Original Article. Key Words Ambulatory monitoring, circadian rhythm, fatigue, motor activity, neoplasms, sleep disorders

Transcription

1 Vol. 32 No. 3 September 2006 Journal of Pain and Symptom Management 245 Original Article Comparison Between Fatigue, Sleep Disturbance, and Circadian Rhythm in Cancer Inpatients and Healthy Volunteers: Evaluation of Diagnostic Criteria for Cancer-Related Fatigue Roland Fernandes, BSc, Patrick Stone, MA, MD, MRCP, Paul Andrews, BSc, PhD, Rachael Morgan, BSc, and Shanika Sharma, BSc St. George s University of London, London, United Kingdom Abstract The aim of this study was to evaluate whether diagnostic criteria for cancer-related fatigue syndrome (CRFS) could be rigorously applied to cancer inpatients, and to explore the relationship between subjective fatigue and objective measures of physical activity, sleep, and circadian rhythm. Female cancer patients (n ¼ 25) and a comparison group of subjects without cancer (n ¼ 25) were studied. Study participants completed a structured interview for CRFS and questionnaires relating to fatigue, psychological symptoms, and quality of life (QoL). Wrist actigraphs worn for 72 hours were used as an objective measure of activity, sleep, and circadian rhythm. Compared to controls, cancer patients were more fatigued, had worse sleep quality, more disrupted circadian rhythms, lower daytime activity levels, and worse QoL. After exclusion of subjects with probable mood disorders, the prevalence of CRFS was 56%. Fatigue severity among the cancer patients was significantly correlated with low QoL, depression, constipation, and decreased self-reported physical functioning. It can be concluded that the diagnostic criteria for CRFS can be applied to cancer inpatients but strict application requires a rigorous assessment of psychiatric comorbidity. Despite cancer inpatients having greater impairments of sleep and circadian rhythm, it was found that fatigue severity did not appear to be related to these impairments. J Pain Symptom Manage 2006;32:245e254. Ó 2006 U.S. Cancer Pain Relief Committee. Published by Elsevier Inc. All rights reserved. Key Words Ambulatory monitoring, circadian rhythm, fatigue, motor activity, neoplasms, sleep disorders Address reprint requests to: Roland Fernandes, BSc, 1 Heathview Road, Thornton Heath, Surrey, CR7 7PN, United Kingdom. m @sgul.ac. uk Accepted for publication: March 12, Ó 2006 U.S. Cancer Pain Relief Committee Published by Elsevier Inc. All rights reserved. Introduction Cancer-related fatigue may be defined as an unusual, persistent, subjective sense of tiredness related to cancer or cancer treatment that interferes with usual functioning. 1 The prevalence of fatigue among cancer patients has been reported to vary widely depending upon /06/$esee front matter doi: /j.jpainsymman

2 246 Fernandes et al. Vol. 32 No. 3 September 2006 the extent of the disease and the nature of the anticancer therapy that the patient is receiving. 2,3 Although fatigue severity undoubtedly varies among different patient groups, at least part of the variation in published prevalence figures relates to a lack of rigorous definition as to what constitutes a case of cancer-related fatigue. In an effort to address this problem, Cella et al. 4 have proposed diagnostic criteria for Cancer-Related Fatigue Syndrome (CRFS). To be considered a case of fatigue, the subject should fulfill the following criteria: 1) experience six out of 10 fatigue-related symptoms on most days or every day for 2 weeks in the previous month (at least one of which should be significant fatigue); 2) these symptoms should be causing impairment in function; 3) there should be reason for believing that the fatigue is a consequence of cancer or cancer therapy; and 4) the fatigue should not be primarily a consequence of a comorbid psychiatric condition. The prevalence of CRFS using criteria proposed by Cella et al. among cancer inpatients is not known. The causes of cancer-related fatigue have not yet been identified. It has been hypothesized that fatigue may be secondary to insomnia and altered circadian rhythms. 5 Davidson et al. 6 reported that those cancer outpatients who complain of fatigue are two and a half times more likely to experience insomnia. Savard et al. 7 investigated the prevalence of sleep disorders and insomnia syndromes in 300 breast cancer survivors and reported that 51% of their sample were experiencing sleep problems and that 19% fulfilled the diagnostic criteria for an insomnia syndrome. Mormont et al. 8 monitored activity levels in patients with metastatic colorectal carcinoma and found that patients with a pronounced circadian rhythm had a better quality of life (QoL) and suffered from less fatigue. Other studies have reported a significant correlation between fatigue severity and self-reported sleep difficulties in breast cancer survivors. 