of patients who voluntarily withdrew from the study was extremely low (n = 12) and comprised only 2% of the total enrolled; therefore, attrition issues, which are important in longitudinal designs, were not problematic to these specific analyses.

The software program we used, Mplus 5.1, uses the principle of maximum likelihood estimation with robust standard errors. The process of appropriately estimating mixture models involved a thorough evaluation of each solution (because of the complexity of the models). An important element in the estimation process was to carefully evaluate each run for local solutions.42 We did that by trying different starting values for each model by overriding the defaults and re-estimating the models multiple times to replicate the log-likelihood.

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. Conflict of Interest Disclosures
  8. References

Baseline Characteristics of Longer Term Survivors With Advanced Cancer

In total, 142 patients survived into a second year postdiagnosis of advanced-stage cancer. Key demographics of the study patients are shown in Table 1. The mean age of longer term survivors was 61.7 years, 28% were women, and the average years of education were 12.6 years. Thirty-seven percent of patients were African American, 60% were white, and 3% were other. The majority of patients had incomes below the poverty line and were living in low-income neighborhoods (median income, $37,407). The mean POMS depressed mood and anxiety scores were 7.3 and 7.9, respectively. The mean FACIT-Sp spirituality score was 28.5. Longer term survivors with advanced cancer did not differ significantly from short-term survivors (patients who lived for <6 months after a late-stage diagnosis; n = 90) on baseline demographic or psychospiritual adjustment variables. A smaller percentage of longer term survivors versus short-term survivors had lung cancer or unknown primary cancers, as expected from previous research.

Table 1. Baseline Characteristics of Longer Term Survivors With Advanced Cancer
VariableNo. (%)Mean ± SDMedianRange
  1. POMS indicates Profile of Mood States; MHMC, MetroHealth Medical Center; FACIT-Sp, Functional Assessment of Chronic Illness Therapy spiritual well being scale.

Age, y 61.7 ± 10.4 40-84
Women40 (28)   
African Americans53 (37)   
Years of education 12.6 ± 2.6 3-21
Median neighborhood income, $000  37.4076.336-83.343
POMS anxiety score 7.9 ± 6.3 0-24
POMS depressed mood score 7.3 ± 7.7 0-32
FACIT-Sp score 28.5 ± 7.5 0-48
Adjuster variables
 Site of care MHMC75 (53)   
 In intervention group71 (50)   
 Weeks from diagnosis  to enrollment 9.6 ± 10.2 0-51

Table 1 also reports data on the 3 adjuster variables that were included in our analytical models. Fifty-three percent of longer term survivors had been recruited in the MHMC ambulatory cancer clinic, and 47% had been recruited at the VAMC ambulatory cancer clinic. Both of these systems provide care for disadvantaged and underserved populations. Fifty percent of longer term survivors had been randomized to the intervention arm of the study. The mean number of weeks from diagnosis to enrollment in this sample was 9.6 weeks (standard deviation, 10.2 weeks). The wide range in weeks is explained by the small number of patients enrolled in the initial year of study implementation during initial screening for patients who had been diagnosed within the past year.

Patterns of Psychospiritual Adaptation in Longer Term Survivors

One-class analysis

Observed and estimated means of anxiety, depressed mood, and spiritual well being based on a traditional latent growth model across the 4 time points (baseline, 3 months, 6 months, and 12 months) are shown in Figure 2. To adequately capture change for anxiety and depression, we used 3 growth factors (an intercept, a linear slope, and a quadratic slope); for spiritual well being, we used 2 growth factors (an intercept and a linear slope). For anxiety and depressed mood, the overall pattern of change was a linear decrease from baseline to the 6-month time point followed by a nonlinear change from 6 months to 12 months after baseline. For spiritual well being, there was a small increase in the observed mean from baseline to the 12-month time point that was explained adequately without including a second order factor. After evaluating several comparison models using best-fit criteria, we selected a nonlinear model for anxiety and depressed mood and a linear model for spiritual well being. Then, these models were used to determine the number of latent groups in the sample.

thumbnail image

Figure 2. These graphs illustrate the observed and estimated means of anxiety, depressed mood, and spiritual well being (SPWB) based on the conventional latent growth curve model.

