• childhood cancer survivors;
  • sedentary lifestyle;
  • intervention strategies;
  • physical activity


  1. Top of page
  2. Abstract
  6. Acknowledgements
  7. Conflict of Interest Disclosures
  8. References


Although physical activity may modify the late effects of childhood cancer treatment, from 20% to 52% of adult survivors are sedentary. The authors of this report sought to identify modifiable factors that influence survivors' participation in physical activity.


Structural equation modeling of data were derived from the Childhood Cancer Survivors Study of adult survivors (current mean age, 30.98 years; mean years since diagnosis, 23.74; mean age at diagnosis, 9.25 years) who were diagnosed between 1970 and 1986.


Approximately 40% of the variance in male survivors' recent participation versus nonparticipation in physical activity was explained directly and/or indirectly by self-reported health fears (P = .01), perceived primary-care physician (PCP) expertise (P = .01), baseline exercise frequency (P ≤ .001), education level (P = .01), self-reported stamina (P = .01), cancer-related pain (P ≤ .001), fatigue (P ≤ .001), age at diagnosis (P = .01), cancer-related anxiety (P ≤ .001), motivation (P = .01), affect (P = .01), and discussion of subsequent cancer risk with the PCP (P ≤ .001) (N = 256; chi-square test statistic = 53.38; degrees of freedom [df] = 51; P = .38, Comparative Fit Index [CFI] = 1.000; Tucker Lewis Index [TLI] = 1.000; root mean square of approximation [RMSEA] = 0.014; weighted root mean square residual [WRMR] = 0.76). Thirty-one percent of the variance in women' recent physical activity participation was explained directly and/or indirectly by self-reported stamina (P ≤ .001), fatigue (P = .01), baseline exercise frequency (P = .01), cancer-related pain (P ≤ .001), cancer-related anxiety (P = .01), recency of visits with PCP (<0.001), quality of interaction with the PCP (P = .01), and motivation (P ≤ .001; N = 366; chi-square test statistic = 67.52; df = 55; P = .12; CFI = 0.98; TLI = 0.98; RMSEA = 0.025; WRMR = 0.76).


Gender-tailored intervention strategies in which providers specifically target motivation, fear, and affect may support physical activity in childhood cancer survivors. Cancer 2009. © 2008 American Cancer Society.

Increases in childhood cancer survival have shifted the care paradigm from a cure to a long-term emphasis on treatment-related morbidity and quality of life. Among the potential consequences of cancer therapy are pain, fatigue, obesity, osteoporosis, cardiomyopathy, and neuromusculoskeletal complications.1-4 Lifelong physical activity initiated during cancer therapy is essential to modify or prevent these late effects.5-8

Recent reports indicate that from 20% to 52% of adult survivors of childhood cancer are sedentary, despite their risk of late effects9-12; this rate is similar to that in the general population.13, 14 However, survivors of acute lymphoblastic leukemia (ALL), the most common pediatric cancer, were more likely than members of the general population to report no leisure-time physical activity (odds ratio [OR], 1.74; 95% confidence interval [95% CI], 1.56-1.94).15 Survivors who were treated with >20 grays of cranial radiotherapy were at particular risk. When the comparison was adjusted for age, race, and ethnicity, ALL survivors were substantially more likely than their healthy counterparts to be inactive (females: OR, 1.86; 95% CI, 1.50-2.31; males: OR, 1.84; 95% CI, 1.45-2.32).

Factors that positively or negatively influence childhood cancer survivors' physical activity have not been documented. To describe modifiable influences on survivors' physical activity participation, we selected a broad-based health behavior model adapted to childhood cancer survivors14-16 to inform variable selection, relations, and structural equation modeling (SEM).


  1. Top of page
  2. Abstract
  6. Acknowledgements
  7. Conflict of Interest Disclosures
  8. References

Conceptual Model

The Interaction Model of Client Health Behavior16-18 (Fig. 1) has been used18-21 to identify survivor, provider, and contextual factors that can be targeted by interventions to modify the late effects of cancer therapy. The model comprises 3 broad elements: 1) the intrapersonal (affect, motivation, cognitive appraisal) and contextual (demographic/social influences, health history, resources) characteristics that uniquely define individuals relative to their health, disease, and treatment; 2) the therapeutic content and process between providers and patients that can support or impede behavior change; and 3) behaviorally related health outcomes. All variables within the datasets that corresponded to the conceptual model17, 18 were evaluated in preliminary models (Fig. 1); only those variables that contributed significantly (P≤.01) to the explained variance in the physical activity outcome were retained in the final models.

thumbnail image

Figure 1. The Interaction Model of Client Health Behavior applied to physical activity in childhood cancer survivors. PCP indicates primary-care physician; CA, cancer; dx, diagnosis.

Download figure to PowerPoint

Data Source and Sample

The Childhood Cancer Survivor Study (CCSS) was designed to enroll a cohort of survivors sufficiently large and diverse to allow meaningful analysis of the delayed effects of treatment. The surveys and sampling methods have been described in detail by Robison et al22 and are available for review online at accessed on September 12, 2008. In 1994, participants completed a baseline questionnaire that addressed demographic and health-related information; in addition, participating centers provided detailed abstracts of the medical records for consenting subjects. The original CCSS study contacted 20,346 eligible participants from 26 institutions across the US and Canada who had survived ≥5 years after treatment for diagnosed malignant disease (before age 21 years) between 1970 and 1986. A subsequent ancillary study, the Health Care Needs Survey, randomly sampled 1600 of approximately 12,872 survivors who remained alive; 978 survivors (61%) completed and returned the survey, and 838 of the 978 survivors (86%) returned the total-cohort second follow-up 2 survey within the same data collection period. Those 838 survivors were the subjects of this analysis.