9e11 One might expect that the prevalence of sleep disorders among cancer inpatients would be even higher and that this would be a significant contributor to fatigue severity in this population, but this hypothesis has not been tested directly. We decided to undertake a study to assess the applicability/practicality of the diagnostic criteria proposed by Cella et al. for CRFS among female cancer inpatients and to test the hypotheses that fatigue severity would be related to decreased daytime activity, sleep disorders, and alterations in circadian rhythm. Because fatigue is a common complaint among the general population, we decided to study a comparison group of female volunteers without cancer as well. Methods Subjects This was a comparison between a convenience sample of female inpatients with cancer and volunteers with no history of cancer. It was decided to investigate only women to make the study population more homogeneous and because there are known differences between the prevalence of insomnia between the sexes. 12 Inclusion criteria were the ability to give informed consent and an expectation of remaining as an in-patient for three or more nights. Patients with a prognosis of less than one month, who were being barrier nursed, or had a booked procedure during the duration of the study were excluded. Noncancer participants were recruited from a pool of women known to the authors and were selected so that the median age of the group was likely to be comparable to that of the cancer group. Consequently, the eligibility criteria for this group was an age between 35 and 90, the ability to give informed consent, understand the questionnaire, and wear the actigraph for 72 consecutive hours. Women who currently or previously had a diagnosis of cancer were excluded. The study was approved by the Local Research Ethics Committee and informed consent was obtained from all participants. Procedure Both patients and volunteers underwent the following assessments. Assessment of Fatigue Diagnostic Interview Guide for Cancer-Related Fatigue. To determine whether or not a participant should be considered as a case of fatigue, the Diagnostic Interview Guide for

3 Vol. 32 No. 3 September 2006 Fatigue, Sleep Disturbance & Circadian Rhythm 247 Cancer-Related Fatigue 4 was administered. This is a short 13-item interview that categorizes patients as either cases or noncases of CRFS according to diagnostic criteria proposed by Cella et al. The diagnostic interview has been successfully used in several previous studies of cancer-related fatigue. 13,26,27 The interview schedule is read to the participants from a predetermined script, and answers are in the form of a yes or no response, thus allowing little scope for inter-interviewer variation. Nonetheless, for consistency all three interviewers were shown how to use the interview by the supervisor of the project (PS). For a rigorous application of these criteria, it is necessary to identify whether subjects have a comorbid psychiatric disorder that may be contributing toward the fatigue and this can only really be achieved by undertaking a structured psychiatric interview. Practical and resource constraints meant that strict application of these criteria was not possible. However, to exclude patients with significant psychiatric disorders from being inappropriately labeled as cases of CRFS, we screened all subjects using the Hospital Anxiety and Depression Scale (HADS) (see below). European Organization for Research and Treatment of Cancer Fatigue Subscale. The primary measure of fatigue severity in this study was the three-item fatigue subscale of the European Organization for Research and Treatment of Cancer 30-item QoL (EORTC QLQ-C30) questionnaire. 14 The fatigue subscale consists of three items (During the past week.did you need to rest?, Have you felt weak?, Were you tired?). Each item can be answered on a fourpoint scale (not at all, a little, quite a bit, or very much). Scores are transformed to a0e100 scale with 0 representing no fatigue and 100 representing maximum fatigue. Bidimensional Fatigue Scale. The Bidimensional Fatigue Scale (BFS) was used as a secondary measure of fatigue severity. This scale was devised by Chalder et al. 15 and has been used to measure fatigue in large community 16,17 and primary care studies. 18 The scale consists of 11 items, seven of which form a physical fatigue subscale (BFS-P) and four of which form a mental fatigue subscale (BFS-M). A combined score for the total scale can also be constructed (BFS-T). Each item is scored as 1 point if the symptom is worse than usual or much worse than usual, or scored as 0 points if better than usual or same as usual. Thus, the BFS-T can range between 0 (minimum fatigue) and 11 (maximum fatigue). Assessment of QoL. Other domains of QoL were assessed using the EORTC QLQ-C30. This is a well-validated health-related QoL instrument. 14 The questionnaire consists of 30 items incorporating five functional scales, eight symptom scales, and a global QoL scale. Scores are transformed onto a 0e100 scale, where higher scores are indicative of a higher level of functioning or a higher level of symptom disturbance. Assessment of Mood. Mood was assessed using the HADS. 19 This scale was originally developed for use among hospital inpatients and has been widely used in patients with cancer. 20,21 It consists of seven items that evaluate anxiety and seven items that measure depression. Each item can be answered on a fourpoint scale (0e3). Scores on each subscale can thus range between 0 (no symptoms) and 21 (numerous and severe symptoms). A total score can also be generated. When used as a screening instrument, the recommended cut-offs are a score of 11 or greater on either subscale 22 or a score of 19 or greater for the total score. 20 Assessment of Sleep and Circadian Rhythm Objective Measures. Activity, sleep, and circadian rhythms were assessed using wrist actigraphy. Actimeters (Ambulatory Monitoring Inc., New York) are small, piezoelectric accelerometers that are worn on the wrist. They detect movement of the limb and then summarize the data in 1-minute epochs. Associated software analyzes the data using an algorithm to determine whether the individual is awake or asleep. Actigraphy has been validated against polysomnography 23,24 as an objective measure of sleep/wake patterns. The 24-hour autocorrelation coefficient (R 24 ) is a measure of circadian rhythm. It is calculated by comparing activity data during each 1-minute epoch of a 24-hour period with the activity levels during subsequent epochs. A