Download figure to PowerPoint

Class differentiation

Identifying the correct number of classes43 was an iterative process, which we began by examining a model that contained a single class (homogeneity hypothesis analysis) followed by models that contained 2, 3, and 4 classes. The 1-class, 2-class, and 4-class models were rejected in favor of the 3-class model. On the basis of fit indices, the 1-class and 2-class models underestimated the number of classes, whereas the 4-class model overestimated the number of classes (Table 2). The 3-class model had the smallest values for all 3 fit indices (Bayesian information criterion, sample-size–adjusted Bayesian information criterion, and Akaike information criterion) and an entropy (a measure of classification quality from 0 to 1 in which 1 indicates perfect classification). The adoption of a 3-class model was supported by all fit indices, by very good classification quality, and by the ability for meaningful interpretation of trajectories.

Table 2. Fit Indices for Latent Class Solutions
 Index* 
No. of ClassesBICSSABICAICEntropy
  • BIC indicates Bayesian information criterion; SSABIC, sample-size–adjusted Bayesian information criterion; AIC, Akaike information criterion.

  • *

    Smaller numbers represent better fit for all indices.

  • Entropy refers to accuracy of classification, with higher values representing better classification.

19922.769834.179840.00 
29783.879666.809674.500.979
39689.559550.519559.670.936
49781.929633.219642.990.889

The average posterior probabilities of class membership are presented in Table 3. A posterior probability that is high in 1 group and a low in the others indicates good classification. For example, for the poor-worsening group, the posterior probability of .973 indicates that the patients have a very high probability of membership in this class.

Table 3. Average Posterior Probabilities of Class Membership
 Posterior P
ClassPoor- WorseningModerate- ImprovingHigh- Stable
Low-worsening.973.027.000
Moderate-improving.026.972.002
High-stable.000.007.993

In addition to using several fit criteria to assess models with a different number of classes, we evaluated models on the basis of substantive results. For example, models that were too complex and did not provide important information were discarded for models that had theoretical relevance and parsimony. Model evaluation is performed in part by log-likelihood value comparison using the chi-square test statistic for the fit of nested models.44 The model with regressions of class on covariates (log-likelihood, −4539.7; n = 58 parameters) was significantly different only from no covariates (log-likelihood, −4892; n = 28 parameters). Additional specifications of the model (growth factors regressed on covariates in addition to regression of class on covariates and regression of growth factor on covariates only) did not significantly improve the model.

On the basis of our study sample, we identified 3 distinct trajectory groups that can be described as follows: 1) an increasing anxiety and depressed mood with a low spiritual well being trajectory (low-worsening; 14 patients; 10% of the sample); 2) a rapidly decreasing, depressed mood and anxiety with an increasing spiritual well being trajectory (moderate-improving; 42 patients; 30% of the sample); and 3) a low and slightly decreasing anxiety and depressed mood trajectory with high levels of spiritual well being (high-stable; 86 patients; 60% of the sample). The solutions of the optimal model for each outcome are presented graphically in Figure 3, and the estimated means and standard errors of the growth factor parameters are presented in Table 4.

thumbnail image

Figure 3. These graphs illustrate the model-estimated means for the 3 class growth mixture solutions for anxiety, depressed mood, and spiritual well being (SPWB).

Download figure to PowerPoint

Table 4. Parameter Estimates for the Three-Class Model
 InterceptLinear SlopeQuadratic Slope
Psychospiritual Trajectory GroupEstSEPEstSEPEstSEP
  1. Est indicates estimate; SE, standard error.

Anxiety
 Low-worsening13.571.56.0000.860.42.042Fixed at 0  
 Moderate-improving12.410.92.000−3.641.25.0040.590.27.030
 High-stable4.520.78.000−2.290.65.0000.490.13.000
Depressed mood
 Low-worsening16.292.04.0001.540.58.008Fixed at 0  
 Moderate-improving12.811.48.000−3.271.37.0180.480.51.343
 High-stable3.441.16.003−1.861.16.1100.420.24.078
Spiritual well being
 Low-worsening22.731.64.000−0.181.10.865 
 Moderate-improving25.981.19.0000.570.27.036 
 High-stable31.640.92.000−0.000.18.969 