Dependent Variable

Consistent with previous CCSS reports,15 physical activity participation was measured as a binary outcome derived from the Behavioral Risk Factor Surveillance Study.23 Survivors were asked (yes = 1; no = 2), “During the past month, did you participate in any physical activities or exercises, such as running, calisthenics, golf, bicycling, swimming, wheelchair basketball, or walking, for exercise?”

Directly Observed, Independent Variables

Eleven directly observed, independent variables were associated significantly with recent physical activity as covariates in the final analytical models. All were self-reported by survivors: 1) their primary-care physician's (PCP's) familiarity with cancer-related problems (1 = familiar, 2 = not familiar); 2) current pain resulting from cancer or its treatment (1 = no pain, 5 = excruciating pain); 3) frequency of fatigue (1 = all the time, 6 = none of the time); 4) whether survivors had discussed the risk of recurrent cancer with their PCP (1 = yes, 2 = no); 5) baseline frequency of aerobic exercise (defined as sufficient to induce sweating or breathing hard, lasting ≥20 minutes 0 days per week or 7 days per week); 6) age at diagnosis; 7) current anxiety as a result of cancer or its treatment (1 = no anxiety, 5 = extreme anxiety); 8) current highest school grade completed; 9) whether the survivor had seen a primary care physician since cancer treatment ended (1 = yes, 2 = no); 10) modified from the Multidimensional Health Locus of Control Scales,24 for intrinsic motivation, survivors rated 4 items (eg, “I am in control of my health.” “The main thing that affects my health is what I myself do”) on a 6-point Likert scale (1 = strongly disagree, 6 = strongly agree; α = .79); 11) for extrinsic motivation, survivors rated 4 items (eg, “Health professionals control my health”; “Regarding my health, I can only do what my physician tells me to do”) on a 6-point Likert scale (1 = strongly disagree, 6 = strongly agree; α = .80).

Latent Independent Variables

In structural equation modeling, latent variables (eg, depression) are measured indirectly by using a set of observed variables.25 Our final analyses identified 4 latent measures that contributed (directly or indirectly) to the explained variance in physical activity participation. The strength of the latent measures (ie, the cohesiveness and fit of their contributory observed variables) was assessed by confirmatory factor analysis. Four strong latent variables (factor score determinacy values >0.80) were derived from the observed variables: survivor-provider interaction (0.94), fear (0.98), affect (0.96), and stamina (0.98).


Derived from 5 observed items on the Physical Function measure of the 36-item Short Form Health Survey (SF-36),26 survivors were asked about the extent to which their physical health limited their ability to climb several flights of stairs, climb 1 flight of stairs, walk more than 1 mile, walk several blocks, and walk 1 block (3 = not limited at all, 2 = limited a little, 1 = limited a lot; α = .924).


Three observed variables measured on a 5-point Likert scale assessed the extent of survivors' fear (1 = not at all concerned, 5 = extremely concerned) about their future health, the return of their cancer, and the discovery of a health problem during a routine check-up (α = .76).


Derived from 3 observed measures on the SF-36 Mental Health subscale26 and rated on a 6-point Likert scale (1 = all of the time, 6 = none of the time), survivors assessed the frequency of feeling unhappy, downhearted and blue, and not cheerful (α = .78).

Survivor-provider Interaction

Four items asked survivors to rate on a 5-point Likert scale (1 = not at all, 5 = extremely) the extent to which they believed that their physician took enough time to answer their questions, they could ask their physician questions about cancer, their fears and concerns had been addressed by their physician, and their PCP could handle cancer-related problems (α = .78).

Statistical Analysis

Structural equation models of the data were analyzed using Mplus 4.2 software.25 Subsamples of men (N = 256) and women (N = 366) with complete data comprised the final sample for the SEM analysis. We chose to use samples with complete data rather than data imputation to avoid potentially distorting coefficients of association and correlation relating variables.27 A sample size >200 is considered large in SEM.28 The best-fitting model is 1 that is sound conceptually, has parameter estimates that correspond significantly to the hypothesized relations, meets the established SEM fit criteria (Figs. 2 and 3), and offers the highest percentage of explained variance for the outcome.

thumbnail image

Figure 2. Predictors of physical activity participation in male survivors. Note that a nonsignificant chi-square text statistic (χ2) measures the absolute fit of the model to the data but is sensitive to sample size (see Bentler, 199075). The Comparative Fit Index (CFI) and the Tucker Lewis Index (TLI) test the proportionate improvement in fit by comparing the target model to an independent base model; a value of 0.90 is minimally acceptable (see Bollen, 199076), values approximating 0.95 indicate a good fit, and values at or close to 1.000 indicate an excellent fit (see Browne and Cudeck, 199377). The root mean square of approximation (RMSEA) represents closeness of fit, and values approximating 0.06 and 0.00 demonstrate a close and an exact fit of the model, respectively (see Browne and Cudeck, 199377 and Hu and Bentler, 199978). The weighted root mean square residual (WRMR) is a better indicator of fit for binary measures, and acceptable values are ≤0.80. CA indicates cancer; DX, diagnosis; df, degrees of freedom.

Download figure to PowerPoint

thumbnail image

Figure 3. Predictors of physical activity participation in female survivors. Note that a nonsignificant chi-square text statistic (χ2) measures the absolute fit of the model to the data but is sensitive to sample size (see Bentler, 199075). The Comparative Fit Index (CFI) and the Tucker Lewis Index (TLI) test the proportionate improvement in fit by comparing the target model to an independent base model; a value of 0.90 is minimally acceptable (see Bollen, 199076), values approximating 0.95 indicate a good fit, and values at or close to 1.000 indicate an excellent fit (see Browne and Cudeck, 199377). The root mean square of approximation (RMSEA) represents closeness of fit, and values approximating 0.06 and 0.00 demonstrate a close and an exact fit of the model, respectively (see Browne and Cudeck, 199377 and Hu and Bentler, 199978). The weighted root mean square residual (WRMR) is a better indicator of fit for binary measures, and acceptable values are ≤0.80 (df indicates degrees of freedom).