4 248 Fernandes et al. Vol. 32 No. 3 September 2006 strong circadian rhythm is indicated by a good correlation between activity levels during epochs separated by 24 hours. Subjective Measures. Data were also obtained from self-report sleep diaries. Using the sleep diaries, a down interval was defined as a period of time in which the participant reported to be trying to sleep. An up interval was defined as a period between two successive down periods, and represented a period of wakefulness. This system of methodology has been used previously in studies investigating sleep. 25 Statistical Methods Data analysis was performed using the Statistical Package for the Social Sciences (SPSS) for Windows version Data were summarized using nonparametric statistics (medians and ranges). Spearman rank correlation coefficients were used to examine the relationship between continuous variables. The Mann- Whitney U test was used for comparisons between groups. Results Sample Characteristics (Table 1) A total of 50 subjects were studied (25 patients and 25 volunteers). There was no statistical difference in terms of age between the two groups (patients median age 67 years, range 46e90; volunteers median age 63 years, range 54e79; P ¼ 0.020). Cancer patients had a variety of diagnoses, with the most common primaries being breast (8/25), lung (5/25), colon (4/25), and skin (3/25) cancer. Most (18/25, 72%) had metastatic disease. Four patients were receiving chemotherapy, and one patient was receiving radiotherapy at the time of study entry. Patients had been on the ward for a median of 24 days (range 2e196) prior to recruitment into the study. Differences Between Patients and Volunteers Prevalence of Fatigue. Using the diagnostic criteria proposed by Cella et al. (if patients with probable mood disorders according to the HADS are excluded), the prevalence of CRFS among the inpatients was 14/25 (56%). It is not possible to directly apply the Table 1 Clinical Characteristics of Cancer Inpatients Median age (years) 67 (range 46e90) Diagnosis Breast 32% Lung 20% Colon 16% Skin 12% Treatment at time of study entry Chemotherapy 16% Radiotherapy 4% Treatment history Chemotherapy 76% Radiotherapy 40% Metastases 72% Medication Analgesics 80% Steroids 56% Hypnotics 20% Antidepressants 16% criteria to the control group (since by definition it is not possible to have CRFS if you do not have cancer). However, if Cella et al. s other diagnostic criteria are applied to the control group (i.e., severe fatigue plus five other fatigue-related symptoms causing significant distress or impairment of function in the absence of comorbid psychiatric conditions), then the prevalence of fatigue syndrome in the controls was 5/25 (20%). These prevalence figures are significantly different (P ¼ 0.014). Sleep Quality, Circadian Rhythm, and Activity. Fig. 1 shows illustrative actigraphic data obtained from one of the cancer patients and one of the control subjects. It is apparent that the cancer patient is less active during the day, more active during the night, and has a less pronounced circadian rhythm. Data from all of the participants are summarized in Table 2. With the exception of sleep latency (i.e., the time taken to fall asleep after going to bed), all of the actigraphic parameters are significantly different between the two groups. QoL and Mood. Using the recommended cutoffs, 20,22 28% (7/25) of cancer patients were probable cases of depression and there were no probable cases in the control group. This difference was statistically significant (P ¼ 0.002). The proportion of those with a probable mood disorder, (i.e., a combined score on the HADS of 19 or greater), was also

5 Vol. 32 No. 3 September 2006 Fatigue, Sleep Disturbance & Circadian Rhythm 249 Fig. 1. Linear actigrams display quantity of activity against time on a histogram. The x-axis reflects the time (with midnight at the center) and the y-axis reflects the activity counts per epoch. The height of each black bar is proportional to movement intensity. Epochs scored as sleep by the actimeters are shown with thickened lines on the x-axis. Down periods obtained from data from sleep diaries can be programmed onto the actigram and are shown in the figure as highlighted gray boxes. The autocorrelation coefficient plot reflects the correlation between activity at any particular time with activity levels at different time lags. The correlation coefficient is displayed on the y-axis and can vary between e1.0 and þ1.0. The different time lags are shown on the x-axis. Subjects with a good circadian rhythm would be expected to have a sinusoidal plot with a maximum autocorrelation coefficient occurring at 24-hours time lag.