Examination of between-class comparisons among the growth factors using the Wald chi-square (χ2) test of parameter equalities indicated that there were significant differences in the initial level and rate of change within each outcome. The significant differences between groups for anxiety were as follows: high-stable versus moderate-improving intercept, χ2(1,N = 142) = 37.36 (P < .001); high-stable versus low-worsening intercept, χ2(1,N = 142) = 29.66 (P < .001); high-stable versus low-worsening slope, χ2(1,N = 142) = 14.48 (P < .001); and moderate-improving versus low-worsening, χ2(1,N = 142) = 9.73 (P < .01). For depressed mood, the significant differences were as follows: high-stable versus moderate-improving intercept, χ2(1,N = 142) = 16.81 (P < .001); high-stable versus low-worsening intercept, χ2(1,N = 142) = 33.19 (P < .001); and high-stable versus low-worsening slope, χ2(1,N = 142) = 7.12 (P < .01). For spiritual well being, the significant differences were as follows: high-stable versus moderate-improving intercept, χ2(1,N = 142) = 10.45 (P < .001); and high-stable versus low-worsening intercept, χ2(1,N = 142) = 23.28 (P < .001). There were no significant differences in the models that had equality constraints in the slopes of spiritual well being.

Differences in Age and Education

Next, we examined the relation between patient demographics (age, race, and education) and psychospiritual trajectory groups. Results from this analysis are shown in Table 5. Patients in the high-stable trajectory group were more likely to be older (odds ratio [OR], 1.09) and to have more years of education (OR, 1.39) than patients in the moderate-improving group. No significant differences were observed when the low-worsening trajectory group was compared with either the moderate-improving group or the high-stable group; however, the study had low power to detect a relation because of the small sample size.

Table 5. Adjusted Odds Ratios and 95% Confidence Intervals for Baseline Predictors Between 3 Trajectory Groups
 Adjusted OR (95% CI)
Predictors measured at baselineModerate-Improving vs Low-WorseningHigh-Stable vs Moderate-ImprovingHigh-Stable vs Low-Worsening*
  • OR indicates odds ratio; CI, confidence interval.

  • *

    Note: There were 14 patients in the low-worsening group, 42 patients in the moderate-improving group, and 86 patients in the high-stable group.

  • P < .05.

  • P < .001.

Age, per0.94 (0.88-1.01)1.09 (1.02-1.15)1.03 (0.95-1.11)
Race: White vs African American1.49 (0.47-4.68)0.98 (0.38-2.561.47 (0.42-5.10)
Education, per y0.90 (0.72-1.13)1.39(1.15-1.66)1.26 (0.96-1.63)

The logistic regression plots presented in Figure 4 illustrate the effects of age and education on the probability distribution of the 3 psychospiritual adaptation groups. The probability of being in the moderate-improving group was much higher for younger and less educated patients. This probability decreases as age and education increase. The converse is true for the high-stable group; as age and education increase, so does the probability of belonging to the high-stable group. These findings suggest a qualitative distinction between the moderate-improving and high-stable groups that is associated with age and education. Those who entered the study when they were middle-aged and less educated were more likely to have higher baseline anxiety and depressed mood and more moderate levels of spiritual well being than those who were older and more educated at baseline. Neither age nor education was related to the probability of being in the low-worsening group.

thumbnail image

Figure 4. These graphs illustrate multinomial logistic regression of class regressed on age and education.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. Conflict of Interest Disclosures
  8. References

The primary objective of this study was to explore patterns of psychospiritual adaptation in middle-aged and older longer term survivors with incurable, late-stage cancer from the early treatment phase into the second year after diagnosis. We sought to identify heterogeneity in trajectories of psychospiritual well being27 and adaptation by considering patterns of anxiety, depressed mood, and spiritual well being across 4 time points over a 12-month period. Our results indicate that longer term survivors with advanced cancer can be described as belonging to distinctly different groups based on the trajectories of anxiety, depressed mood, and spiritual well being. By using general growth mixture models, we identified 3 groups of longer term cancer survivors with low-worsening, moderate-improving, and high-stable psychospiritual well being. To our knowledge, this is the first time that these groups have been identified in survivors with advanced cancer. This is especially important in identifying a group of advanced cancer survivors (10%) who may be at greatest risk for poor psychospiritual adaptation over time.

Our analyses demonstrated that the identification of advanced cancer patients who are at risk for increasing levels of anxiety and depressed mood and decreasing spiritual well being is not achieved well by using a univariate measure. Change in quality-of-life outcomes over time is a multivariate problem and is understood most effectively and appropriately within this context. The current results demonstrate that identifying the high-risk group can be problematic if only the initial levels of anxiety and depressed mood are included. With the inclusion of spiritual well being in the model, the high-risk group of longer term survivors can be identified by presenting the lowest levels of spiritual well being.