Download figure to PowerPoint


  1. Top of page
  2. Abstract
  6. Acknowledgements
  7. Conflict of Interest Disclosures
  8. References


There were no demographic or cancer-related differences between the total sample with missing data (N = 838) and the SEM subsamples with no missing data (Table 1). Significant differences observed between the sex subsamples and the total sample (using Bonferonni correction factor) were observed in the subsamples with complete data: Men reported higher income categories; women were more likely to have visited their PCP since they completed cancer treatment (Table 1). The typical respondent was representative of the entire CCSS cohort22, 29, 30: white, woman, college graduate, earning from $20,000 to $60,000, and covered by health insurance (Table 1). Approximately 25% percent of respondents in the total sample reported no leisure-time physical activity; women in the SEM subsample were slightly more active (P value nonsignificant). Subsample women were more likely to be fearful of finding a problem at check-up than were subsample men. Subsample men and women reported less fatigue than the total sample, with men reporting less fatigue than women. Subsample women were less motivated extrinsically than subsample men (Table 2). Because sex differences in physical activity are reported in the general population31-34 and were observed in the study's independent variables (Table 2), we estimated separate models for men and women survivors rather than testing model equivalency between sexes.

Table 1. Descriptive Summary of the Total Sample and Structural Equation Modeling Subsamples*
CharacteristicTotal Sample: No. (%)Chi-Square PSEM Subsambles: No. (%)Chi-Square P
  • SEM indicates Structural Equation Modeling; NA, not applicable; GED, General Education Degree; SD, standard deviation.

  • *

    Numbers vary because of missing data.

  • Differences between total sample, total men, and total women

  • Differences between total sample, SEM men, and SEM women.

Sex   NA   
 Men385 (45.9)NANA NANA 
 Women453 (54.1)NANA NANA 
Race   .235  .484
 White624 (74.4)288 (74.8)336 (74.7) 225 (74.5)280 (75.7) 
 Black64 (7.7)24 (6.2)40 (8.9) 18 (6)28 (7.6) 
 Hispanic96 (11.5)41 (10.6)55 (12.2) 34 (11.3)46 (12.4) 
 Other51 (6.1)32 (8.3)19 (4.2) 25 (8.3)16 (4.3) 
Diagnosis   .631  .279
 Leukemia, Hodgkin disease, lymphoma593 (66.7)224 (64.4)275 (66.1) 176 (63.5)225 (65.6) 
 Solid tumor205 (23.1)80 (23)102 (24.5) 62 (22.4)87 (25.4) 
 Bone cancer91 (10.2)44 (12.6)39 (9.4) 39 (14.1)31 (9) 
Personal annual income, $US   <.001  <.001
 None94 (11.5)18 (4.8)76 (17.4) 16 (5.4)67 (18.6) 
 <19,999278 (34.2)105 (27.9)173 (39.6) 80 (26.8)138 (38.2) 
 20,000-39,999212 (26.1)104 (27.6)108 (24.7) 81 (27.1)88 (24.4) 
 40,000-59,999123 (15.1)68 (18)55 (12.6) 51 (17.1)46 (12.7) 
 60,000-79,99943 (5.3)31 (8.2)12 (2.7) 29 (9.7)10 (2.8) 
 80,000-99,99927 (3.3)23 (6.1)4 (0.9) 18 (6)4 (1.1) 
 >100,00037 (4.5)28 (7.4)9 (2.1) 24 (8)8 (2.2) 
Marital status   .106  .097
 Ever married329 (39.4)136 (35.5)193 (42.7) 106 (35.2)161 (43.4) 
 Never married506 (60.6)247 (64.5)259 (57.3) 195 (64.8)210 (56.6) 
Current health insurance   .994  .746
 Yes732 (88.1)336 (88)396 (88.2) 269 (89.7)325 (88.1) 
 No99 (11.9)46 (12)53 (11.8) 31 (10.3)44 (11.9) 
Education, y   .997  .955
 1-82 (0.2)1 (0.3)1 (0.2) 1 (0.3)0 (0) 
 9-1218 (2.2)11 (2.9)7(1.6) 5 (1.7)5 (1.4) 
 Completed high school/GED106 (12.8)47 (12.4)59 (13.2) 33 (11.1)45 (12.3) 
 Training after high school other than college42 (5.1)22 (5.8)20 (4.5) 15 (5.1)15 (4.1) 
 Some college263 (31.8)120 (31.6)143 (31.9) 100 (33.7)110 (30) 
 College graduate268 (32.4)122 (32.1)146 (32.6) 96 (32.3)126 (34.3) 
 Postgraduate education129 (15.6)57 (15)72 (16.1) 47 (15.8)66 (18) 
Physical activity during past mo   .987  .726
 Yes629 (75.1)290 (75.3)339 (74.8) 230 (76.2)287 (77.2) 
 No209 (24.9)95 (24.7)114 (25.2) 72 (23.8)85 (22.8) 
CharacteristicTotal Sample: Mean ± SD, yF-Ratio PSEM Subsamples: Mean ± SD, yF-Ratio P
Age30.98 ± 7.5031.06 ± 7.5430.92 ± 7.47.97630.99 ± 7.4231.05 ± 7.41.987
Age at diagnosis9.25 ± 5.879.38 ± 5.789.13 ± 5.95.8389.39 ± 5.789.25 ± 5.97.942
Time since diagnosis21.74 ± 4.5421.67 ± 4.5221.79 ± 4.55.92721.60 ± 4.4021.81 ± 4.58.847
Table 2. Comparison of Total Sample and Structural Equation Modeling Subsamples on Study Measures
    SEM Subsamples   
 Total SampleMen, N=256Women, N=363F-Ratio PBonferroni Post Hoc P
Continuous VariablesNo.MeanSDMeanSDMeanSDBetween Total and SEM SubsamplesBetween Total and SEM SubsamplesBetween SEM Men and SEM Women
  • SEM indicates Structural Equation Modeling; SD, standard deviation; PCP, primary-care physician; T, total sample; SF, SEM women; NS, nonsignificant; SM, SEM men.