6 250 Fernandes et al. Vol. 32 No. 3 September 2006 Study Variable Table 2 Comparison of Patients (n ¼ 25) and Controls (n ¼ 25) Using the Mann-Whitney Test a Median Score in Patients Median Score in Controls P-Value for Difference HADS Depression Subscale Score 6 (11) 2 (4) HADS Anxiety Subscale Score 4 (7) 5 (5) Combined HADS Score 13 (23) 8 (9) BFS-Mental Subscale Score 1 (3) 0 (1) BFS-Physical Subscale Score 6 (3) 0 (3) <0.001 Total BFS Score 7 (6) 0 (3) <0.001 EORTC QLQ-C30 Fatigue (47.22) (22.22) <0.001 Nausea/vomiting (70.84) (00.00) <0.001 Pain (70.83) (16.67) <0.001 Dyspnea (66.66) (33.33) <0.001 Insomnia (66.66) (66.67) Appetite loss (100.00) (00.00) <0.001 Constipation (83.33) (00.00) <0.001 Diarrhea (66.66) (00.00) Quality of life (37.50) (41.67) <0.001 Physical functioning (40.00) (20.00) <0.001 Role functioning (54.17) (50.00) <0.001 Emotional functioning (37.50) (33.33) <0.001 Cognitive functioning (37.50) (33.33) <0.001 Social functioning (37.50) (00.00) <0.001 Mean activity (up interval) b (65.31) (19.20) <0.001 Mean activity (down interval) c (41.32) (10.99) % Sleep (up interval) d (19.20) 3.24 (3.89) <0.001 % Sleep (down interval) e (31.24) (9.66) Sleep efficiency (%) f (25.20) (7.82) Sleep latency (minutes) g (39.92) (20.67) Wake after sleep onset (minutes) h (60.32) (32.00) Autocorrelation coefficient (R 24 ) i 0.28 (0.23) 0.46 (0.14) <0.001 EORTC QLQ-C30 ¼ European Organization for Research and Treatment of Cancer Quality of Life Questionnaire C-30. a Interquartile ranges in brackets. b Mean activity score (counts/minute) during the time spent not trying to sleep. c Mean activity score (counts/minute) during the time spent trying to sleep. d Percent minutes scored as sleep during the time spent not trying to sleep. e Percent minutes scored as sleep during the time spent trying to sleep. f Proportion of actual time spent asleep when attempting to sleep. g Minutes spent trying to sleep until the start of the first 20-minute block of sleep. h Number of wake minutes during the actual time spent asleep. i Twenty-four hour autocorrelation coefficient reflecting circadian rhythm. found to be statistically different in the cancer (8/25, 32%) and cancer-free (2/25, 8%) groups (P ¼ 0.004). All of the QoL domains, with the exception of the insomnia score, were also significantly different between the two groups. noncases (n ¼ 11) of fatigue. The only significant differences between the two groups were in the anxiety subscale (HADA, P ¼ 0.002), depression subscale (HADD, P ¼ 0.007), and total score (P ¼ 0.001) of the HADS. Correlation Between Subjective Fatigue and Other Study Variables (Table 3) In the cancer patients, fatigue severity was significantly correlated with the depression subscale and total scores on the HADS questionnaire, and, with constipation, decreased physical functioning, and low QoL on the EORTC QLQ-C30. We also compared the difference in the study variables in the patient group between the cases (n ¼ 14) and the The Performance of the Diagnostic Interview Guide for Cancer-Related Fatigue Cella et al. suggest that a case of cancer-related fatigue should be defined by the presence of six or more fatigue-related symptoms, at least one of which should be significant fatigue, reduced physical functioning, and the absence of comorbid psychiatric disorders. We analyzed the data to explore the effect that altering this diagnostic threshold would

7 Vol. 32 No. 3 September 2006 Fatigue, Sleep Disturbance & Circadian Rhythm 251 Table 3 Spearman Rank (Rs) Correlations of the Study Variables Against the EORTC QLQ-C30 Fatigue Score Patients Controls Rs P-Value Rs P-Value HADS Depression Subscale Score <0.001 HADS Anxiety Subscale Score Combined HADS Score <0.001 BFS-Mental Subscale Score BFS-Physical Subscale Score Total BFS Score EORTC QLQ-C30 Nausea/vomiting Pain Dyspnea Insomnia <0.001 Appetite loss Constipation Diarrhea QoL Physical functioning < Role functioning Emotional functioning <0.001 Cognitive functioning <0.001 Social functioning Mean activity (up interval) Mean activity (down interval) % Sleep (up interval) % Sleep (down interval) Sleep efficiency (%) Sleep latency (minutes) Wake after sleep onset (minutes) Autocorrelation coefficient (R 24 ) EORTC QLQ-C30 ¼ European Organization for Research and Treatment of Cancer Quality of Life Questionnaire C-30. have on the prevalence of CRFS (Table 4). If significant fatigue were the only fatigue-related symptom required for the diagnosis of CRFS, then the prevalence in our sample would have been 96%. However, if having significant fatigue causing impairment of functioning attributable to cancer or its treatment and in the absence of comorbid psychiatric disorders were taken as the criteria for diagnosing CRFS, then the prevalence in our Table 4 The Effect of Varying the Threshold for Diagnosing Cases of CRFS using Cella et al. s Diagnostic Criteria Cella et al. Criteria Prevalence of CRFS Criterion A1 96% (24/25) Criteria A1 þ B 96% (24/25) Criteria A1 þ B þ C 96% (24/25) Criteria A1 þ B þ C þ D 64% (16/25) Criteria A1 þ 1 other A symptom þ B þ C þ D 60% (15/25) Criteria A1 þ 2 other A symptoms þ B þ C þ D 60% (15/25) Criteria A1 þ 3 other A symptoms þ B þ C þ D 60% (15/25) Criteria A1 þ 4 other A symptoms þ B þ C þ D 60% (15/25) Criteria A1 þ 5 other A symptoms þ B þ C þ D 56% (14/25) Criteria A1 þ 6 other A symptoms þ B þ C þ D 40% (10/25) Criteria A1 þ 7 other A symptoms þ B þ C þ D 20% (5/25) Criteria A1 þ 8 other A symptoms þ B þ C þ D 16% (4/25) Criteria A1 þ 9 other A symptoms þ B þ C þ D 0% (0/25) A1 ¼ presence of significant fatigue; other A symptoms ¼ presence of up to nine other fatigue-related symptoms (e.g., feeling weak all over, having difficulty concentrating, not feeling refreshed after sleep); B ¼ fatigue has an effect on functioning; C ¼ symptoms are a consequence of cancer or cancer therapy; D ¼ symptoms are not primarily due to a comorbid psychiatric disorder.

