Fax: (404) 303-9520
Adult cancer survivors: How are they faring?†
Article first published online: 28 OCT 2005
Copyright © 2005 American Cancer Society
Supplement: Cancer Survivorship: Resilience Across the Lifespan
Volume 104, Issue Supplement 11, pages 2565–2576, 1 December 2005
How to Cite
Baker, F., Denniston, M., Smith, T. and West, M. M. (2005), Adult cancer survivors: How are they faring?. Cancer, 104: 2565–2576. doi: 10.1002/cncr.21488
Adapted from a paper presented at Cancer Survivorship: Resilience Across the Lifespan, Washington, DC, June 2–4, 2002.
- Issue published online: 18 NOV 2005
- Article first published online: 28 OCT 2005
- Manuscript Accepted: 13 JUL 2005
- Manuscript Received: 16 JUN 2005
- American Cancer Society
- psychosocial problems;
- Cancer Problems in Living Scale (CPILS);
- prostate carcinoma;
- lung carcinoma;
- breast carcinoma;
- colorectal carcinoma
This study identified the psychosocial problems that 752 patients from 3 states who had been diagnosed with 1 of the 10 most commonly occurring cancers indicated concerned them the most. Approximately 1 year after being diagnosed with cancer, 68.1% of patients were concerned with their illness returning, and more than half were concerned with developing a disease recurrence (59.8%) or had fears regarding the future (57.7%). In addition to these psychological problems focused on fear, approximately two-thirds (67.1%) of patients were concerned about a physical health problem, fatigue, and loss of strength. Two other physical health problems that concerned more than two-fifths of patients were sleep difficulties (47.9%) and sexual dysfunction (41.2%). More problems were reported by younger survivors (ages 18–54 yrs), women, nonwhites, those who were not married, and those with a household income of less than $20,000 a year. Those patients currently in treatment for cancer reported on average significantly more problems (P < 0.001) and on average had a higher Cancer Problems in Living Scale (CPILS) total score (P < 0.001) compared with those not currently in treatment. In a comparison of respondents with one of the four most common cancers, the most concerns regarding problems in living and highest mean CPILS scores were reported by those diagnosed with lung cancer, followed by survivors of breast cancer, colorectal cancer, and prostate cancer. Cancer 2005. © 2005 American Cancer Society.
Using as a definition of a cancer survivor anyone who has been diagnosed with cancer from the time of the diagnosis through the balance of their life, an estimated 10.1 million cancer survivors currently are alive in the U.S.1 As more people survive cancer, greater attention is being given to quality-of-life outcomes and how individuals adapt to this chronic disease. As public health programs move from their past concentration on treatment, early detection, and the prevention of cancer to an expanded research focus on survivorship and quality of life, there is a greater need for research to identify the physical, psychosocial, and economic factors affecting the well-being of cancer survivors.
Some studies have found that cancer survivors experience high levels of physical, emotional, and social problems relative to the general population2 and that these symptoms can persist for some time after treatment has ended, sometimes for decades.3 However, in general, studies have found that a cancer survivor's quality of life typically returns to a level comparable to that of individuals with no history of cancer soon after treatment completion.3–6 Although the psychosocial functioning of the majority of cancer survivors has not been found to differ significantly from that of the general population, research suggests that a notable percentage of survivors do continue to experience elevated levels of emotional distress.7 Anxiety, depression, fears of disease recurrence, and concerns regarding passing the disease on to their offspring surface as persistent emotional and psychologic issues of cancer survivors. Cancer survivors also have reported symptoms consistent with posttraumatic stress disorder (PTSD), such as avoidance of cancer-related thoughts, reexperiencing negative experiences, and increased arousal.8–10 At least one study indicated that between 5–10% of cancer survivors would be likely to meet Diagnostic and Statistical Manual of Mental Disorders—Fourth Edition (DSM–IV) PTSD diagnostic criteria.11
Most of the studies assessing psychosocial functioning after the completion of cancer treatment have focused on breast cancer survivors, although some studies have included mixed samples or have focused on survivors of specific cancer sites other than the breast (see Kornblith12 for a summary of cancer survivor studies by diagnosis). To our knowledge, few studies to date have compared psychosocial sequelae across a number of different cancer sites.
Sexual problems and issues of fertility have been found to be most prevalent among survivors of breast cancer,13, 14 testicular cancer,15, 16 and prostate cancer.17 Some survivors have reported cognitive dysfunction after treatment, which appears to be related to high-dose chemotherapy.18–20
Cancer survivors also face a host of economic, financial, insurance, and employment issues after the completion of their treatment and the transition from cancer patient to cancer survivor. Reports of job discrimination and concerns about changing jobs due to fear of the loss of health insurance and medical care have been documented.21–28 Van Tulder et al.5 compared employment and insurance issues among survivors of Hodgkin disease and a group of hospital volunteers. Compared with the hospital volunteers, the cancer survivors reported working fewer hours and having greater difficulty obtaining life insurance, although they did not demonstrate significant differences in employment or obtaining health insurance.