  • *

    Differences between the total sample (with missing data) and SEM men and women with no missing data.

PCP takes time to answer questions8223.351.103.471.  
Can ask questions about cancer8283.821.214.013.653.741.23.206  
PCP addresses fears/concerns8253.531.243.651.173.531.25.368  
Fear about future health8322.861.082.751.092.951.04.073  
Fear of cancer recurrence8332.391.  
Fear problem at check-up8322.
Pain as a result of cancer8161.360.751.340.731.410.79.460  
Cancer anxiety8141.560.861.470.781.640.91.052  
Exercise in last 7 d8312.272.192.492.302.102.10.092  
Climb several stairs7612.620.642.720.592.570.66.015  
Climb 1 flight of stairs7632.840.442.870.412.830.46.520  
Walk >1 mile7592.650.652.710.592.620.68.230  
Walk several blocks7632.750.572.800.492.720.62.227  
Walk 1 block7642.870.422.890.382.860.45.680  
Very depressed7445.311.075.401.  
Feel blue7394.921.  
Feel tired7383.861.; T/SF=.008.001
Not happy7414.341.194.431.  
Intrinsic motivation82518.193.6418.573.5517.883.72.070  
Extrinsic motivation8257.953.298.363.257.543.29.009T/SF=NS.002
    SEM Subsamples   
Total SampleMenWomen
Categorical/Nominal VariablesNo.%No.%No.%Chi-square P*  
Visited PCP since cancer treatment completed       .055  
 Yes 28134.18929.714038.6   
 No 54465.92170.322361.4   
PCP discussed risk of recurrent cancer       .004  
 Yes 18323.94416.59828   
 No 58376.122383.525272   
Confident of PCP knowledge of problems after cancer       .088  
 Yes 31344.71174812538.1   
 No 14220.34317.68325.3   
 No check-up 24635.18434.412036.6   

The Physical Activity Model for Men

The likelihood of men survivors' physical activity participation during the past month was predicted by the perception that their PCP was familiar with cancer-related problems, greater fear regarding future health, more education, and greater baseline frequency of aerobic exercise (Fig. 2, Table 3). Younger age at diagnosis, infrequent fatigue, little or no cancer pain, and infrequent negative affect predicted greater stamina. Less fear concerning future health, infrequent fatigue, and less anxiety about cancer predicted a more positive affect. Greater fear was predicted by more cancer pain, having discussed the risk of cancer in the future with their PCP, higher levels of extrinsic motivation, and lower levels of intrinsic motivation. Perceptions of less stamina, greater cancer anxiety, physical activity nonparticipation, and greater cancer pain predicted frequent fatigue. Greater fear concerning future health, greater cancer pain, and less intrinsic motivation predicted greater cancer anxiety (a correlate of physical activity participation).

Table 3. Structural Equation Modeling Results for Men and Women Survivors
VariableEstimateSEEstimate/SE*Standard YX
  • SE indicates standard error; PCP, primary-care physician; RX, treatment.

  • *

    Z scores of 1.96 were significant at P=.05, and Z score of 2.58 were significant at P=.01.

  • The standard YX is an approximation of the strength of the relative contribution of the background variable to the outcome (either the latent construct or the path outcome) obtained by using data that adjust for the differences in measurement scales.

 Physical activity
  Baseline exercise frequency−0.2270.053−4.327−0.413
  Provider knowledge0.4030.1343.0060.280
 Age at diagnosis−0.0170.005−3.058−0.228
  Cancer pain−0.1260.029−4.349−0.224
  Cancer discussion−0.7170.167−4.286−0.303
  Cancer pain0.2950.0674.4180.256
  Intrinsic motivation−0.0470.016−2.967−0.186
  Extrinsic motivation0.0760.0184.1260.278
  Cancer anxiety−0.3730.055−6.828−0.298
  Cancer anxiety−0.6250.094−6.666−0.447
  Cancer pain−0.3770.137−2.747−0.263
  Physical activity−0.3710.091−4.087−0.439
 Cancer anxiety
  Cancer pain0.2700.0505.3860.264
  Intrinsic motivation−0.0400.015−2.714−0.179
 Physical activity
  Baseline exercise frequency−0.1220.038−3.220−0.211
  Cancer pain−0.2450.036−6.755−0.342
  Intrinsic motivation0.0270.0073.8350.173
  Baseline exercise frequency0.0440.0152.8530.161
  Extrinsic motivation0.0870.0155.9210.343
  Seen PCP since cancer RX0.4080.1043.9180.243
 Patient-provider interaction
  Extrinsic motivation0.0650.0213.1300.206
Intrinsic motivation    
 Survivor-provider interaction0.6390.2023.1570.177
 Cancer anxiety
  Cancer pain0.3460.0447.8870.305
  Cancer pain−0.3950.091−4.333−0.255

The Physical Activity Model for Women

A greater likelihood that women survivors had recently participated in physical activity (Table 3, Fig. 3) was predicted directly by greater reported stamina, less fatigue, and more frequent baseline aerobic exercise. Higher levels of intrinsic motivation (P = .01) were a significant indirect predictor of physical activity through the effect on stamina. Greater stamina was predicted by less fatigue, less cancer-related pain, more frequent exercise at baseline, and greater intrinsic motivation. Greater fear regarding health was associated with having seen a PCP since completion of cancer therapy, greater fatigue, and greater extrinsic motivation. A more positive perception of survivor-PCP interaction was predicted by less fear regarding future health and greater extrinsic motivation. Greater intrinsic motivation was predicted by more positive perceptions of interaction with the provider. Greater fatigue was predicted by greater cancer-related pain, and greater anxiety was predicted by greater cancer-related pain and greater fear concerning future health.