8 252 Fernandes et al. Vol. 32 No. 3 September 2006 sample would have been 64%. If all nine of the other fatigue-related symptoms listed in the interview schedule were required for a diagnosis of CRFS, then the prevalence in our sample would have been 0%. Discussion Because fatigue is a common complaint among the general population, it is important that rigorous criteria are used to diagnose CRFS in patients with cancer. Otherwise, there is a risk that the prevalence of fatigue will be overestimated. The criteria proposed by Cella et al. go some way in addressing this methodological problem. If patients in our study had simply been asked to report the presence of significant fatigue, we would have found the prevalence of such fatigue to be 96%. However, excluding patients with comorbid psychiatric disorders and only including patients with at least five other fatigue-related symptoms reduced our prevalence figure for CRFS to 56%. This compares to prevalence figures of 17% for randomly selected community cancer patients identified in a telephone survey, 26 23% among patients following bone marrow transplantation, 27 26% among breast cancer survivors, 13 and 20% among the control group in our own study. An important limitation of our study is the failure to control for the effect of hospitalization and treatment on both sleep and fatigue. To fully assess the relative impact that hospitalization and the presence of a cancer diagnosis have on the prevalence of fatigue and sleep disorders it would have been necessary to include other control groups in our study (e.g., healthy volunteers at home and in a hospital environment, and hospitalized patients with noncancer diagnoses). However, these refinements were not within the resources of the present study. Similarly, the decision to use a relatively heterogeneous sample of patients (with respect to primary diagnosis and treatment history) was made on practical and resource grounds. The effect of anticancer treatment on fatigue prevalence has recently been demonstrated by Andrykowski et al. 13 They reported that CRFS increased from 10% to 26% following adjuvant treatment for breast cancer. Other potential confounders of the current study include the differential administration of analgesics, sedatives, intravenous fluids, and the environmental conditions on the ward where the patients were staying. All these factors may influence the prevalence of fatigue and/or insomnia, and future studies should attempt to control for such variables. It is interesting to note that although there were numerous differences between the patients and controls with regard to most QoL domains, there was no significant difference in insomnia scores between the two groups. Nonetheless, the actigraphic measures of sleep (with the exception of sleep latency) were significantly different between the two groups. This suggests that simply asking subjects Have you had problems sleeping?, as the EORTC QLQ- C30 does, is insufficient to detect all but the grossest of sleep disturbances. The actimeters produced a very rich data source for each of the subjects, they were well tolerated, and provided objective evidence of decreased activity, reduced sleep, and impaired circadian rhythms. Contrary to expectation, there were no significant correlations between subjective fatigue severity and any of the actigraphic variables. The strongest association (Rs ¼ 0.4, P ¼ 0.087) was between fatigue and circadian rhythm (R 24 ). It is possible that the failure to find a statistically significant correlation between these variables was due to the small sample size that we used. A previous study by Mormont et al. 8 involving 200 ambulatory metastatic colorectal cancer patients reported a statistically significant but low magnitude association (Rs ¼ 0.26, P ¼ 0.001) between fatigue and R 24. This figure is in keeping with our own results. We found that mood disturbance (particularly depression) was one of the variables most strongly associated with fatigue severity in both the patients and the controls. Previous studies have also reported that depression is one of the strongest correlates of fatigue in both the general population 15,16 and among cancer patients. 2 The classification proposed by Cella et al. recommends that patients with comorbid psychiatric disorders that may be contributing to fatigue should be excluded from the diagnosis of CRFS. To thoroughly assess this diagnostic criterion, it would have been necessary to undertake structured psychiatric interviews with the study participants. This was beyond the resources of the current study; however, we were able to use the