The research concerning adult cancer survivors has been limited not only by its focus on a restricted number of cancer sites but also by being based primarily on patients who come to academic centers for diagnosis and treatment. Therefore, the samples employed are not representative of many population groups who for various reasons are not treated at these facilities. A key perspective missing in the majority of this research is the absence of longitudinal studies that follow survivors with repeated assessments to determine how they fare over time.29 Although research on the psychosocial status of pediatric oncology patients has employed repeated measures designs, the past research regarding adult cancer patients has more often been cross-sectional rather than longitudinal. The patient report data in this research literature also has tended to be limited by dependence on retrospective questions. In recent years, the scope of research on adult cancer survivors has been broadening, so that there are more prospective, longitudinal studies covering more cancer sites and employing population-based samples.
The data presented in this article were collected during the implementation pilot study for the Study of Cancer Survivors-I (SCS-I), a national, population-based, prospective, longitudinal investigation of the needs and quality of life of adult survivors of the 10 most common types of cancer, which was conducted by the Behavioral Research Center of the American Cancer Society. This research effort was undertaken to overcome methodologic weaknesses that had limited the usefulness and generalizability of earlier studies, including the use of relatively small, unrepresentative samples of cancer survivors primarily focusing on those who were treated at academic cancer centers. This project was developed as part of a program of research that represented an attempt to develop the largest, most diverse studies of both recent and long-term survivors and their family members.
This implementation pilot followed a questionnaire pilot study that tested a battery of instruments that measured demographic characteristics, treatment, physical health, symptoms, personal and environmental resource variables, and quality of life. The refined questionnaire that resulted from this questionnaire pilot also included a measure of problems in living that offered respondents an opportunity to rate a variety of physical, psychological, and community reintegration problems with which they might be having difficulty. The current study presents results from the administration of this measure of psychosocial morbidity to the large number of cancer survivors included in the SCS-I implementation pilot and focuses on the psychosocial need of cancer survivors at one point in time (i.e., approximately 1 year after diagnosis).
Two questions will be addressed in the current study:
- 1What problems do adult cancer survivors experience in dealing with their cancer and its treatment 1 year after diagnosis?
- 2How do demographic characteristics and cancer type relate to ratings of psychosocial problems?
Pilot study design
The SCS-I begins in the first year of survivorship and extends into the tenth year. The study procedures have been reviewed and approved by the Emory University Institutional Review Board (IRB) as well as by IRBs from public health departments and universities in the states in which data were collected.
In selecting a sample of cancer patients, researchers have an advantage not available for other chronic diseases in that cancer is a reportable disease and most states maintain a registry of people diagnosed with cancer. State and Surveillance, Epidemiology, and End Results (SEER) (a registry support program of the National Cancer Institute) cancer registries participating in SCS-I identify, sample, and invite subjects to take part in the study. The SCS-I sample in each state includes U.S. adults diagnosed within a specified 12–month period with 1 of the 10 most common cancers (prostate, female breast, lung, colorectal, bladder, non-Hodgkin lymphoma, skin melanoma, kidney, ovarian, and uterine). To be eligible for SCS-I, participants have to be at least 18 years of age at the time of diagnosis and be able to provide written, informed consent prior to entry into the study. Cases are stratified by type of cancer and age, and also race/ethnicity in states in which there is sufficient racial/ethnic diversity. Subjects who consent to participate in the study are mailed questionnaires 4 times: at approximately 1 year, 2 years, 5 years, and 10 years after diagnosis.
To refine the study's materials and procedures, an implementation pilot for SCS-I was conducted in Iowa, Minnesota, and Georgia. In Minnesota, a sample was selected from the state cancer registry. For sample selection in Iowa, the researchers contracted with the SEER registry at the University of Iowa. In Georgia, because the Emory University SEER program covered only Atlanta and a few rural counties, an alternative strategy using a two-stage cluster sample design was employed to obtain a sample that was more representative of the entire state. Specifically, the state was divided into four regions based on geography and urban status, and a stratified random sample of cancer-reporting facilities was selected with probability proportional to size sampling. The researchers sent medical abstractors to invite the selected facilities to participate in the study and to ascertain cases for sampling. Physician and patient informed consent was obtained, and baseline questionnaires were sent to consented patients approximately 1 year after diagnosis.
Initially, 5605 subjects were selected for study participation. Of these, 818 were found to be deceased and 200 were found to be ineligible at the time of physician or patient consent. Because it was not always known at what point in the consent process a sampled individual was discovered to be deceased or ineligible, for the purposes of calculating consent rates, the assumption was made that half were discovered at the time of physician consent and half at the time of patient consent, resulting in an estimate of 5096 (5605 –409- 100) live/eligible cases at the time of physician consent. Physician consent was not obtained for 444 cases, leaving 4652 cases for whom physician consent was obtained, for a physician consent rate of live/eligible cases of 91.3% (4652 of 5096 eligible cases). Subtracting 409 of the cases found to be deceased and 100 found to be ineligible that were estimated to have been discovered at the time of patient consent from the 4652 cases for whom physician consent to contact was not denied results in 4143 cases. At the time patient consent was sought, 876 declined to be included in the study, 690 could not be located, and 592 did not respond to either mail or telephone requests for consent, yielding an overall patient consent rate of 47.9% (1985 of 4143 cases) for live/eligible cases with physician consent.
Due to various problems encountered in implementing the pilot study, the time since diagnosis for these respondents varied from 3 months to 29 months. Because the baseline data collection point for the study is intended to be approximately 1 year past diagnosis, only those cases who were within 1 year ± 4 months of the diagnosis that made them eligible for the study were included in this analysis.