  1. Top of page
  2. Abstract
  6. Acknowledgements
  7. Conflict of Interest Disclosures
  8. References

Although, in this study, we identified some of the same direct predictors (education, fatigue, fear, stamina) of physical activity participation that were reported for adult cancer survivors,35, 36 the impact of these variables were sex-specific, and they simultaneously and indirectly predicted physical activity participation through their impact on other variables. A higher level of education37 predicted physical activity among men. Perceptions of greater stamina predicted physical activity participation among women survivors, consistent with the hypothesis38 that low exercise capacity accounts directly for a more sedentary lifestyle among survivors. As with survivors of adult cancer,39, 40 fatigue predicted less physical activity among women; conversely, physical activity participation predicted less fatigue in men.

Both men and women who reported more frequent exercise at baseline were more likely to report recent physical activity, suggesting that physical activity was a previously established behavior. For both sexes, cancer-related pain, anxiety, fear, and fatigue were powerful, indirect influences on physical activity participation. Pain contributed to increased fatigue and anxiety and reduced stamina. Perceptions of decreased stamina reinforced perceptions of increased fatigue. Higher fear levels increased cancer-related anxiety.

In contrast to anxiety, which is a negative influence on positive health behavior, fear and worry can support positive health behavior.41 Consistent with survivors of adult cancer,40 greater fear concerning health predicted physical activity participation among men childhood cancer survivors.

Having discussed cancer recurrence with their PCP predicted men's greater fear. Men often view physician visits as necessary only in a crisis30, 42, 43; therefore, physician contact may reflect specific problems that contribute to a more negative affect. In contrast, women tend to view interaction with their physician as supportive.44, 45

Perceptions of survivor-provider interaction were relevant to recent physical activity in the model for women but not in the model for men: The provider's perceived competence was predictive for men but not for women. These findings further demonstrate significant sex differences in healthcare expectations in which men seek minimal physician input but value physician competence, whereas women rely on physician input for their healthcare decisions and value the quality of the relationship with the physician.30, 42-45

Both models revealed complex relations between fear, motivation, and affect. Men who were more motivated extrinsically had greater fear about their future health. It has been established that extrinsically motivated individuals are more worried and fearful about their health, see themselves as less in control of health matters, and rely more on health professionals for direction.18, 19, 46 Higher levels of intrinsic motivation, supported by positive perceptions of the survivor-provider relationship, predicted greater stamina in women and indirectly predicted physical activity participation through stamina. Individuals with greater intrinsic motivation more often are self-reliant and self-directed, rather than provider-directed,19, 21 in their health behavior choices.

Clinical Implications

Regardless of the length of time since a survivor's cancer diagnosis and treatment,29, 47 PCPs are encouraged to inquire specifically about any treatment-related symptoms, particularly pain, fatigue, and anxiety. These symptoms may share common biologic mechanisms48, 49 and, until they are addressed, may remain significant barriers to physical activity in survivors. Physical activity programs have reduced fatigue and decreased depression and anxiety in survivors of adult cancer,6, 7, 50-52 and they may have a similar effect on childhood cancer survivors.

Pediatric cancer patients experience problems with pain, physical function, and fatigue throughout therapy53-55; as survivors, many must confront these persistent symptoms as well as the late effects of radiation and chemotherapy.56, 57 To modify these late effects, it is important to introduce individually tailored physical activity during treatment and support survivors to maintain their physical activity throughout their lives. Exercise programs during treatment include passive range-of-motion exercises, progressing to strength and endurance activities.58-62 The concurrent demands of treatment, together with the patient's physical, psychological, and motivational preparedness, must be considered in tailoring physical activities.63

As in the general population, sex should be considered in tailoring physical activity interventions to survivors of childhood cancer.64-74 For some men survivors, personalized risk information conveyed by print, video, and web-based media71, 74 may help to increase health concern and support motivation for physical activity in the absence of frequent physician interaction.69-71, 74 In some women survivors, direct contact with the PCP may be more effective in supporting physical activity than a media intervention. Finally, interventions that are tailored individually to the influence of fear, affect, and motivation65, 67, 69 are more likely to be effective in supporting changes in behavior than interventions that only address changing knowledge.

In conclusion, the study sample reflects a subset of the overall CCSS population—those who responded to the Health Care Needs and CCSS Follow-up 2 Surveys; therefore, survivors who were included in the current analysis may not be fully representative of the population from which they were derived. The physical activity outcome and the independent measures were based on self-reported data.

Despite the study limitations, we are able to report 3 novel findings regarding factors that predict the likelihood of childhood cancer survivors' physical activity and suggestions for considering these factors in managing survivors' long-term care: 1) physical activity is influenced by pain, anxiety, fatigue, and stamina independent of sex; 2) the sexes differ in which factors predict physical activity and how they do so; and 3) affect, fear, motivation, and survivor-physician interaction are modifiable predictors of the likelihood of physical activity. Interventions that consider these multiple factors have the potential to support positive behavior change in childhood cancer survivors.