Although our identification of 3 groups is unique in the population under study and the methodologies used, these groupings are supported in concept by the work of Blank and Bellizzi in a different cancer context.13, 45, 46 Those authors assessed adaptation in 509 middle-aged and older men who were diagnosed with early stage prostate cancer. Survivors in that study had been diagnosed between 1 year and 8 years before data collection. On the basis of the men's responses to open-ended questions about how prostate cancer had changed their lives, the authors conceptualized 3 categories of effect (no change, positive change, and negative change) and reported representation of these effects across age groups (ages 47-59 years, 60-69 years, and ≥70 years).13, 46 Similar to our findings, negative change was observed in the smallest percentage (10%-13%) and was not associated with age. Furthermore, whereas more survivors in the older age groups reported no change, a higher percentage of survivors in the middle-aged group (ages 47-59 years) reported a positive change in their lives. Although that study was strikingly different in its design and in the population assessed,13, 45, 46 the results provide important insights into the interpretation of findings on psychospiritual adaptation in middle-aged and older long-term survivors with advanced cancer.

In our descriptive analyses, we observed that the majority of advanced cancer survivors (60%) were likely to be members of the group with high-stable psychospiritual well being over time. This group reported lower levels of anxiety and depressed mood and higher levels of spiritual well being in the early treatment phase and maintained this higher level of psychospiritual well being into the second year of survivorship. Although the moderate-improving group (30%) started with a level of psychospiritual well being that was only slightly better than the (high-risk) low-worsening group, this group exhibited marked improvement over time.

The identification of these 3 distinct groups provides potential new insights into observed variation in patients' preferences for support and engagement in psycho-oncology interventions and underscores the need for interventions that are tailored to patients' individual needs and preferences for support over time (see, eg, Rose et al,16 Radziewicz et al33). Patients in these groups may perceive a need for different degrees of contact and support from practitioners and may focus on different problems both initially and over time. For example, patients in the low-worsening group may need more frequent and ongoing support in coping with more serious psychological or existential issues and in managing emotional and physical symptoms.

Patients in the high-stable group may be least likely to perceive the need for additional support in coping either in the early treatment phase or over time, in longer term survivorship. It will be important to examine the support needs and preferences of patients in this group, which constituted the majority of patients in our study. The finding that patients with advanced cancer who have high-stable psychospiritual well being were more likely to be older and to have more years of formal education is consistent with previous studies that assessed anxiety, depressed mood, and spiritual well being in cancer patients.10, 13, 23 Even so, it cannot be assumed that all older or more educated patients in this group would not seek or use supportive interventions. We previously reported that older patients with advanced cancer largely were similar to middle-aged patients in their preferences for or actual engagement in a coping and communication support intervention, at least in the early treatment phase for late-stage cancer.16

In contrast to the high-stable group, the identification of a unique group of longer term survivors with low-worsening psychospiritual well being emphasizes the need for early detection of patients who are at risk for poor outcomes from the outset and over time. Patients with advanced cancer in this group may need continuous access and follow-up support to assist with coping and adaptation over time. It is likely that these patients were the most vulnerable in disposition and coping resources9, 13, 47 before their cancer diagnosis and that the added stress and burden over the long term increasingly became overwhelming, perhaps especially in the absence of support. Covariates, including patient age, race, and years of education, were not associated with classification in the low-worsening group; because they constituted only 10% of our sample, it is clear that future analyses with larger sample sizes will be needed to better characterize this at-risk group.

The identification of patients with moderate-improving psychospiritual well being is a potentially important new discovery in describing adaptation over time with late-stage cancer. Compared with the high-stable group, survivors in the moderate-improving group were more likely to be younger, ie, middle-aged, and less educated. It is unclear the extent to which survivors in this group either 1) initially experienced heightened distress and then regained previous levels of psychospiritual well being over time; or 2) uniquely adapted to the challenges of long-term survivorship, perhaps surpassing levels of well being before they had cancer (see, eg, Blank and Bellizzi,13 Bellizzi48). In long-term survivorship with advanced cancer, coping and adaptation often necessitate clarifying and shifting life goals and goals of care while simultaneously modifying strategies for optimization with compensation from the early treatment phase through end of life.13, 16, 49, 50 Survivors with moderate-improving psychospiritual well being, the majority of whom were middle-aged and less educated in the current study, may benefit most from psycho-oncology interventions that are individualized and supportive of such efforts to adapt over time.