Table 1 presents the demographic characteristics of the 752 respondents who met the time from diagnosis criteria. Consistent with a plan to oversample cancer patients who were younger than 55 years at the time of diagnosis, half of the sample was comprised of individuals who were ages 18–54 years at the time of diagnosis. Somewhat more than half of the cases (59.3%) were female. The sample was 11.7% nonwhite and less than 1% Hispanic. The respondents were nearly equally divided between those who were residents of Iowa, Minnesota, and Georgia at the time of their diagnosis. Nearly 75% of the sample were currently married or living with a partner. Approximately 26% of the respondents had at least a college education, and 46% reported a high school education or less as their highest level of education. Only approximately one-fifth of the sample had a yearly household income of less than $20,000, and nearly 15% reported a yearly household income of at least $75,000.
|Ages 18–54 yrs at diagnosis||50.0|
|State of residence at the time of diagnosis|
|Current marital status|
|Single, never married||5.1|
|< High school graduate||13.0|
|High school graduate/GED||33.0|
|Some college/vocational school||28.1|
|≥ College graduate||25.9|
|Prefer not to answer||9.6|
The medical characteristics of these respondents are shown in Table 2. Although the sample of survivors originally were selected so as to oversample those types of cancer that had higher rates of mortality, those survivors who completed a questionnaire at baseline and who met the time from diagnosis criteria set for this analysis did not follow that pattern. In fact, the three cancer sites with the best survival rates had the highest numbers of respondents. One-fifth of respondents (20.5%) reported having been diagnosed with cancer more than once, and approximately the same proportion (21.0%) were currently receiving treatment. The mean time since diagnosis for this sample was 1.03 years, with a standard deviation (SD) of 0.19 years.
|Type of cancer reported||No.||Percentage|
|Melanoma of the skin||83||11.1|
|Diagnosed more than once||154||20.5|
|Currently in treatment||158||21.0|
|Time since diagnosis (yrs)||752||1.03 (SD of 0.19)|
Among other measures, the baseline questionnaire included the Cancer Problems in Living Scale (CPILS), a 29-item scale asking respondents to rate each of the problems listed by choosing one of the following three possible responses: “Not a problem for me” (scored as 0), “Somewhat of a problem for me” (scored as 1), or “A severe problem for me” (scored as 2). This scale offers a way of assessing the problem burden experienced by the cancer survivor at various times and has been shown to be a useful measurement for assessing the needs of the cancer survivor as she or he deals with cancer and its treatment.30, 31 Two additional items were added to the CPILS for SCS-I: “Not being able to get the information I need on cancer” and “Problems communicating with my spouse or partner. ” At baseline, the CPILS asked how much of a problem each item had been since the patient was diagnosed with cancer.
Because only a purposively selected subset of the data were used in these analyses, the data were analyzed as a convenience sample using SPSS statistical software (SPSS Inc., Chicago, IL), and no attempt was made to generate population estimates, although the original sample was population-based. In reporting results from the CPILS, the responses were dichotomized so that the percentages reported in the current study are for those respondents who indicated that the item had been “somewhat of a problem” or a “severe problem” for them. Chi-square tests were performed to compare the frequency of problem occurrence by selected demographic and medical characteristics, and odds ratios were obtained for these comparisons. The CPILS total score was calculated by using the previously described 0 to 2 scoring of items. CPILS total scores were calculated only for those cases with missing data for no more than five items. The total number of problems reported was simply a count of the number of items that a respondent had reported as having been a problem for them.
Medical characteristics by cancer type reported are shown in Table 3. Note that only those who reported a single type of cancer are included in the analyses for this table. Although the mean time since diagnosis did not vary much across the 10 cancers (1.00 year to 1.06 years), the proportion of respondents currently receiving treatment (from 8.6% for skin melanoma to 49.0% for female breast cancer) and the proportion that reported having been diagnosed with cancer more than once (from 5.9% for uterine cancer to 32.6% for urinary bladder cancer) varied considerably by cancer type.
|Type of cancer reported||No.a||Mean time since diagnosis (SD)||Currently receiving treatment (%)||Diagnosed more than once (%)|
|Female breast||100||1.01 (0.19)||49.0||22.0|
|Melanoma of the skin||70||1.02 (0.20)||8.6||20.0|
As shown in Table 4, the mean number of problems reported was 9.4 (SD of 7.1), and the mean CPILS total score was 11.4 (SD of 10.0). Younger survivors, female survivors, survivors who were not currently married or living with a partner, and survivors with a lower annual income were found to have statistically significantly higher means for both the number of problems reported and the CPILS total score. Nonwhite survivors had a statistically significantly higher mean than white survivors for CPILS total score but not for the number of problems reported. No statistically significant differences in the mean number of problems reported or the mean CPILS total score were found when comparing survivors with more than a high school education with those with less education. Those respondents currently receiving treatment for cancer reported on average significantly more problems (P < 0.001) and had on average a higher CPILS total score (P < 0.001) than those who were not currently receiving treatment.