  1. Top of page
  2. Abstract
  6. Acknowledgements
  7. Conflict of Interest Disclosures
  8. References

We acknowledge the contributions of Sharon Naron, MPA, ELS and Vani Shanker, PhD (for editorial assistance) and Kelly Shempert and Dawn Silcott (for illustrations).

Conflict of Interest Disclosures

  1. Top of page
  2. Abstract
  6. Acknowledgements
  7. Conflict of Interest Disclosures
  8. References

Supported by grants from the National Institutes of Health (NIH) (NINR RO3 NR009203; (Cheryl L. Cox, principal investigator); the Robert Wood Johnson Foundation (Kevin C. Oeffinger, principal investigator); the NIH- National Cancer Institute (NCI U24 CA55727; (Leslie L. Robison, principal investigator); and the American Lebanese Syrian Associated Charities (ALSAC).


  1. Top of page
  2. Abstract
  6. Acknowledgements
  7. Conflict of Interest Disclosures
  8. References
  • 1
    Harila-Saari AH, Vainionpaa LK, Kovala TT, Tolonen EU, Lanning BM. Nerve lesions after therapy for childhood acute lymphoblastic leukemia. Cancer. 1998; 82: 200-207.
  • 2
    Mattano LA, Sather HN, Trigg ME, Nachman JM. Osteonecrosis as a complication of treating acute lymphoblastic leukemia in children: a report from the Children's Cancer Group. J Clin Oncol. 2000; 18: 3262-3272.
  • 3
    Reilly J, Bentham JC, Ralston JM, et al. Reduced energy expenditure in preobese children treated for acute lymphoblastic leukemia. Pediatr Res. 1998; 44: 557-562.
  • 4
    Reinders-Messelink H, Schoemaker M, Snijders T, et al. Motor performance of children during treatment for acute lymphoblastic leukemia. Med Pediatr Oncol. 1999; 33: 545-550.
  • 5
    Centers of Disease Control and Prevention, National Centers for Chronic Disease Prevention and Health Promotion. Physical Activity and Health: A Report of the Surgeon General. Hyattsville, Md: US Department of Health and Human Services; 1996.
  • 6
    Courneya KS. Exercise interventions during cancer treatment: biopsychosocial outcomes [review]. Exerc Sport Sci Rev. 2001; 29: 60-64.
  • 7
    Dimeo FC, Stieglitz RD, Novelli-Fischer U, Fetscher S, Keul J. Effects of physical activity on the fatigue and psychologic status of cancer patients during chemotherapy. Cancer. 1999; 85: 2273-2277.
  • 8
    Young-McCaughan S, Mays MZ, Arzola SM, et al. Research and commentary: change in exercise tolerance, activity and sleep patterns, and quality of life in patients with cancer participating in a structured exercise program. Oncol Nurs Forum. 2003; 30: 441-454.
  • 9
    Butterfield RM, Park ER, Puleo E, et al. Multiple risk behaviors among smokers in the Childhood Cancer Survivors Study cohort. Psychooncology. 2004; 13: 619-629.
  • 10
    Demark-Wahnefried W, Werner C, Clipp EC, et al. Survivors of childhood cancer and their guardians. Cancer. 2005; 103: 2171-2180.
  • 11
    Pinto BM, Trunzo JJ. Health behaviors during and after a cancer diagnosis. Cancer. 2005; 104( 11 suppl): 2614-2623.
  • 12
    Tercyak KP, Donze JR, Prahlad S, Mosher RB, Shad AT. Multiple behavioral risk factors among adolescent survivors of childhood cancer in the Survivor Health and Resilience Education (SHARE) Program. Pediatr Blood Cancer. 2006; 47: 825-830.
  • 13
    Centers for Disease Control and Prevention (CDC). Prevalence of physical activity, including lifestyle activities among adults—United States, 2000-2001. MMRW Morb Mortal Wkly Rep. 2003; 52: 764-769.
  • 14
    US Department of Health and Human Services. Leisure-Time Physical Activity Among Adults: United States, 1997-1998. Hyattsville, Md: US Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Health Statistics; 2002.
  • 15
    Florin TA, Fryer GE, Miyoshi T, et al. Physical inactivity in adult survivors of childhood acute lymphoblastic leukemia: a report from the Childhood Cancer Survivor Study. Cancer Epidemiol Biomarkers Prev. 2007; 16: 1356-1363.
  • 16
    Carter K, Kulbok P. Evaluation of the interaction model of client health behavior through the first decade of research. Adv Nurs Sci. 1995; 18: 62-73.
  • 17
    Cox CL. An interaction model of client health behavior: theoretical prescription for nursing. Adv Nurs Sci. 1982; 5: 41-56.
  • 18
    Cox CL. Online exclusive: a model of health behavior to guide studies of childhood cancer survivors. Oncol Nurs Forum. 2003; 30: E92-E99.
  • 19
    Cox CL, McLaughlin RA, Rai SN, Steen BD, Hudson MM. Adolescent survivors: a secondary analysis of a clinical trial targeting behavior change. Pediatr Blood Cancer. 2005; 45: 144-154.
  • 20
    Cox CL, McLaughlin RA, Steen BD, Hudson MM. Predicting and modifying substance use in childhood cancer survivors: application of a conceptual model. Oncol Nurs Forum. 2006; 33: 51-60.
  • 21
    Cox CL, Montgomery M, Rai SN, McLaughlin R, Steen BD, Hudson MM. Supporting breast self-examination in female childhood cancer survivors: a secondary analysis of a behavioral intervention. Oncol Nurs Forum. 2008; 35: 423-430.
  • 22
