Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2014 Jul 1.
Published in final edited form as: Health Psychol. 2013 Feb 25;32(7):768–774. doi: 10.1037/a0031322

Lagged Relationships Among Sleep Disturbance, Fatigue, and Depressed Mood During Chemotherapy

Heather SL Jim 1, Paul B Jacobsen 2, Kristin M Phillips 3, Robert M Wenham 4, William Roberts 5, Brent J Small 6
PMCID: PMC3700563  NIHMSID: NIHMS451889  PMID: 23437852

Abstract

Objective

Recent research suggests that sleep disturbance, fatigue, and depressed mood form a symptom cluster in patients treated with chemotherapy. To date, however, no studies have examined lagged relationships among these symptoms during chemotherapy, a time when symptom variability is high. The aim of the current study was to examine lagged changes among daily symptoms during platinum-based chemotherapy.

Methods

Participants were 78 women with gynecologic cancer (mean age 63, SD=11; 91% Caucasian; 97% non-Hispanic). Sleep disturbance was assessed via wrist actigraphy, while fatigue and depressed mood were assessed via daily diary in the week after participants’ first chemotherapy infusion. Latent change score models (LCS) were used to examine lagged relationships between symptom pairs.

Results

High levels of sleep disturbance (i.e., minutes awake at night) were associated with earlier subsequent peaks in fatigue, while high levels of fatigue were associated with higher subsequent levels of depressed mood.

Conclusions

These findings suggest that sleep disturbance, fatigue, and depressed mood occur in a cascade pattern during chemotherapy, in which increases in sleep disturbance contribute to fatigue, which in turn contributes to depressed mood. Interventions targeting symptoms early in the cascade, such as sleep disturbance, may provide benefits across multiple downstream symptoms.

Keywords: neoplasms, gynecologic neoplasms, sleep, fatigue, depression

A growing body of research indicates that sleep disturbance, fatigue, and depressed mood are common and distressing problems among cancer patients undergoing chemotherapy. Symptoms such as these typically follow a predictable “rollercoaster” pattern in which they are highest in the week after a chemotherapy infusion, then gradually decrease until the next infusion (Berger, 1998; Jim et al., 2011). Sleep disturbance, fatigue, and depression tend to be particularly severe in patients treated with the intravenous combination of platinum and taxane that is currently recommended for many gynecologic cancers (Morgan et al., 2009). Data indicate that approximately 88% of women undergoing this type of chemotherapy report moderate to severe fatigue while on treatment, while 80% report sleep disturbance and 25% report depressed mood (Butler et al., 2004; Goncalves, Jayson, & Tarrier, 2008; Palesh et al., 2010).

Research suggests that sleep disturbance, fatigue, and depressed mood form a symptom cluster in cancer patients undergoing chemotherapy. These symptoms show high cross-sectional correlations with one another and low correlations with other symptoms, such as headaches (Bender, Ergyn, Rosenzweig, Cohen, & Sereika, 2005; Donovan & Jacobsen, 2007). Not only do these symptoms demonstrate significant cross-sectional relationships (Berger, Wielgus, Hertzog, Fischer, & Farr, 2009; Byar, Berger, Bakken, & Cetak, 2006; Donovan & Jacobsen, 2007; Jacobsen et al., 1999), they also appear to change together over time (Liu et al., 2009). For example, Roscoe and colleagues (Roscoe et al., 2002) found that increases in fatigue from the second to the fourth chemotherapy infusions were associated with increases in depressed mood and objectively-measured sleep disturbance in women with breast cancer. In addition, intraday increases in fatigue are associated with intraday increases in depressed mood in women undergoing chemotherapy for ovarian cancer (Badr, Basen-Engquist, Carmack Taylor, & De Moor, 2006). In a recent study of gynecologic cancer patients, we found that daily increases in fatigue were associated with concurrent increases in daily depressed mood and objectively-measured night awakenings in the week following each of the first three chemotherapy infusions (Jim, et al., 2011).

Findings such as these have generated interest in determining why these symptoms co-occur. One possibility is that chemotherapy gives rise to multiple symptoms simultaneously. Another possibility is that these symptoms occur in a cascade, in which increases in one symptom contribute to increases in others. Several studies have investigated the possibility that symptoms influence one another. For example, Stepanski and colleagues (Stepanski et al., 2009) used structural equation modeling to examine relationships among symptoms in a heterogeneous sample of cancer patients. They found that sleep disturbance mediated the relationship between depressed mood and fatigue. Similarly, Banthia and colleagues (Banthia, Malcarne, Ko, Varni, & Sadler, 2009) found that mood and sleep predicted fatigue in breast cancer survivors. Conversely, Huang and Lin (Huang & Lin, 2009) found that depressed mood mediated the influence of sleep disturbance on fatigue in patients with hepatocellular carcinoma. Because all three of these studies were cross-sectional, the temporal influences of symptoms on one another remain unclear. In contrast, a study of lagged relationships between symptoms in post-treatment breast cancer survivors found that self-reported nighttime sleep was associated with next day fatigue and depressed mood, but that fatigue and depressed mood did not predict sleep that night (Rumble et al., 2010).