The findings in these analyses are a stimulant for longer longitudinal designs and suggest possible directions for future research. Despite the need of further studies to validate these results, we have demonstrated that longer term survivors are described best as belonging to qualitatively different groups, at least in the variables that represent psychospiritual well being. Furthermore, the findings suggest that these groups vary not only in the outcomes examined but also in association with important patient attributes. Older and more educated survivors are more likely to be in the high-stable group given the significant relation of age and education.

Our analyses have several important limitations. The generalizability of results is a consideration, because the patients with advanced cancer in this study were part of a randomized controlled trial that was testing a coping and communication support intervention over time. The great majority of patients were economically disadvantaged, and approximately 33% were African Americans, although race did not prove to be associated with any of the 3 trajectories for psychospiritual well being. Whether these results would apply to other patient groups is not clear. In these analyses, we included longer term survivors only according to our definition (living into a second year and completing a 12-month interview, on average, 15 months after diagnosis), and we did not test other possible covariates in our models. Our study sample was relatively small, and we may have lacked power to detect other important differences. Future analyses of this type would benefit from larger sample sizes and extended follow-up with patients who are living long term with advanced cancer. Finally, our study represents an exploratory analysis; replication and validation of the model is needed in different populations and clinical settings.

Despite these limitations, our results open new avenues of study for optimizing psychospiritual adaptation among longer term cancer survivors. It is our hope that these findings will contribute to understanding the processes of adaptation and identification of potential subgroups of longer term survivors of late-stage cancer who are at the greatest risk for poor psychospiritual adaptation. These findings also raise questions about different group trajectories along other important dimensions, including decision-making preferences and care practices, among patients living longer term with advanced cancer. The current findings may inform the development of tailored interventions with variable degrees of intensity and frequency for longer term survivors who are treated in ambulatory clinics that provide care for disadvantaged and underserved populations.

Acknowledgements

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. Conflict of Interest Disclosures
  8. References

We acknowledge the contributions of research support staff on this project (Steven Lewis, Mary Ellen Lawless, Nasim Seifi, Kathryn Engelhardt, and Mary Hutchinson).

Conflict of Interest Disclosures

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. Conflict of Interest Disclosures
  8. References

Supported by grants from the National Cancer Institute (R01-CA10282), the Department of Veterans Affairs Health Services Research and Development (Merit: HR-03-255), and the American Cancer Society (ROG-04-090-01).