    Robison LL, Mertens AC, Boice JD, et al. Study design and cohort characteristics of the Childhood Cancer Survivor Study: a multi-institutional collaborative project. Med Pediatr Oncol. 2002; 38: 229-239.
  • 23
    Centers for Disease Control and Prevention National Center for Chronic Disease Prevention and Health Promotion. Behavioral Risk Factor Surveillance System, 2008. Available at: Accessed May 15, 2008.
  • 24
    Wallston KA, Wallston BS, DeVellis R. Development of the Multidimensional Health Locus of Control (MHLC) Scales. Health Educ Monogr. 1978; 6: 160-170.
  • 25
    Muthen K, Muthen BO. Mplus User's Guide. 4th ed. Los Angeles, Calif: Muthen & Muthen; 2007.
  • 26
    Ware JE, Snow KK, Kosinski M. SF-36 Health Survey: Manual and Interpretation Guide. Lincoln, RI: Quality Metric Incorporated; 2000.
  • 27
    Kalton G, Kasprzyk D. The treatment of missing survey data. Surv Methodol. 1986; 12: 1-16.
  • 28
    Kline RB. Principles and Practice of Structural Equation Modeling (Methodology in the Social Sciences). 2nd ed. New York: The Guilford Press; 2005.
  • 29
    Hudson MM, Mertens AC, Yasui Y, et al. Health status of adult long-term survivors of childhood cancer: a report from the Childhood Cancer Survivor Study. JAMA. 2003; 290: 1583-1592.
  • 30
    Oeffinger KC, Mertens AC, Hudson MM, et al. Healthcare of young adult survivors of childhood cancer: a report from the Childhood Cancer Survivor Study. Ann Fam Med. 2004; 2: 61-70.
  • 31
    Jago R, Anderson CB, Baranowski T, Watson K. Adolescent patterns of physical activity differences by gender, day, and time of day. Am J Prev Med. 2005; 28: 447-452.
  • 32
    Sallis JF. Epidemiology of physical activity and fitness in children and adolescents. Crit Rev Food Sci Nutr. 1993; 33: 403-408.
  • 33
    Thompson AM, Baxter-Jones AD, Mirwald RL, Bailey DA. Comparison of physical activity in male and female children: does maturation matter? Med Sci Sports Exerc. 2003; 35: 1684-1690.
  • 34
    Trost SG, Pate RR, Sallis JF, et al. Age and gender differences in objectively measured physical activity in youth. Med Sci Sports Exerc. 2002; 34: 350-355.
  • 35
    Courneya KS, Vallance JKH, Jones LW, Reiman T. Correlates of exercise intentions in non-Hodgkin's lymphoma survivors: an application of the Theory of Planned Behavior. J Sport Exerc Pathol. 2005; 27: 335-349.
  • 36
    Courneya KS, Friedenreich CM, Reid RD, et al. Predictors of follow-up exercise behavior 6 months after a randomized trial of exercise training during breast cancer chemotherapy. Breast Cancer Res Treat. 2008 [Epub ahead of print].
  • 37
    Patterson RE, Neuhouser ML, Hedderson MM, Schwartz SM, Standish LJ, Bowen DJ. Changes in diet, physical activity, and supplement use among adults diagnosed with cancer. J Am Diet Assoc. 2003; 103: 323-328.
  • 38
    Aziz NM. Cancer survivorship research: challenge and opportunity. J Nutr. 2002; 132: 3494S-3503S.
  • 39
    Courneya KS, Friedenreich CM, Quinney HA, et al. A longitudinal study of exercise barriers in colorectal cancer survivors participating in a randomized controlled trial. Ann Behav Med. 2005; 29: 147-153.
  • 40
    Stolley MR, Sharp LK, Wells AM, Simon N, Schiffer L. Health behaviors and breast cancer: experiences of urban African-American women. Health Educ Behav. 2006; 33: 604-624.
  • 41
    Sallis JF, Prochaska JJ, Tayler WC. A review of correlates of physical activity of children and adolescents. Med Sci Sports Exerc. 2000; 32: 963-975.
  • 42
    Shaw AK, Pogany L, Speechley KN, Maunsell E, Barrera M, Mery LS. Use of healthcare services by survivors of childhood and adolescent cancer in Canada. Cancer. 2006; 106: 1829-1837.
  • 43
    Xu KT, Borders TF. Gender, health, and physician visits among adults in the United States. Am J Public Health. 2003; 93: 1076-1079.
  • 44
    Hall JA, Irish JT, Roter DL, Ehrlich CM, Miller LH. Satisfaction, gender, and communication in medical visits. Med Care. 1994; 32: 1216-1231.
  • 45
    Hall JA, Roter DL. Patient gender and communication with physicians: results of a community-based study. Womens Health. 1995; 1: 77-95.
  • 46
    Deci EL, Ryan RM. Handbook of Self-Determination Research. Rochester, NY: University of Rochester Press; 2002.
  • 47
    Meeske K, Siegel S, Globe D, Mack W, Bernstein L. Prevalence and correlates of fatigue in long-term survivors of childhood leukemia. J Clin Oncol. 2005; 23: 5501-5510.
  • 48
    Cleeland CS, Bennett GJ, Dantzer R, et al. Are the symptoms of cancer and cancer treatment due to a shared biologic mechanism? Cancer. 2003; 97: 2919-2925.
  • 49
    Lee BN, Dantzer R, Langley KE, et al. A cytokine-based neuroimmunologic mechanism of cancer-related symptoms. Neuroimmunomodulation. 2004; 11: 279-292.
  • 50
    Courneya KS, Friedenreich CM. Physical exercise and quality of life following cancer diagnosis: a literature review. Ann Behav Med. 1999; 21: 171-179.
  • 51