Efforts to understand relationships among sleep disturbance, fatigue, and depressed mood have been hampered by weaknesses in study design and statistical methodology. As noted above, many of the studies examining relationships between symptoms have been cross-sectional. Even when longitudinal designs are employed, symptoms are often measured infrequently, with several weeks or more between assessments. Among studies measuring symptoms daily or multiple times per day, sophisticated techniques for analyzing time-series data have been lacking, so that symptoms are often averaged across assessments. The result is a loss of statistical power and reduced ability to describe highly variable patterns of symptom change.

To address these issues, we used advanced methodological and statistical techniques to model complex relationships among symptoms during chemotherapy, a time when symptoms fluctuate widely both within and across days (Jim, et al., 2011). Wrist actigraphy and daily diaries were used to collect data on sleep disturbance, fatigue, and depressed mood in women undergoing chemotherapy for gynecologic cancer. Our first study used piecewise regression analyses to describe daily and intraday changes in these symptoms as well as concurrent relationships among symptom changes (Jim, et al., 2011). We found that in the weeks after chemotherapy infusions, increases in nighttime sleep disturbance were associated with increases in fatigue, while increases in fatigue were associated with increases in depressed mood. Nighttime sleep was not associated with depressed mood, however. The aim of the current study was to examine lagged relationships among symptoms demonstrating concurrent associations in our previous study. To date, no studies have examined lagged relationships among symptoms specifically in patients undergoing chemotherapy. Examination of lagged relationships is important to help understand the directionality of symptom influences on one another. This understanding can then be used clinically to prioritize management of symptoms that contribute to other symptoms. Based on previous data indicating statistically significant concurrent symptom relationships (Jim, et al., 2011), we examined lagged relationships among two pairs of symptoms: 1) night awakenings and fatigue and 2) fatigue and depressed mood.

Lagged symptom changes were examined using a Latent Change Score models (LCS) approach (Ferrer & McArdle, 2010; McArdle, 2009; McArdle et al., 2004). With this approach, we can compare models corresponding to four specific hypotheses regarding each symptom pair, each representing alternative versions of the potential dynamic relationships. The corresponding hypotheses were: 1) there is no relationship among symptom changes, 2) changes in Symptom A are a leading indicator in that they precede changes in Symptom B, 3) changes in Symptom B are a leading indicator of changes in Symptom A, and 4) a form of dynamic coupling exists such that Symptom A and Symptom B both influence changes in one another. This statistical method has been used extensively in previous research to evaluate the dynamic relationships between outcomes such as cognition and lifestyle activities and perceived control and self-reported health (Ghisletta, Bickel, & Lovden, 2006; Infurna, Gerstorf, & Zarit, 2011; Small, Dixon, McArdle, & Grimm, 2011).

Method

Participants and Procedures

As part of a larger study of symptoms during cancer treatment (Jim, et al., 2011), women with gynecologic cancer were recruited for an IRB-approved study examining side effects of platinum-based chemotherapy. Eligibility criteria were that participants: 1) be at least 18 years of age; 2) be scheduled to receive intravenous platinum-based chemotherapy for gynecologic cancer at Moffitt Cancer Center; 3) were not treated with chemotherapy for at least two months prior to recruitment; 4) be free of documented or observable psychiatric or neurologic disorders that could interfere with study participation (e.g., Parkinson’s disease, schizophrenia); 5) be able to speak and read English; and 6) provide written informed consent. Patients were recruited between September 2007 and July 2009.

Eligibility was determined by chart review and consultation with the attending physician. Eligible women were recruited and informed consent was obtained during an outpatient clinic visit before the start of chemotherapy. All participants completed a baseline demographic assessment and began actigraphic monitoring at this time. Participants continuously wore the actigraph and completed daily diaries of bedtime, rising time, fatigue, and depressed mood until seven days after their first infusion, when they returned the actigraph and diaries by mail.

Eighty women agreed to participate. Two participants did not provide any data and thus are not included in analyses; one became ineligible due to a change in treatment plans and the other elected to discontinue study participation. An additional two participants provided partial data before becoming ineligible or electing to discontinue study participation; these participants were included in analyses. Thus, the final sample consisted of 78 participants.