References

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. Conflict of Interest Disclosures
  8. References
  • 1
    Rowland JH. Survivorship research: past, present, and future. In: GanzPA, ed. Cancer Survivorship: Today and Tomorrow. New York, NY: Springer; 2007: 28-42.
  • 2
    HewittM, GreenfieldS, StovallE, eds. From Cancer Patient to Cancer Survivor: Lost in Transition. Washington, DC: National Academies Press; 2006.
  • 3
    Stein KD, Syrjala KL, Andrykowski MA. Physical and psychological long-term and late effects of cancer. Cancer. 2008; 112( 11 suppl ): 2577-2592.
  • 4
    Rowland JH, Hewitt M, Ganz P. Cancer survivorship: a new challenge in delivering quality cancer care. J Clin Oncol. 2006; 24: 5101-5114.
  • 5
    Rowland JH, Yancik R. Cancer survivorship: the interface of aging, comorbidity and quality of life. J Natl Cancer Inst. 2006; 98: 504-505.
  • 6
    Sugimura H, Yang P. Long-term survivorship in lung cancer: a review. Chest. 2006; 129: 1088-1097.
  • 7
    Dow KH. Seventh National Conference on Cancer Nursing Research keynote address: challenges and opportunities in cancer survivorship research. Oncol Nurs Forum. 2003; 30: 455-469.
  • 8
    Fisher PS. Survivorship. Clin J Oncol Nurs. 2006; 10: 93-98.
  • 9
    Newsom JT, Knapp JE, Schultz R. Longitudinal analysis of specific domains of internal control and depressive symptoms in patients with recurrent cancer. Health Psychol. 1996; 15: 323-331.
  • 10
    Rose JH, O'Toole E, Einstadter D, Shenko C, Love T, Dawson N. Patient age, well-being, care perspectives and practices in the early treatment phase for late-stage cancer. J Gerontol A Biol Sci Med Sci. 2008; 63: 960-968.
  • 11
    Brown JE, King MT, Butow PN, Dunn SM, Coates AS. Patterns over time in quality of life, coping and psychological adjustment in late stage melanoma patients: an application of multilevel models. Qual Life Res. 2000; 9: 75-85.
  • 12
    Rose JH, O'Toole EE, Dawson NV, et al. Perspectives, preferences, care practices and outcomes among older and middle-aged patients with late-stage cancer. J Clin Oncol. 2004; 22: 4907-4917.
  • 13
    Blank TO, Bellizzi KM. A gerontological perspective on cancer and aging: multiple trajectories, multiple influences, and differential outcomes. Cancer. 2008; 112: 2569-2576.
  • 14
    Barbareschi G, Sanderman R, Tuinstra J, van Sonderen E, Ranchor AV. A prospective study on educational level and adaptation to cancer, within 1 year after the diagnosis, in an older population. Psycho-Oncol. 2008; 17: 373-382.
  • 15
    Wardle J, McCaffery K, Nadel M, Atkin W. Socio-economic differences in cancer screening participation: comparing cognitive and psychosocial explanations. Soc Sci Med. 2004; 59: 249-261.
  • 16
    Rose JH, Radziewicz R, Bowman KF, O'Toole EE. A coping and communication support intervention tailored to older patients diagnosed with late stage cancer. Clin Interv Aging. 2008; 3: 77-95.
  • 17
    Jacobson PB, Donovan Z, Swaine Z, Watson I. Management of anxiety and depression in adult cancer patients: toward an evidence-based approach. In: ChangA, GanzP, HayesT, et al, eds. Oncology: An Evidence-Based Approach. New York, NY: Springer-Verlag; 2006: 1552-1579.
  • 18
    Massie MJ. Prevalence of depression in patients with cancer: a critical review. J Natl Cancer Inst Monogr. 2004; 32: 57-71.
  • 19
    Stark DPH, House A. Anxiety in cancer patients. Br J Cancer. 2000; 83: 1261-1267.
  • 20
    [No authors listed] NCCN practice guidelines for the management of psychosocial distress. National Comprehensive Cancer Network. Oncology (Williston Park). 1999; 13( 5A): 113-147.
  • 21
    National Comprehensive Cancer Network. Clinical practice guidelines in oncology: distress management v.1.2009. Available at: http://www.nccn.org/Professionals/Physicians_gls/PDF/distress.pdf Accessed on November 30, 2008.
  • 22
    Brady MJ, Peterman AH, Fitchett G, Mo M, Cella D. A case for including spirituality in quality of life measurement in oncology. Psycho-Oncol. 1999; 8: 417-428.
  • 23
    Balboni TA, Vanderwerker LC, Block SD, et al. Religiousness and spiritual support among advanced cancer patients and associations with end-of-life treatment preferences and quality of life. J Clin Oncol. 2007; 25: 555-560.
  • 24
    Tarakeshwar N, Vanderwerker LC, Paulk ME, et al. Religious coping is associated with the quality of life of patients with advanced cancer. J Palliat Med. 2006; 9: 646-657.
  • 25
    National Consensus Project. Clinical Practice Guidelines for Palliative Care. Available at: http://www.nationalconsensusproject.org. Accessed December 20, 2008.