|Characteristics||No. of problems||CPILS total score|
|No.||Mean (SD)||No.||Mean (SD)|
|All survivors||744||9.4 (7.1)||732||11.4 (10.0)|
|Age in yrs|
|18–54||353||11.1 (7.2)a||351||13.4 (10.4)a|
|≥ 55||349||7.7 (6.6)||339||9.1 (8.8)|
|Female||420||10.4 (7.1)a||413||12.5 (10.0)b|
|Male||291||9.8 (9.8)||286||8.1 (7.0)|
|White||658||9.2 (7.0)||654||11.0 (9.8)c|
|Non-white||86||10.8 (8.2)||78||13.9 (7.5)|
|Married/marriage-like||548||9.0 (6.8)b||545||10.7 (9.7)b|
|Not married||193||10.8 (7.1)||184||13.4 (10.8)|
|< $20,000/yr||134||11.4 (8.1)b||127||15.1 (12.5)b|
|≥ $20,000/yr||520||9.2 (7.0)||518||10.8 (9.4)|
|Highest educational level achieved|
|≤ HS graduate||341||9.4 (7.2)||332||11.6 (10.5)|
|> HS graduate||398||9.4 (7.1)||395||11.2 (6.3)|
|Currently receiving cancer treatment|
|Yes||158||11.6 (7.4)a||156||14.6 (11.4)a|
|No||487||8.7 (6.9)||480||10.3 (9.3)|
Table 5 shows the overall occurrence of these 31 problems in living for this sample of survivors. Note that the order of the items on this table is by occurrence and is not the same as the order in which they appear on the questionnaire.
|Item||Percentage identifying the item as a problem|
|Feeling fearful that my illness will return||68.1|
|Fatigue, loss of strength||67.1|
|Concern about relapsing||59.8|
|Fears about the future||57.7|
|Being less physically able to have sexual intercourse||41.2|
|Having difficulty in making long-term plans||39.7|
|Uncomfortable with changes in my physical appearance||39.6|
|Preoccupation with being ill||38.5|
|Continued major problems with my health||36.5|
|Diminished ability to concentrate||35.7|
|Difficulty in meeting my medical expenses||28.3|
|Difficulty in returning to former roles||26.1|
|Being less able to provide for the financial needs of my family||25.6|
|Problems communicating with my spouse or partner||20.1|
|Being concerned about infection and crowds||19.9|
|Difficulty in obtaining adequate insurance||16.5|
|Difficulty in pursuing the career of my choice||16.3|
|Not being able to change jobs for fear of losing my health insurance coverage||15.4|
|Being treated as different from others||15.1|
|Problems with family/children||14.1|
|Not being able to get the information I need about cancer||12.8|
|Concern about being physically unable to have children||4.7|
Another way to examine how problems differed for selected demographic groups is to identify how individual problems vary in occurrence across these groups. Tables 6–8 show the odds ratio for particular groups for problems whose occurrence was related to age (ages 18–54 yrs vs. age ≥ 55 yrs), gender, race (white vs. nonwhite), marital status (married/marriage-like vs. all other), household income (< $20,000/yr vs. ≥ $20,000/yr), and highest educational level achieved (≤ high school graduate vs. > high school education) at a significance level of P < 0.01. The odds ratio indicates how much more likely one group is to have reported a particular problem than the group with which it is being compared. This level of significance was chosen because multiple tests have been performed. For example, the odds ratio of 4.2 for “concern about being physically unable to have children” means that survivors ages 18–54 years were more than 4 times as likely to report this problem as those age 55 years or older, whereas the odds ratio of 0.5 for “ being less physically able to have sexual intercourse” means that women were half as likely as men to report this problem. Problems were more often reported among those who were younger at the time of diagnosis, were female, were nonwhite, were not married or in a marriage-like relationship, had a lower household income, or had higher education. The only item for which this was not the case was sexual dysfunction, which was reported more often by men and by those who were currently married or living with a partner.
|Problems that differed by age (OR for ages 18–54 yrs)||OR (95% CI)|
|Concern about being physically unable to have children||4.2 (1.66–10.5)|
|Difficulty pursuing the career of my choice||2.5 (1.76–3.63)|
|Unable to change jobs for fear of losing health insurance||2.5 (1.71–3.51)|
|Feeling angry||1.9 (1.54–2.35)|
|Being less able to provide for my family's financial needs||1.9 (1.48–2.41)|
|Job discrimination||1.7 (1.10–2.60)|
|Problems with family or children||1.7 (1.27–2.38)|
|Not being able to get the information I need about cancer||1.7 (1.22–2.33)|
|Concern about relapsing||1.6 (1.38–1.86)|
|Fears about the future||1.6 (1.35–1.81)|
|Being treated as different from others||1.6 (1.17–2.06)|
|Problems communicating with spouse or partner||1.6 (1.25–2.07)|
|Guilt feelings||1.5 (1.22–1.94)|
|Fearful that my illness will return||1.5 (1.27–1.70)|
|Feeling vulnerable||1.5 (1.26–1.80)|
|Feeling isolated||1.4 (1.12–1.70)|
|Feeling helpless||1.4 (1.16–1.66)|
|Preoccupation with being ill||1.4 (1.18–1.67)|
|Diminished ability to concentrate||1.3 (1.06–1.49)|
|Uncomfortable with changes in my physical appearance||1.3 (1.13–1.58)|
|Difficulty in meeting my medical expenses||1.3 (1.11–1.63)|