Measures

Demographic and clinical data

Age, race/ethnicity, marital status, education level, and annual household income were assessed in all participants via self-report. Disease stage, chemotherapy regimen, days since surgery, and recurrence status were assessed in participants via medical chart review.

Daily diaries

Participants were asked to rate their fatigue and depressed mood at 10 am, 2 pm, and 6 pm during assessment days. These times were selected because they occurred at regularly-scheduled intervals when participants were likely to be awake. The actigraph beeped at these times to remind participants to complete their ratings. Fatigue and depressed mood were rated on an eleven-point scale (0=no fatigue/depression at all, 10=as fatigued/depressed as I could be). Ratings were averaged for each day to produce one mean fatigue score and one mean depressed mood score per participant per day. Participants also recorded their bedtimes and rising times during assessment days.

Actigraphy

Wrist actigraphy was used to objectively measure sleep. The Actiwatch®-Score (MiniMitter, Bend, OR) actigraph was used. Participants wore the actigraph continuously on the non-dominant wrist during the assessment period. The actigraph uses a piezoelectric accelerometer to monitor and store the degree and intensity of motion, averaging over every minute. The American Academy of Sleep Medicine (AASM) has noted that actigraphy is reliable and valid for detection of sleep (Morgenthaler et al., 2007). In the current study, actigraphy variables were calculated in combination with patient recordings of bedtime and rising time. Although actigraphy records a variety of sleep parameters (Berger et al., 2008), nighttime minutes awake after sleep onset (WASO) was selected for analysis in the current study based on its significant concurrent relationship with daily fatigue (Jim, et al., 2011).

Statistical Analysis

Latent change score (LCS) modeling was conducted to examine lead-lagged relationships among pairs of symptoms (McArdle, 2009). LCS modeling is a type of structural equation modeling in which a symptom score (e.g., symptom A[t]) is considered to be the sum of the symptom score on the preceding day (e.g., symptom A[t-1]) plus a latent change score (e.g. ΔA[t]). Relationships can then be modeled between ΔA[t] and a different symptom on the previous day (e.g., symptom B[t-1]). Relationships are assumed to be time invariant. Univariate LCS models were applied first to generate starting values for bivariate LCS models. Four sets of bivariate models were then created for each symptom pair (i.e., nighttime awakenings and fatigue, fatigue and depressed mood). The first model was a no coupling model, in which changes in one symptom were not dependent on the prior status of the second and vice versa. This model served as a baseline and statistical fit parameters (i.e., χ2, CFI) were compared to later models to evaluate whether the addition of subsequent paths improved model fit (Lowery, 2011). In the next two models, either a path predicting change in symptom B from symptom A or a path predicting change in symptom A from symptom B were added. A final dual coupling model specified that changes in each symptom were predicted by the prior status of the other. LCS analyses were conducting using Mplus (Muthen & Muthen, 2004) with syntax generated by Zhang (Zhang, 2005).

Results

Sample Descriptives

The sample (N=78) had a mean age of 63 years (range 33-87). The majorities of participants were Caucasian (91%), non-Hispanic (97%), married (64%), had completed high school (89%), and had an annual household income of $40,000 a year or more (52%). Participants were diagnosed with ovarian (41%), endometrial (28%), uterine (10%), cervical (6%), or other gynecologic cancers (14%). Nineteen percent had Stage I disease, 14% Stage II, 51% Stage III, and 15% Stage IV. All patients received platinum-based chemotherapy (91% carboplatin, 5% cisplatin, 4% oxaliplatin) as a single agent (4%) or in combination with paclitaxel (78%), docetaxel (17%), gemcitabine (9%), bevacizumab (4%), and/or topotecan (1%). Patients were a median of 45 days from surgery (range 13-4868) and 67% were undergoing first-line chemotherapy (i.e., had not recurred).

Univariate Models of Sleep Disturbance, Fatigue, and Depressed Mood

Univariate LCS models were created for each of the three symptoms: WASO [χ2(28)=53.34, p<.01, CFI=.63], fatigue [χ2(28)=92.10, p<.01, CFI=.85], and depressed mood [χ2(28)=92.73, p<.01, CFI=.87]. Table 1 displays the parameter estimates for each of the univariate models. As a guide to this table, the intercept mean μ0 represents the value on the day of the first chemotherapy infusion. The slope mean (μs) and proportion (β) describe constant changes over time and those that are dependent on previous levels of symptoms, respectively. The intercept variance (σ02) and slope variance (σs2) gauge the extent of individual differences in initial symptoms, as well as constant change over time. Residual variance (σε2) represents the amount of unexplained variance, and the relationship between intercept and slope is scaled as a correlation (ρ, μ0, μs). To describe the univariate models graphically, implied daily symptom means are displayed in Figure 1. Higher scores indicate greater symptomatology; thus, increases were seen in all symptoms with the greatest increase observed in WASO.