  • 26
    Farrell BR. Overview of the domains of variables relevant to end-of-life care. J Palliat Med. 2005; 8( suppl 1): S22-S29.
  • 27
    Lin HR, Bauer-Wu SM. Psycho-spiritual well-being in patients with advanced cancer: an integrative review of the literature. J Adv Nurs. 2003; 44: 69-80.
  • 28
    Keating N, Norrendam M, Landrum M, Huskamp H, Meara E. Physical and mental health status of older long-term cancer survivors. J Am Geriatr Soc. 2005; 53: 2145-2152.
  • 29
    Johnson KS, Kuchibhatla M, Tulsky JA. What explains racial differences in the use of advance directives and attitudes toward hospice care? J Am Geriatr Soc. 2008; 56: 1953-1958.
  • 30
    Institute of Medicine (IOM). Cancer Care for the Whole Patient: Meeting Psychosocial Health Needs. Washington, DC: National Academies Press; 2007.
  • 31
    Epstein RM, Street RL. Patient-Centered Communication in Cancer Care: Promoting Healing and Reducing Suffering. Bethesda, Md: National Cancer Institute; 2007.
  • 32
    Institute of Medicine/National Research Council (IOM/RRC). From Cancer Patient to Cancer Survivor: Lost in Transition. Washington, DC: National Academies Press; 2006.
  • 33
    Radziewicz R, Rose JH, Bowman KF, Berila R, Spuckler A, O'Toole EE. Establishing treatment fidelity in a coping and communication support telephone intervention for aging advanced cancer patients and their family caregivers. Cancer Nurs. 2009; 32: 193-202.
  • 34
    Shacham S. A shortened version of the Profile of Mood States. J Pers. 1983; 47: 305-306.
  • 35
    Peterman AH, Fitchett G, Brady MJ, Hernandez L, Cella D. Measuring spiritual well-being in people with cancer: the Functional Assessment of Chronic Illness Therapy—Spiritual Well-being Scale (FACIT-Sp). Ann Behav Med. 2002; 24: 49-58.
  • 36
    Muthen B, Shedden K. Finite mixture modeling outcomes using the EM algorithm. Biometrics. 1999; 55: 463-469.
  • 37
    Muthen B, Muthen LK. Integrating person-centered and variable-centered analyses: growth mixture modeling with latent trajectory classes. Alcohol Clin Exp Res. 2000; 24: 882-891.
  • 38
    Magnusson, D. The logic and implications of a person-oriented approach. In: CairnsRB, BergmanR, KaganJ, eds. Methods and Models for Studying the Individual. Thousand Oaks, Calif: Sage Publications; 1998: 33-64.
  • 39
    Muthen LK, Muthen BO. Mplus User's Guide, 5th ed. Los Angeles, Calif: Muthen & Muthen; 1998-2007.
  • 40
    Muthen B. Latent variable analysis: growth mixture modeling and related techniques or longitudinal data. In: KaplanD, ed. Handbook of Quantitative Methodology for the Social Sciences. Newbury Park, Calif: Sage Publications; 2004: 345-368.
  • 41
    Muthen B. Latent variable mixture modeling. In: MarcoulidesGA, SchumackerRE, eds. Advanced Structural Equation Modeling: New Developments and Techniques. Mahwah, NJ: L. Erlbaum Associates; 2000: 1-33.
  • 42
    Hipp JR, Bauer DJ. Local solutions in the estimation of growth mixture models. Psychol Methods. 2006; 11: 36-53.
  • 43
    Nylund K, Asparouhov T, Muthen BO. Deciding on the number of classes in latent class analysis and growth mixture modeling: a Monte Carlo simulation study. Structural Equation Modeling. 2007; 14: 535-569.
  • 44
    Muthen L, Muthen B. Chi-square difference testing using the S-B scaled chi-square. Available at: http://www.statmodel.com/chidiff.shtml. Accessed December 3, 2008.
  • 45
    Blank TO, Bellizzi KM. After prostate cancer: predictors of well-being among long-term prostate cancer survivors. Cancer. 2006; 106: 2128-2135.
  • 46
    Blank TO, Bellizzi K, Murphy K, Ryan K. How do men “make sense” of their prostate cancer? Age and treatment factors [abstract]. Gerontologist. 2003; 43( special issue 1): 342-343.
  • 47
    Miller DL, Manne SL, Taylor K, Keates J, Dougherty J. Psychological distress and well-being in advanced cancer: the effects of optimism and coping. J Clin Psychol Med Settings. 1996; 3: 115-130.
  • 48
    Bellizzi KM. Expressions of generativity and post-traumatic growth in adult cancer survivors. Int J Aging Hum Dev. 2004; 58: 267-287.
  • 49
    Baltes PB, Baltes MM. Psychological perspectives on successful aging: the model of selective optimization with compensation. In: BaltesPB, BaltesMM, eds. Successful Aging: Perspectives From the Behavioral Sciences. Cambridge, Mass: Cambridge University Press; 1990: 1-34.
  • 50
    Baltes MM, Carstensen LL. Social-psychological theories and their applications to aging: from individual to collective. In: BengstonVL, SchaieKW, eds. Handbook of Theories of Aging. New York, NY: Springer Publishing; 1999: 209-226.