|Problems that differed by gender (OR for females)||OR (95% CI)|
|Problems with family or children||2.2 (1.48–3.31)|
|Being concerned about infection and crowds||1.8 (1.32–2.40)|
|Uncomfortable with changes in my physical appearance||1.7 (1.36–2.07)|
|Fears about the future||1.7 (1.39–1.98)|
|Fatigue, loss of strength||1.6 (1.32–1.86)|
|Sleep difficulties||1.5 (1.22–1.77)|
|Being fearful that my illness will return||1.5 (1.23–1.75)|
|Diminished ability to concentrate||1.4 (1.12–1.69)|
|Eating difficulties||1.4 (1.10–1.84)|
|Feeling dependent||1.4 (1.16–1.80)|
|Feeling vulnerable||1.3 (1.08–1.60)|
|Difficulty in making long-term plans||1.3 (1.07–1.57)|
|Concern about disease recurrence||1.3 (1.10–1.58)|
|Feeling angry||1.3 (1.08–1.63)|
|Being less physically able to have sexual intercourse||0.5 (0.43–0.62)|
|Problems that differed by race (OR for nonwhites)||OR (95% CI)|
|Difficulty in obtaining adequate insurance||1.2 (1.08–1.32)|
|Difficulty in pursuing the career of my choice||1.2 (1.10–1.37)|
|Difficulty in meeting my medical expenses||1.1 (1.04–1.19)|
|Difficulty in returning to former roles||1.1 (1.04–1.21)|
|Being concerned about infection and crowds||1.1 (1.05–1.24)|
|Problems that differed by marital status (OR for not married)||OR (95% CI)|
|Concern about being physically unable to have children||2.0 (1.31–3.08)|
|Difficulty in meeting my medical expenses||1.3 (1.14–1.44)|
|Uncomfortable with changes in my physical appearance||1.2 (1.05–1.27)|
|Feeling dependent||1.2 (1.10–1.36)|
|Eating difficulties||1.2 (1.04–1.33)|
|Feeling isolated||1.2 (1.10–1.40)|
|Feeling helpless||1.2 (1.04–1.27)|
|Difficulty in returning to former roles||1.2 (1.09–1.36)|
|Difficulty in obtaining adequate insurance||1.2 (1.06–1.42)|
|Difficulty in pursuing the career of my choice||1.2 (1.04–1.37)|
|Continued major problems with my health||1.1 (1.03–1.25)|
|Difficulty making long-term plans||1.1 (1.04–1.24)|
|Being less physically able to have sexual intercourse||0.9 (0.79–0.93)|
|Problems that differed by level of education achieved (OR for > high school)||OR (95% CI)|
|Concern about being physically unable to have children||2.0 (1.07–3.63)|
|Unable to change jobs for fear of losing my insurance||1.5 (1.12–1.97)|
|Problems that differed by household income (OR for < $20,000/yr)||OR (95% CI)|
|Difficulty in obtaining adequate insurance||1.4 (1.22–1.68)|
|Difficulty in meeting my medical expenses||1.3 (1.16–1.44)|
|Difficulties pursuing the career of my choice||1.3 (1.09–1.45)|
|Feeling dependent||1.2 (1.05–1.28)|
|Being less able to provide for my family's financial needs||1.2 (1.06–1.30)|
|Feeling helpless||1.2 (1.08–1.31)|
|Difficulty in returning to former roles||1.2 (1.09–1.34)|
|Continued major problems with my health||1.2 (1.08–1.30)|
|Fatigue, loss of strength||1.1 (1.03–1.20)|
Table 9 shows the problems whose occurrence was related to current treatment status at the 0.01 level of significance. Problems that had a significant association with treatment status were reported more often by those respondents currently receiving treatment than by those not currently receiving treatment for cancer.
|Problems that differ by treatment status (OR for currently in treatment for cancer)||Odds Ratio (95% CI)|
|Concern about infection and crowds||1.9 (1.44, 2.48)|
|Continued major problems with my health||1.9 (1.41, 2.43)|
|Fatigue, loss of strength||1.8 (1.24, 2.45)|
|Preoccupied with being ill||1.7 (1.34, 2.29)|
|Sleep difficulties||1.7 (1.31, 2.29)|
|Feeling dependent||1.7 (1.33, 2.28)|
|Having difficulty in making long term plans||1.7 (1.28, 2.21)|
|Being treated as different from others||1.6 (1.15, 2.14)|
|Feeling fearful that my illness will return||1.6 (1.17, 2.30)|
|Concern about relapsing||1.5 (1.10, 2.02)|
|Fears about the future||1.5 (1.11, 1.98)|
|Feeling vulnerable||1.5 (1.15, 2.25)|
|Difficulty in returning to former roles||1.5 (1.12, 1.96)|
|Diminished ability to concentrate||1.4 (1.10, 1.90)|
Multiple linear regression modeling was used to determine which of the variables shown in Table 4 were independently predictive of the number of problems reported and the CPILS total score. In addition to the variables shown in Table 4, the Medical Outcomes Study: 36-Item Short Form Health Survey (MOS SF-36) General Health score and the MOS SF–36 Physical Function score also were included as independent variables in the modeling. The demographic variables—age, gender, race, marital status, education, and income—were first entered into the model, and the treatment status and MOS SF-36 scores were then allowed to enter stepwise. Table 10 shows the regression analysis results for the variables that were found to be significant in the final model for each of the two dependent variables. The final models had adjusted R2 values of 0.428 (number of problems reported) and 0.437 (CPILS total score). The same four variables—age, gender, the MOS SF-36 General Health score, and the MOS SF-36 Physical Function score—were found to be significant in the final model for both the number of problems reported and the CPILS total score.