Table 1. Univariate LCS Model Results for Individual Symptoms.

WASO Fatigue Depressed Mood
Proportion β −.49 (.18) −.23 (.14) −.13 (.11)
Intercept mean μ0 −.03 (.17) 3.18 (.35)** 1.32 (.27)**
Slope mean μS .48 (.13)** 1.01 (.55) .34 (.19)
Intercept variance σ02 .06 (.36) 5.16 (1.23)** 4.32 (.88)**
Slope variance σs2 .11 (.08) .41 (.35) .19 (.13)
Residual variance σε2 1.06 (.09)** 1.79 (.15) .98 (.08)
ρ, μ0, μs .19 (.10) .71 (.58) .21 (.35)
Open in a new tab

Note. WASO – nighttime minutes awake after sleep onset. Unstandardized estimates and standard errors shown.

*

p<.05

**

p<.01.

Figure 1.

Figure 1

Open in a new tab

Implied means from the univariate LCS models for self-reported fatigue and depressed mood (panel a) and nightly minutes awake after sleep onset (panel b)

Bivariate Models of Sleep Disturbance, Fatigue, and Depressed Mood

Because WASO displayed a right-tailed distribution, it was recoded (i.e., 1 = 20 minutes or less, 2 = 21-40 minutes, 3 = 41-60 minutes, 4 = 61-90 minutes, 5 = 91-120 minutes, 6 = 121 or more minutes). Table 2 displays a summary of model fit indices for each symptom pair. In this table, the change in chi-square (Δχ2) indicates whether any improvement in model fit is statistically significant, relative to the number of additional parameters that have been added to the model. Regarding WASO and fatigue, the dual coupling model was retained because it provided statistically significant better fit than the other models (p<.01). Within the dual coupling model, the path predicting changes in fatigue from prior WASO was statistically significant (p<.05), while the path predicting changes in WASO from fatigue was not (p=.10). To aid in interpretation of lagged relationships over time, we graphed the dual coupling model of WASO and fatigue by using a hypothetical scenario in which the initial sample means for WASO varied by half a standard deviation and initial sample means for fatigue were held constant (Figure 2a), and vice versa (Figure 2b) in the dual coupling model. As shown in Figure 2a, which depicts the statistically-significant effects of prior WASO on fatigue, poor sleepers reported peak fatigue at day 2, while good sleepers reported peak fatigue at day 6. Figure 2b depicts the non-significant effects of prior fatigue on WASO, in which varying levels of fatigue show minimal linear differences in WASO over time.

Table 2. Bivariate LCS Model Results for Symptom Pairs.

Symptom/Model χ2 (df) Δχ2 (df) CFI
WASO and fatigue
 No coupling 268.79 (100) .71
 WASO → fatigue 249.16 (99) 19.63 (1)** .74
 Fatigue → WASO 268.26 (99) .53 (1) .71
 Dual coupling 241.00 (98) 27.79 (2)** .75
Fatigue and depressed mood
 No coupling 256.16 (100) .86
 Depressed mood → fatigue 240.20 (99) 15.95 (1)** .88
 Fatigue → depressed mood 222.66 (99) 33.50 (1)** .89
 Dual coupling 222.24 (98) 33.92 (2)** .89
Open in a new tab

Note. WASO – nighttime minutes awake after sleep onset. Unstandardized estimates and standard errors shown.

*

p<.05

**

p<.01.

Figure 2.

Figure 2

Figure 2

Open in a new tab

Graphs indicating dynamic relationships between nightly minutes awake after sleep onset (WASO) and lagged fatigue (panel a), fatigue and lagged WASO (panel b) and fatigue and lagged depression (panel c). As a guide to interpreting these figures, for a given pair of symptoms, the y axis indicates model-implied sample means of the lagging symptom when initial levels of the lead symptom are varied by half a standard deviation and initial sample means for the lagged symptom are kept constant.

Regarding fatigue and depressed mood, the model in which fatigue was a leading indicator of changes in depressed mood was retained because it provided significantly better fit than the no coupling model and the model in which depressed mood was a leading indicator of fatigue (ps<.01) and the dual coupling model did not provide better fit (p=.52). A graph of implied means from the single coupling model depicting the statistically-significant effects of prior fatigue on depressed mood (Figure 2c) indicated that patients with high fatigue showed greater subsequent depressed mood over time compared to patients with moderate or low fatigue.