9 Vol. 32 No. 3 September 2006 Fatigue, Sleep Disturbance & Circadian Rhythm 253 HADS to exclude subjects with a probable mood disorder. It is interesting to note that if we had failed to exclude such subjects from our study, the prevalence of fatigue in our patient sample would have risen to 88% (22/25). We found that even after excluding all subjects with a probable mood disorder, there was a significant association (Rs ¼ 0.75, P < 0.001) between depression scores and fatigue severity among those 19 subjects (14 patients and 5 controls) who fulfilled the diagnostic criteria for a fatigue syndrome. Conclusion The prevalence of CRFS among this sample of cancer inpatients was 56%. More research is required to determine how practical the diagnostic criteria proposed by Cella et al. would be if a rigorous exclusion is applied to subjects with comorbid psychiatric disorders. We have found that inpatients with cancer have demonstrable impairments of sleep and circadian rhythm that are likely to have significant effects on QoL. However, in this relatively small study, we did not detect any significant association between fatigue severity and any of the objective measures of activity, sleep, or circadian rhythm. References 1. Mock V. Breast cancer and fatigue: issues for the workplace. AAOHN J 1998;46(9):425e431; quiz 432e Stone P, Richards M, A Hern R, Hardy J. A study to investigate the prevalence, severity and correlates of fatigue among patients with cancer in comparison with a control group of volunteers without cancer. Ann Oncol 2000;11(5):561e Glaus A. The relationship between fatigue and type and stage of cancer. In: Glaus A, ed. Recent results in cancer research. Fatigue in patients with cancer: analysis and assessment. New York: Springer, 1998:105e Cella D, Peterman A, Passik S, et al. Progress toward guidelines for the management of fatigue. Oncology (Huntington) 1998;12(11A):369e Olders H, Winningham M. Select psychiatric and psychological considerations. In: Winningham M, Barton-Burke M, eds. Fatigue in cancer: A multidimensional approach. Boston: Jones and Bartlett, 2000:197e Davidson JR, MacLean AW, Brundage MD, Schulze K. Sleep disturbance in cancer patients. Soc Sci Med 2002;54(9):1309e Savard J, Simard S, Blanchet J, et al. Prevalence, clinical characteristics, and risk factors for insomnia in the context of breast cancer. Sleep 2001;24(5): 583e Mormont MC, Waterhouse J, Bleuzen P, et al. Marked 24-h rest/activity rhythms are associated with better quality of life, better response, and longer survival in patients with metastatic colorectal cancer and good performance status. Clin Cancer Res 2000;6(8):3038e Andrykowski MA, Curran SL, Lightner R. Off-- treatment fatigue in breast cancer survivors: a controlled comparison. J Behav Med 1998;21(1):1e Broeckel JA, Jacobsen PB, Horton J, et al. Characteristics and correlates of fatigue after adjuvant chemotherapy for breast cancer. J Clin Oncol 1998; 16(5):1689e Bower JE, Ganz PA, Desmond KA, et al. Fatigue in breast cancer survivors: occurrence, correlates, and impact on quality of life. J Clin Oncol 2000; 18(4):743e Reyner LA, Horne JA, Reyner A. Gender- and age-related differences in sleep determined by homerecorded sleep logs and actimetry from 400 adults. [erratum appears in Sleep 1995;18(5):391]. Sleep 1995;18(2):127e Andrykowski MA, Schmidt JE, Salsman JM, et al. Use of a case definition approach to identify cancer-related fatigue in women undergoing adjuvant therapy for breast cancer. J Clin Oncol 2005;23: 6613e Aaronson NK, Ahmedzai S, Bergman B, et al. The European Organization for Research and Treatment of Cancer QLQ-C30: a quality-of-life instrument for use in international clinical trials in oncology. J Natl Cancer Inst 1993;85(5):365e Chalder T, Berelowitz G, Pawlikowska T, et al. Development of a fatigue scale. J Psychosom Res 1993;37:147e Lawrie S, Pelosi A. Chronic fatigue syndrome in the community: prevalence and associations. Br J Psychiatry 1995;166:793e Pawlikowska T, Chalder T, Hirsch S, et al. A population based study of fatigue and psychological distress. Br Med J 1994;308:743e David A, Pelosi A, McDonald E, et al. Tired, weak, or in need of rest: fatigue among general practice attenders. Br Med J 1990;301(6762): 1199e Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta Psychiatr Scand 1983; 67(6):361e Razavi D, Delvaux N, Farvacques C, Rabaye E. Screening for adjustment disorders and major

10 254 Fernandes et al. Vol. 32 No. 3 September 2006 depressive disorders in cancer in-patients. Br J Psychiatr 1990;156: Hopwood P, Howell A, Maguire P. Screening for psychiatric morbidity in patients with advanced breast cancer: validation of two self-report questionnaires. Br J Cancer 1991;64:353e Snaith R, Zigmond A. The Hospital Anxiety and Depression Scaledmanual. Windsor: NFER Nelson, Sadeh A, Hauri P, Kripke D, Lavie P. The role of actigraphy in the evaluation of sleep disorders. Sleep 1995;18:288e Cole RJ, Kripke DF, Gruen W, et al. Automatic sleep/wake identification from wrist activity. [see comment]. Sleep 1992;15(5):461e Korszun A, Young EA, Cary Engleberg N, et al. Use of actigraphy for monitoring sleep and activity levels in patients with fibromyalgia and depression. J Psychosom Res 2002;52:439e Cella D, Davis K, Breitbart W, et al. Cancerrelated fatigue: prevalence of proposed diagnostic criteria in a United States sample of cancer survivors. J Clin Oncol 2001;19(14):3385e Sadler IJ, Jacobsen PB, Booth-Jones M, et al. Preliminary evaluation of a clinical syndrome approach to assessing cancer-related fatigue. J Pain Symptom Manage 2002;23(5):406e416.