|Significant variables||Dependent = number of problems reported Model adjusted R2 = 0.428|
|Beta coefficient||SE||t||P value|
|Age < 54 yrs at diagnosis||3.481||0.533||6.526||0.000|
|MOS SF-36 General Health score||−0.264||0.027||−9.612||0.000|
|MOS SF-36 Physical Function score||−0.122||0.029||−4.171||0.000|
|Dependent = CPILS total score Model adjusted R2 = 0.437|
|Significant variables||Beta coefficient||SE||t||P value|
|Age < 54 yrs at diagnosis||4.676||0.738||6.332||0.000|
|MOS SF-36 General Health score||−0.345||0.038||−9.123||0.000|
|MOS SF-36 Physical Function score||−0.212||0.040||−5.242||0.000|
Table 11 shows the mean number of problems reported and the mean CPILS total scores for survivors of the four most commonly occurring cancers: female breast, colorectal, prostate, and lung. These analyses were restricted to those who reported having been diagnosed with only one type of cancer. The most concerns about problems in living were reported by those respondents diagnosed with lung cancer, followed by survivors of breast cancer, colorectal cancer, and prostate cancer. One-way analysis of variance tests indicated differences in means that were statistically significantly different across survivors of these four cancers for both measures. Post hoc testing yielded the following results: 1) prostate cancer survivors had significantly lower means for both the number of problems reported and CPILS total score compared with survivors of breast or lung cancer; 2) breast cancer survivors differed significantly with regard to both measures only from prostate cancer survivors; 3) colorectal cancer survivors differed significantly from lung cancer survivors with regard to the mean CPILS total score but not with regard to the mean number of problems reported; and 4) lung cancer survivors had a significantly higher mean CPILS total score than did prostate or colorectal cancer survivors but differed significantly with regard to the mean number of problems reported from only the prostate survivors. All other pairwise comparisons were not found to be statistically significant.
|Type of cancer||No. of problemsa||CPILS Total Scorea|
|No.||Mean (SD)||No.||Mean (SD)|
|Female breast||99||10.4 (7.3)||97||12.3 (9.7)|
|Colorectal||57||8.6 (6.8)||57||10.4 (9.4)|
|Lung||64||11.7 (7.4)||63||15.0 (11.5)|
|Prostate||84||6.6 (6.1)||82||7.9 (7.9)|
Table 12 shows the overall top 10 problems and gives their ranking and prevalence among survivors of the 4 most common cancers. The top 10 problems for each of the 4 cancers did not appear to differ markedly from the top 10 problems for the overall group, although there was at least 1 problem in the overall top 10 that did not occur in the top 10 for each of the 4 cancers separately.
|All cancers||Breast cancer||Colorectal cancer||Prostate cancer||Lung cancer|
|Problem and overall rank||Rank||% with problem||Rank||% with problem||Rank||% with problem||Rank||% with problem|
|1. Fearful my illness will return||2||70.7||1||67.9||2||48.8||2||74.2|
|2. Fatigue, loss of strength||1||79.8||3||60.7||4||43.2||1||79.7|
|3. Concern about relapsing||4||58.8||2||63.6||3||46.8||4||61.3|
|4. Fears about the future||3||63.6||4||58.9||5||37.8||3||62.5|
|5. Sleep difficulties||5||55.6||a||6||36.9||6||58.7|
|6. Sexual dysfunction||a||a||1||74.1||a|
|7. Difficulty making long-term plans||7||53.1||5||38.6||a||5||59.4|
|8. Uncomfortable with changes in my physical appearance||6||37.9||8b||35.1||a||10||46.9|
|9. Preoccupied with being ill||10||40.8||6||37.5||9||25.9||7||54.7|
|10. Feeling vulnerable||8||45.9||a||10||25.6||a|
The diagnosis of cancer usually leads to rapid treatment of the disease, involving surgery, radiation therapy, chemotherapy, and other less common treatments or combinations of these interventions. Even after the threat of cancer is diminished through successful treatment, the patient is likely to remain concerned about recurrence of this life-threatening illness.32 Notwithstanding that modern treatment methods have resulted in cancer patients living longer, the future remains uncertain and the patient remains fearful of disease recurrence. Lee-Jones et al.33 recommended obtaining more data to estimate the extent of these concerns among cancer patients. This study found that after surviving for 1 year past diagnosis, more than two-thirds of cancer patients were concerned with their illness returning, nearly 60% were concerned about disease recurrence, and nearly 58% were fearful about their future. This finding is consistent with the description by Muzzin et al., who characterized the continuing fear that a cancer survivor feels as a sword of Damocles that continues to hang over the patient and family for the rest of the person's life.34
At 1 year past diagnosis, in addition to the previously mentioned psychologic problems focusing on fear of the future, approximately two-thirds of the survivors in the current study reported that they experienced physical problems including fatigue and loss of strength (67.1%), and nearly half said they had sleep difficulties (47.1%), whereas two-fifths of the respondents indicated that they had problems regarding sexual dysfunction (41.2%). Of course, all three of these complications also could have a psychological component.