Discussion

The current study examined lagged relationships among sleep disturbance, fatigue, and depressed mood in the week after gynecologic cancer patients’ first chemotherapy infusions. Results indicate that increased nighttime awakenings are associated with earlier subsequent peaks in fatigue, while increased fatigue is associated with greater subsequent depressed mood. These results are noteworthy because they suggest that there is a temporal sequence of symptom onset during platinum-based chemotherapy for gynecologic cancer, one of the most arduous chemotherapy regimens for cancer.

Regarding the relationship between nighttime awakenings and fatigue, although a dual coupling model best fit the data, the path predicting lagged changes in fatigue from nighttime awakenings was significant, but the path predicting lagged changes in nighttime awakenings from fatigue was not. These data are interesting because they suggest that while nighttime sleep disturbance and fatigue exert reciprocal influences on one another, the influence of nighttime sleep disturbance on fatigue is stronger than vice versa. Specifically, sleep disturbance appears to affect the timing, rather than the magnitude, of fatigue after chemotherapy. This finding suggests that there are multiple mechanisms of fatigue after chemotherapy, with poor sleep contributing to early fatigue and other factors contributing to later fatigue.

The finding that fatigue contributes to greater subsequent depressed mood, but not vice versa, is noteworthy because it argues against the possibility that fatigue in cancer patients is merely a manifestation of depressed mood (Jacobsen, Donovan, & Weitzner, 2003). Fatigue and depressed mood tend to be highly correlated, but our study suggests that these constructs are distinct and that their relationship results from the distressing nature of cancer-related fatigue. As such, our findings are congruent with the clinical syndrome approach to assessing cancer-related fatigue which allows for the possibility that fatigue arising from cancer or treatment can cause clinically-significant distress or impairment (Cella, Peterman, Passik, Jacobsen, & Breitbart, 1998) .

The current study lends support for a cascade model of symptoms during chemotherapy, in which initial sleep disturbance predicts subsequent fatigue, which in turn predicts subsequent depressed mood. These findings are consistent with other data suggesting that sleep disturbance occurs first in a cascade of symptoms including fatigue and depressed mood. For example, previous studies indicate that post-chemotherapy peaks in nighttime sleep disturbance occur before peaks in fatigue and depressed mood (Berger, et al., 2009; Jim, et al., 2011). Nighttime sleep disturbance predicts next-day fatigue and depressed mood in post-treatment breast cancer survivors (Rumble, et al., 2010). In a study by Broeckel and colleagues (Broeckel, Jacobsen, Horton, Balducci, & Lyman, 1998), severe fatigue was associated with increased risk of a concurrent psychiatric disorder (e.g., depression), but not past psychiatric disorder. In addition, interventions addressing sleep disturbance tend to show improvements in both fatigue and depressed mood (Carlson & Garland, 2005; Dirksen & Epstein, 2008; Ritterband et al., 2011; Savard, Simard, Ivers, & Morin, 2005). On the other hand, interventions addressing biochemical changes associated with depressed mood lead to changes in mood, but usually not in fatigue or sleep (Morrow et al., 2003; Roscoe, et al., 2002; Roscoe et al., 2005). Taken together with previous literature, our findings suggest that interventions targeting early increases in nighttime sleep disturbance may have beneficial effects across multiple distressing symptoms in cancer patients during chemotherapy.

Strengths of the current study include intensive assessment of symptoms during chemotherapy for gynecologic cancer, as well as use of advanced statistical techniques to model lagged changes in symptoms over time. Limitations of the current study should also be noted, however. The sample size was relatively small, which together with patient attrition precluded examination of lagged symptoms over multiple chemotherapy administrations. Although all participants received platinum-based chemotherapy, there was sample heterogeneity in terms of the chemotherapy regimens received, time since surgery, and previous treatment with chemotherapy. All of these factors could be expected to influence the severity of symptom reports (Jim, et al., 2011). Due to the complexity of the LCS models, as well as the fact that we were interested in lagged relationships among symptoms rather than symptom severity per se, we elected not to control for these clinical variables in the models. We believe this decision increases the generalizability of findings to a variety of patients undergoing platinum-based chemotherapy for gynecologic cancer. Nevertheless, future research should examine whether clinical variables affect lagged relationships among symptoms. Despite evidence of significant intraday variability in fatigue in this sample (Jim, et al., 2011), we elected to analyze daily averages of fatigue and depressed mood ratings rather than multiple ratings per day. This decision may obscure significant changes in fatigue over the course of the day; future research should explore the effects of nighttime sleep on intraday changes in fatigue.