    Courneya KS, Mackey JR, Bell GJ, Jones LW, Field CJ, Fairey AS. Randomized controlled trial of exercise training in postmenopausal breast cancer survivors: cardiopulmonary and quality of life outcomes. J Clin Oncol. 2003; 21: 1660-1668.
  • 52
    Mock V, Frangakis C, Davidson NE, et al. Exercise manages fatigue during breast cancer treatment: a randomized controlled trial. Psychooncology. 2005; 14: 464-477.
  • 53
    Cox C, Rai SN, Rosenthal D, Phipps S, Hudson MM. Subclinical late cardiac toxicity in childhood cancer survivors: impact on quality of life. Cancer. 2008; 112: 1835-1844.
  • 54
    Karimova EJ, Rai SN, Howard SC, et al. Femoralhead osteonecrosis in pediatric and young adult patients with leukemia or lymphoma. J Clin Oncol. 2007; 25: 1525-1531.
  • 55
    Vainionpaa L, Kovala T, Tolonen U, Lanning M. Vincristine therapy for children with acute lymphoblastic leukemia impairs conduction in the entire peripheral nerve. Pediatr Neurol. 1995; 13: 314-318.
  • 56
    Kaste SC, Rai SN, Fleming K, et al. Changes in bone mineral density in survivors of childhood acute lymphoblastic leukemia. Pediatr Blood Cancer. 2006; 46: 77-87.
  • 57
    Oeffinger K, Eshelman D, Tomlinson G, et al. Grading of late effects in young adult survivors of childhood cancer followed in an ambulatory adult setting. Cancer. 2000; 88: 1687-1695.
  • 58
    Aznar S, Webster AL, San Juan AF, et al. Physical activity during treatment in children with leukemia: a pilot study. Appl Physiol Nutr Metab. 2006; 31: 407-413.
  • 59
    Hinds PS, Hockenberry M, Rai SN, et al. Clinical field testing of an enhanced-activity intervention in hospitalized children with cancer. J Pain Symptom Manage. 2007; 33: 686-697.
  • 60
    Marchese VG, Chiarello LA, Lange BJ. Effects of physical therapy intervention for children with acute lymphoblastic leukemia. Pediatr Blood Cancer. 2004; 42: 127-133.
  • 61
    San Juan AF, Fleck SJ, Chamorro-Vina C, et al. Effects of an intrahospital exercise program intervention for children with leukemia. Med Sci Sports Exerc. 2007; 39: 13-21.
  • 62
    San Juan AF, Fleck SJ, Chamorro-Vina C, et al. Early-phase adaptations to intrahospital training in strength and functional mobility of children with leukemia. J Strength Condition Res. 2007; 21: 173-177.
  • 63
    Stull VB, Snyder DC, Demark-Wahnefried W. Lifestyle interventions in cancer survivors: designing programs that meet the needs of this vulnerable and growing population. J Nutr. 2007; 137: 243S-248S.
  • 64
    de Nooijer J, Lechner L, Candel M, de Vries H. Short- and long-term effects of tailored information versus general information on determinants and intentions related to early detection of cancer. Prev Med. 2004; 38: 694-703.
  • 65
    Dijker AJ, Koomen W, Kok G. Interpersonal determinants of fear of people with AIDS: the moderating role of predictable behavior. Basic Appl Soc Psychol. 1997; 19: 61-79.
  • 66
    Eakin EG, Glasgow RE, Riley KM. Review of primary care-based physical activity intervention studies: effectiveness and implications for practice and future research. J Fam Pract. 2000; 49: 158-168.
  • 67
    Geller ES. Designing an Effective Fear Appeal. Available at: Accessed August 20, 2007.
  • 68
    Glasgow RE, Goldstein MG, Ockene JK, Pronk NP. Translating what we have learned into practice. Principles and hypotheses for interventions addressing multiple behaviors in primary care. Am J Prev Med. 2004; 27( 2 suppl): 88-101.
  • 69
    Hoskins SL, van Hooff JC. Motivation and ability: which students use online learning and what influence does it have on their achievement? Br J Educ Technol. 2005; 36: 177-192.
  • 70
    Lee MH, Tsai CC. Exploring high school students' and teachers' preferences toward the constructivist Internet-based learning environments in Taiwan. Educ Stud. 2005; 31: 149-167.
  • 71
    Passig D, Levin H. Gender preferences for multimedia interfaces. J Comp Assist Learning. 2000; 16: 64-71.
  • 72
    Revere D, Dunbar PJ. Review of computer-generated outpatient health behavior interventions: clinical encounters “in absentia.” J Am Med Inform Assoc. 2001; 8: 62-79.
  • 73
    Skinner CS, Strecher VJ, Hospers H. Physicians' recommendations for mammography: do tailored messages make a difference? Am J Public Health. 1994; 84: 43-49.
  • 74
    Young BJ. Gender differences in student attitudes toward computers. J Res Comput Educ. 2000; 33: 204-216.
  • 75
    Bentler PM. Comparative fit indexes in structural models. Psychol Bull. 1990; 107: 238-246.
  • 76
    Bollen K. Overall fit in covariance structure models: 2 types of sample size effects. Psychol Bull. 1990; 107: 256-259.
  • 77
    Browne M, Cudeck R. Alternative ways of assessing model fit. In: BollenKA, LongJA, eds. Testing Structural Equation Models. Newbury Park, Calif: Sage; 1993: 136-162.
  • 78
    Hu L, Bentler P. Cutoff criteria for fit indices in covariance structure analysis: conventional criteria versus new alternatives. Struct Equation Model. 1999; 6: 1-55.