These data support the importance of considering the age and stage of life at which cancer occurs and the type and number of problems that the survivor faces. In the current study, younger cancer patients (ages 18–54 yrs) reported experiencing an average of 3 more problems than older patients (age 55 yrs and older). Mor et al.35 suggested that the reason older persons with cancer may demonstrate fewer and less severe psychosocial problems is that older patients have fewer demands on their time and resources compared with younger patients. They also suggested that the elderly are likely to have diminished expectations of life that might mitigate the negative effects of specific complications of cancer and its treatment. Another way of considering this difference in problem burden is in terms of the role responsibilities that people have at different stages of life. Younger individuals with cancer are more likely to still be employed and to have dependent family members and the related concerns of an earlier life stage.
In addition to age, concerns about problems related to living with cancer were found to differ for other demographic variables including gender, race, education, marital status, and household income. The findings in the current study are consistent with other studies that have shown that women with cancer are more willing than men to admit that they have psychologic problems.36–38 Vernon et al.39 also found that colorectal cancer patients who were female and had less formal education demonstrated more psychological distress. Similarly, Grosfeld et al.40 observed that young and single patients were more vulnerable to psychologic distress.
The finding in the current report that nonwhite cancer patients showed more concern regarding certain problems in living demonstrates the importance of considering the special needs of minority patients in their psychosocial response to cancer. Meyerowitz et al.41 have offered some insight into why members of minority groups might experience cancer differently by pointing out that minority status involves two components:1) the experience of having one's culture and ethnic identity unfamiliar to many people, and 2) the discrimination and prejudice that can accompany minority status for people of color in the U.S.
Conceptually, the impact of cancer and its treatment is usually considered in terms of a general stress model, in which cancer-related stressors predict distress and other psychosocial outcomes of this experience for cancer survivors.42, 43 These stressors include the related illness and treatment symptoms and functional limitations associated with the particular cancer type and the therapies the patient receives. The stress of the disease and its treatment affect physical health and functioning, which in turn affect the survivor's abilities to deal with various aspects of life. The current study deals with only a few elements of the fuller model of cancer-related stressors that will be tested in the full roll out of the SCS-I with its multiple questionnaires, including additional measures of physical health and functioning, symptoms, internal and external resources, and quality-of-life outcomes across time. The longitudinal study will allow for the examination of variables in addition to sociodemographic and medical characteristics, including personal and social variables such as self-esteem, dispositional optimism, coping, and social support. However, the regression analyses conducted on these preliminary data do demonstrate the relevance of including measures of general health and physical functioning as well as indicators of gender, age, marital status, and race/ethnicity in accounting for the number and types of problems in living encountered by cancer survivors.
As noted earlier, to our knowledge only a few studies to date have compared multiple types of cancer in terms of the psychosocial problems associated with each type. In the current study, in which such comparisons were possible, fatigue was reported as a common problem by those survivors who had types of cancer that usually are treated with chemotherapy, such as breast cancer. Conversely, patients with prostate cancer, who are less likely to receive chemotherapy, were less concerned with fatigue but more concerned about sexual dysfunction, which is understandable because impotence may occur as an iatrogenic outcome of prostatectomy, which is a common treatment for this disease.
The current study data demonstrated that on average, among survivors of the four most common cancers, lung cancer survivors experienced the most problems. They also had the highest mean CPILS total score, whereas prostate cancer survivors reported the fewest problems and had the lowest CPILS total score. To our knowledge, little is known regarding the experience of surviving lung cancer because of the low survival rate (a 1-yr survival rate of 42% and a 5-yr survivial rate of 15% have been reported for all stages of the disease44), but these data begin to help to develop a picture of the survivors of this cancer as facing more problems than survivors of other types of cancer. In speculating about why this is, it is worth noting that the amount of respiratory distress that lung cancer survivors experience must play a role in presenting the survivors with an array of problems in managing the tasks of daily living. Also, given the very high rate of mortality from this disease, it is not surprising that lung cancer survivors would have considerable concern and anxiety with regard to the future.
Those problems ranked among the top 10 reported did not differ very much across the 10 types of cancer studied herein, although some differences were found for survivors of the 4 most common types of cancer. The problems in the top 10 for the 4 most common cancers that do not appear in the overall top 10 are, for breast cancer, the diminished ability to concentrate (9th); for colorectal cancer, being less able to provide for the financial needs of my family (7th), difficulty in meeting my medical expenses (tie for 8th), and feeling angry (10th); for prostate cancer, the diminished ability to concentrate (7th) and feeling angry (8th); and for lung cancer, feeling dependent (8th) and feeling helpless (9th). More differences were noted when examining the ranking of problems across the four cancers. For example, although sexual dysfunction ranks as the 6th problem overall (41.2%), it not surprisingly is the number 1 problem (74.1%) identified by survivors of prostate cancer. The most striking differences were noted in the percentages of survivors across the four cancers who reported a particular problem. For example, although fear of disease recurrence is the number 1 or 2 problem for each of the four cancers, the percentage of survivors reporting this problem was found to range from a high of 74.2% for lung cancer survivors to a low of 48.8% for prostate cancer survivors. Other problems that demonstrated a large variation in occurrence across the four cancers were fatigue, concern regarding disease recurrence, fears about the future, and preoccupation with being ill.