In summary, the current study is one of the first to examine lagged relationships among common and distressing symptoms in cancer patients treated with chemotherapy. Data suggest a cascade model of symptoms, in which nighttime sleep disturbance contributed to lagged increases in fatigue, which in turn contributed to lagged increases in depressed mood. Findings suggest a approach to symptom management interventions, such as targeting symptoms early in the cascade (e.g., nighttime sleep disturbance) to produce beneficial effects across multiple symptoms. Future research efforts to replicate these findings in patients diagnosed with other types of cancer will help improve our current understanding and management of cancer-related symptoms to achieve better patient outcomes.

Acknowledgements

This study was supported by the National Cancer Institute grant number R03-CA126775. Dr. Jim is supported in part by the National Cancer Institute grant number K07-CA138499. Author contributions: Jim: study design, statistical analysis, data interpretation, manuscript writing; Jacobsen: study design, data interpretation, manuscript writing; Phillips: data interpretation, manuscript writing; Wenham: referral of patients; manuscript writing; Roberts: referral of patients; Small: statistical analysis, data interpretation, manuscript writing. The authors also wish to acknowledge the contributions of the Moffitt Cancer Center Survey Methods Shared Resource.

Contributor Information

Heather S.L. Jim, Department of Health Outcomes and Behavior, Moffitt Cancer Center

Paul B. Jacobsen, Department of Health Outcomes and Behavior, Moffitt Cancer Center

Kristin M. Phillips, Department of Health Outcomes and Behavior, Moffitt Cancer Center

Robert M. Wenham, Department of Gynecologic Oncology, Moffitt Cancer Center

William Roberts, Department of Gynecologic Oncology, Moffitt Cancer Center.

Brent J. Small, Department of Aging Studies, University of South Florida

References

  1. Badr H, Basen-Engquist K, Carmack Taylor CL, De Moor C. Mood states associated with transitory physical symptoms among breast and ovarian cancer survivors. J Behav Med. 2006;29(5):461–475. doi: 10.1007/s10865-006-9052-9. [DOI] [PubMed] [Google Scholar]
  2. Banthia R, Malcarne VL, Ko CM, Varni JW, Sadler GR. Fatigued breast cancer survivors: the role of sleep quality, depressed mood, stage and age. Psychol Health. 2009;24(8):965–980. doi: 10.1080/08870440802110831. doi: 905088851 [pii] 10.1080/08870440802110831. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Bender CM, Ergyn FS, Rosenzweig MQ, Cohen SM, Sereika SM. Symptom clusters in breast cancer across 3 phases of the disease. Cancer Nurs. 2005;28(3):219–225. doi: 10.1097/00002820-200505000-00011. doi: 00002820-200505000-00011 [pii] [DOI] [PubMed] [Google Scholar]
  4. Berger AM. Patterns of fatigue and activity and rest during adjuvant breast cancer chemotherapy. Oncol Nurs Forum. 1998;25(1):51–62. [PubMed] [Google Scholar]
  5. Berger AM, Wielgus K, Hertzog M, Fischer P, Farr L. Patterns of circadian activity rhythms and their relationships with fatigue and anxiety/depression in women treated with breast cancer adjuvant chemotherapy. Support Care Cancer. 2009 doi: 10.1007/s00520-009-0636-0. [DOI] [PubMed] [Google Scholar]
  6. Berger AM, Wielgus KK, Young-McCaughan S, Fischer P, Farr L, Lee KA. Methodological challenges when using actigraphy in research. J Pain Symptom Manage. 2008;36(2):191–199. doi: 10.1016/j.jpainsymman.2007.10.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Broeckel JA, Jacobsen PB, Horton J, Balducci L, Lyman GH. Characteristics and correlates of fatigue after adjuvant chemotherapy for breast cancer. J Clin Oncol. 1998;16(5):1689–1696. doi: 10.1200/JCO.1998.16.5.1689. [DOI] [PubMed] [Google Scholar]
  8. Butler L, Bacon M, Carey M, Zee B, Tu D, Bezjak A. Determining the relationship between toxicity and quality of life in an ovarian cancer chemotherapy clinical trial. J Clin Oncol. 2004;22(12):2461–2468. doi: 10.1200/JCO.2004.01.106. [DOI] [PubMed] [Google Scholar]
  9. Byar KL, Berger AM, Bakken SL, Cetak MA. Impact of adjuvant breast cancer chemotherapy on fatigue, other symptoms, and quality of life. Oncol Nurs Forum. 2006;33(1):E18–26. doi: 10.1188/06.ONF.E18-E26. doi: 10.1188/06.ONF.E18-E26. [DOI] [PubMed] [Google Scholar]
  10. Carlson LE, Garland SN. Impact of mindfulness-based stress reduction (MBSR) on sleep, mood, stress and fatigue symptoms in cancer outpatients. Int J Behav Med. 2005;12(4):278–285. doi: 10.1207/s15327558ijbm1204_9. [DOI] [PubMed] [Google Scholar]
  11. Cella D, Peterman A, Passik S, Jacobsen P, Breitbart W. Progress toward guidelines for the management of fatigue. Oncology (Williston Park) 1998;12(11A):369–377. [PubMed] [Google Scholar]
  12. Dirksen SR, Epstein DR. Efficacy of an insomnia intervention on fatigue, mood and quality of life in breast cancer survivors. J Adv Nurs. 2008;61(6):664–675. doi: 10.1111/j.1365-2648.2007.04560.x. [DOI] [PubMed] [Google Scholar]
  13. Donovan KA, Jacobsen PB. Fatigue, depression, and insomnia: evidence for a symptom cluster in cancer. Semin Oncol Nurs. 2007;23(2):127–135. doi: 10.1016/j.soncn.2007.01.004. doi: S0749-2081(07)00014-9 [pii] 10.1016/j.soncn.2007.01.004. [DOI] [PubMed] [Google Scholar]
  14. Ferrer E, McArdle JJ. Longitudinal modeling of developmental changes in psychological research. Curr Dir Psychol Sci. 2010;19(3):149–154. [Google Scholar]
  15. Ghisletta P, Bickel J-F, Lovden M. Does activity engagement protect against cognitive decline in old age? Methodological and analytical considerations. J Gerontol B Psychol Sci Soc Sci. 2006;61B:P253–P261. doi: 10.1093/geronb/61.5.p253. [DOI] [PubMed] [Google Scholar]
  16. Goncalves V, Jayson G, Tarrier N. A longitudinal investigation of psychological morbidity in patients with ovarian cancer. Br J Cancer. 2008;99(11):1794–1801. doi: 10.1038/sj.bjc.6604770. doi: 6604770 [pii] 10.1038/sj.bjc.6604770. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Huang TW, Lin CC. The mediating effects of depression on sleep disturbance and fatigue: symptom clusters in patients with hepatocellular carcinoma. Cancer Nurs. 2009;32(5):398–403. doi: 10.1097/NCC.0b013e3181ac6248. doi: 10.1097/NCC.0b013e3181ac6248. [DOI] [PubMed] [Google Scholar]
  18. Infurna FJ, Gerstorf D, Zarit SH. Examining dynamic links between perceived control and health: Longitudinal evidence for differential effects in midlife and old age. Dev Psychol. 2011;47(1):9–18. doi: 10.1037/a0021022. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Jacobsen PB, Donovan KA, Weitzner MA. Distinguishing fatigue and depression in patients with cancer. Semin Clin Neuropsychiatry. 2003;8(4):229–240. doi: S1084361203000492 [pii] [PubMed] [Google Scholar]
  20. Jacobsen PB, Hann DM, Azzarello LM, Horton J, Balducci L, Lyman GH. Fatigue in women receiving adjuvant chemotherapy for breast cancer: characteristics, course, and correlates. J Pain Symptom Manage. 1999;18(4):233–242. doi: 10.1016/s0885-3924(99)00082-2. [DOI] [PubMed] [Google Scholar]