The sample of patients included in the current study offers a number of strengths, including the fact that the sample is relatively large compared with the smaller opportunistic samples on which the majority of the psychosocial research with cancer patients is based. The sample also includes patients diagnosed with 10 different types of cancer, which are more cancer types than usual for this type of study. However, because these data were gathered as part of an implementation study for a much larger national study, there also are a number of limitations. The implementation pilot was confined to those states that volunteered to participate in this initial effort. Therefore, the data are taken from two upper Midwestern states (Iowa and Minnesota) and one state from the Southeast (Georgia), providing less diversity than would have been desirable for comparisons to be made by race. Accordingly, the current report compared whites and nonwhites because relatively few Hispanics or Asians could be included in the sample given the population of these states, and the black patients surveyed were mainly from the Georgia sample. In addition, the numbers of cases included in each of the 10 cancer types tended to be somewhat small for certain comparisons. A second limitation of the data is the low rate of consent for study participation. This low participation rate limits the generalizability of the current study findings, even though the questionnaire return rate of nearly 80% appears to be acceptable (and is lower than the actual questionnaire response rate for which deceased and ineligible cases would have removed from the denominator). Finally, insufficient data are available concerning those patients who did not participate to allow for a thorough assessment of possible biases.
These limitations will be overcome in the roll out of the larger national study. A greater effort will be made in the full study to secure help from the cancer registries in identifying the characteristics of those patients who were eligible and did not complete questionnaires so they can be compared for bias with those of the patients who did. Because all states that have sufficiently well developed cancer registries and that are willing to participate in the SCS-I will be included, the sample sizes will be in the thousands, resulting in cell sizes large enough for more complicated statistical analyses. This initial data collection, although smaller than what is planned for the future, has allowed for the refinement of procedures and the identification of some important correlations. The pilot study also has identified some hypotheses for testing when a larger sample is available.
With regard to the clinical implications of these data, the CPILS offers a tool for identifying problems in everyday life that concern cancer patients at various points in time after diagnosis. As a needs assessment tool, this measure can be used for setting goals for program development for particular groups of patients. In this pilot study, the CPILS has shown the need to provide psychologic counseling and supportive services to help patients deal with the fears and emotional concerns about the uncertain future that cancer presents to those diagnosed with it. Such services are not readily available and are not likely to be covered sufficiently by insurance. The need for sexual counseling for men who have prostate cancer and for their partners is also supported by these data. Fatigue was identified as a common problem, and recently, more attention is being given to cancer-related fatigue and its treatment. Drugs are now available to help restore platelets, the ameliorative effects of moderate exercise have been identified, and educational programs are available for patients who experience major fatigue problems. This study found that two physical health problems continued to concern cancer survivors—sleep difficulties and diminished ability to concentrate—which suggests a more concerted need for continued medical follow-up and possibly psychologic follow-up as well because both of these problems are often associated with anxiety and other psychologic distress.
The differences between younger and older cancer patients also suggest the need for services that are more focused on the different problems of cancer patients who still have work and family responsibilities. For persons who are still working when diagnosed with cancer, the workplace organization can be a key locus for needed services. More needs to be done to identify the advantages of providing specialized Employee Assistance Program (EAP) services for workers who are dealing with cancer.
As a growing number of people survive cancer over a longer period of time, it is becoming increasingly apparent that extending life is not the only criterion for effective cancer treatment. As more attention is given also to increasing the quality of life of people who receive a cancer diagnosis, it is critical to identify the continued problems in living with which survivors must deal. Data such as those presented in the current study can help establish which problems should be the focus of possible prevention and remediation efforts through effective treatment and care. The findings of the current study clearly indicate that, for more than a quarter of adult cancer survivors at 1 year past diagnosis, these problems include not only continued health problems, but also psychosocial problems such as emotional discomfort, fears about the future, financial problems, and difficulties in returning to former roles. The implementation study from which these data were drawn is only an initial, although important, step in assessing the issues in psychosocial adaptation and quality of life of long-term cancer survivors. The eventual completion of the larger national longitudinal American Cancer Society cohort study of cancer survivors will allow for a greater understanding of how these problems change over time in relation to a complex array of variables.
- 1SEER cancer statistics review, 1975–2001. Bethesda, MD: National Cancer Institute, 2004. Available at URL: http://www.seer.cancer.gov/csr/1975_2001 [accessed June 8, 2005]., , , , .
- 7Long-term psychosocial adaptation of survivors of acute leukemia. Proc Annu Meet Am Soc Clin Oncol. 1995; 14(A1668): 508., , , et al.
- 12Psychosocial adaptation of cancer survivors. In: HollandJC, editor. Psycho-oncology. New York: Oxford University Press, 1998: 223–254..
- 15Psychosocial sequelae in long-term survivors of testicular cancer. J Psychosoc Oncol. 1988; 6: 41–63., , .
- 25Work and cancer health histories: a study of the experiences of recovered patients (white collar study). Oakland, CA: American Cancer Society, California Division, 1976..
- 26Work and cancer health histories: study of experiences of recovered blue-collar workers. Oakland, CA: American Cancer Society, California Division, 1978..
- 28Work life for employees with disabilities: recommendations for promotion. Rehabil Couns Bull. 1997; 40: 181–191., , .
- 44American Cancer Society. Cancer facts and figures, 2005. Atlanta, GA: American Cancer Society, 2005.