  21. Jim HS, Small B, Faul LA, Franzen J, Apte S, Jacobsen PB. Fatigue, Depression, Sleep, and Activity During Chemotherapy: Daily and Intraday Variation and Relationships Among Symptom Changes. Ann Behav Med. 2011 doi: 10.1007/s12160-011-9294-9. doi: 10.1007/s12160-011-9294-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Liu L, Fiorentino L, Natarajan L, Parker BA, Mills PJ, Sadler GR, Ancoli-Israel S. Pre-treatment symptom cluster in breast cancer patients is associated with worse sleep, fatigue and depression during chemotherapy. Psychooncology. 2009;18(2):187–194. doi: 10.1002/pon.1412. doi: 10.1002/pon.1412. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Lowery R. VassarStats Statistical Computation Website. 2011 Retrieved Aug. 3, 2011, from http://faculty.vassar.edu/lowry/tabs.html#csq.
  24. McArdle JJ. Latent variable modeling of differences and changes with longitudinal data. Ann Rev Psychol. 2009;60(1):577–605. doi: 10.1146/annurev.psych.60.110707.163612. [DOI] [PubMed] [Google Scholar]
  25. McArdle JJ, Hamgami F, Jones K, Jolesz F, Kikinis R, Spiro A, 3rd, Albert MS. Structural modeling of dynamic changes in memory and brain structure using longitudinal data from the normative aging study. J Gerontol B Psychol Sci Soc Sci. 2004;59(6):P294–304. doi: 10.1093/geronb/59.6.p294. doi: 59/6/P294 [pii] [DOI] [PubMed] [Google Scholar]
  26. Morgan RJ, Alvarez RD, Armstrong DK, Boston B, Burger R, Chen L, Teng N. NCCN Clinical Practice Guidelines in Oncology: Ovarian Cancer. 2009 Retrieved Sept. 23, 2009, from http://www.nccn.org/professionals/physician_gls/PDF/ovarian.pdf.
  27. Morgenthaler T, Alessi C, Friedman L, Owens J, Kapur V, Boehlecke B, Swick TJ. Practice parameters for the use of actigraphy in the assessment of sleep and sleep disorders: an update for 2007. Sleep. 2007;30(4):519–529. doi: 10.1093/sleep/30.4.519. [DOI] [PubMed] [Google Scholar]
  28. Morrow GR, Hickok JT, Roscoe JA, Raubertas RF, Andrews PL, Flynn PJ, King DK. Differential effects of paroxetine on fatigue and depression: a randomized, double-blind trial from the University of Rochester Cancer Center Community Clinical Oncology Program. J Clin Oncol. 2003;21(24):4635–4641. doi: 10.1200/JCO.2003.04.070. [DOI] [PubMed] [Google Scholar]
  29. Muthen LK, Muthen BO. MPlus (Version 6.1) Muthen and Muthen; Los Angeles: 2004. [Google Scholar]
  30. Palesh OG, Roscoe JA, Mustian KM, Roth T, Savard J, Ancoli-Israel S, Morrow GR. Prevalence, demographics, and psychological associations of sleep disruption in patients with cancer: University of Rochester Cancer Center-Community Clinical Oncology Program. J Clin Oncol. 2010;28(2):292–298. doi: 10.1200/JCO.2009.22.5011. doi: JCO.2009.22.5011 [pii] 10.1200/JCO.2009.22.5011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Ritterband LM, Bailey ET, Thorndike FP, Lord HR, Farrell-Carnahan L, Baum LD. Initial evaluation of an Internet intervention to improve the sleep of cancer survivors with insomnia. Psychooncology. 2011 doi: 10.1002/pon.1969. doi: 10.1002/pon.1969. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Roscoe JA, Morrow GR, Hickok JT, Bushunow P, Matteson S, Rakita D, Andrews PL. Temporal interrelationships among fatigue, circadian rhythm and depression in breast cancer patients undergoing chemotherapy treatment. Support Care Cancer. 2002;10(4):329–336. doi: 10.1007/s00520-001-0317-0. [DOI] [PubMed] [Google Scholar]
  33. Roscoe JA, Morrow GR, Hickok JT, Mustian KM, Griggs JJ, Matteson SE, Smith B. Effect of paroxetine hydrochloride (Paxil) on fatigue and depression in breast cancer patients receiving chemotherapy. Breast Cancer Res Treat. 2005;89(3):243–249. doi: 10.1007/s10549-004-2175-1. [DOI] [PubMed] [Google Scholar]
  34. Rumble ME, Keefe FJ, Edinger JD, Affleck G, Marcom PK, Shaw HS. Contribution of cancer symptoms, dysfunctional sleep related thoughts, and sleep inhibitory behaviors to the insomnia process in breast cancer survivors: a daily process analysis. Sleep. 2010;33(11):1501–1509. doi: 10.1093/sleep/33.11.1501. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Savard J, Simard S, Ivers H, Morin CM. Randomized study on the efficacy of cognitive-behavioral therapy for insomnia secondary to breast cancer, part I: Sleep and psychological effects. J Clin Oncol. 2005;23(25):6083–6096. doi: 10.1200/JCO.2005.09.548. [DOI] [PubMed] [Google Scholar]
  36. Small BJ, Dixon RA, McArdle JJ, Grimm KJ. Do changes in lifestyle engagement moderate cognitive decline in normal aging? Evidence from the Victoria Longitudinal Study. Neuropsychology. 2011 doi: 10.1037/a0026579. doi: 2011-28657-001 [pii] 10.1037/a0026579. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Stepanski EJ, Walker MS, Schwartzberg LS, Blakely LJ, Ong JC, Houts AC. The relation of trouble sleeping, depressed mood, pain, and fatigue in patients with cancer. J Clin Sleep Med. 2009;5(2):132–136. [PMC free article] [PubMed] [Google Scholar]
  38. Zhang Z. A c++ program to generate the Mplus bivariate latent difference score model codes. 2005 Retrieved May 10, 2011, from http://www.psychstat.org/us/article.php/38.htm.

RESOURCES