Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Psychooncology. Author manuscript; available in PMC 2013 May 1.
Published in final edited form as:
Psychooncology. 2012 May; 21(5): 469–478.
Published online 2011 March 6. doi:  10.1002/pon.1935
PMCID: PMC3563242

Survivor Profiles Predict Health Behavior Intent: The Childhood Cancer Survivor Study



To determine whether unique groups of adult childhood cancer survivors could be defined on the basis of modifiable cognitive, affective and motivation indicators. Secondary objectives were to examine to what extent group membership covaried with more static variables (e.g., demographics, disease, and treatment) and predicted intent for subsequent medical follow-up.


Using latent class analysis of data from 978 participants (ages 18–52 years; mean, 31; SD, 8) in the Childhood Cancer Survivor Study, we classified survivors according to their worries about health, perceived need for follow-up care, health motivation, and background variables. Intent to participate in medical follow-up, as a function of class membership, was tested using equality of proportions.


The best-fitting model (BIC=18,540.67, BLMRT=<0.001) was characterized by three distinctive survivor classes (worried, 19%; self-controlling, 26%; collaborative, 55%) and three significant class covariates (gender, perceptions of health and severity of late effects). A smaller proportion of survivors in the self-controlling group [81%] than in the worried [90%] (P=0.015) and collaborative [88%] (P=0.015) groups intended to obtain a routine medical checkup. A smaller proportion of survivors in the self-controlling group [32%] than in the collaborative [65%] (P=<0.001) and worried [86%] (P=<0.001) groups planned a cancer-related check-up. A smaller proportion of survivors in the collaborative group [65%] than in the worried group [86%] (P=<0.001) were likely to obtain a cancer-related check-up.


Childhood cancer survivors can be classified according to modifiable indicators. The classification is distinctive, predicts intent for future medical follow-up, and can inform tailored interventions.

Keywords: childhood cancer, survivorship, late effects, medical follow-up, pediatric oncology


As childhood cancer survival rates approach 80%–90%,[1, 2] interventions to modify late treatment-related morbidities[36] become increasingly important. Despite clear treatment exposure-based follow-up recommendations,[7] survivors' participation in recommended cancer-related follow-up is sub-optimal.[814]

There is growing support for intervention strategies that target what people think, rather than what they do, to promote health-enhancing behaviors.[15] Motivation, perceived self worth, social/behavioral contexts, and worries/concerns are observable variables that may influence health outcomes as mediators (affecting why and how interventions work) and as moderators (affecting the impact of interventions on different people and under different circumstances).[1517] Not only can these variables be experimentally manipulated[17] and targeted for intervention, but they also may be useful in defining groups that respond better to one model of survivorship care versus another. We therefore used latent class analysis (LCA) to: 1) characterize survivor groups on the basis of potentially modifiable indicators; 2) identify background variables likely to predict group membership; and 3) determine whether group designation predicts intent to participate in medical follow-up.


Late effects are cancer-related sequelae that persist or develop five or more years after completion of cancer therapy. They encompass numerous conditions, including organ dysfunction, psychosocial problems, and subsequent malignancies, that reduce survivors' quality of life and predispose them to early mortality.[18] Investigators from the Childhood Cancer Survivor Study (CCSS) reported a 73% cumulative incidence of chronic health conditions and a 42% cumulative incidence of severe, disabling, or life-threatening conditions or death from chronic illness among survivors 30 years after cancer diagnosis.[18] Childhood cancer survivors' long-term health outcomes are optimal when they receive ongoing comprehensive follow-up care that includes monitoring for and treating the late effects of their cancer therapies.[19] The Children's Oncology Group (COG) has compiled risk-based, exposure-related clinical practice guidelines for screening and management of late effects of treatment for pediatric malignancies.[7] However, CCSS data indicate that most survivors are not engaging in the recommended follow-up care.[8, 1112]

A variety of demographic, medical, and logistic factors are barriers to adequate follow-up care for survivors of childhood cancer. At the time when most children and adolescents are considered cured of their cancer, they generally have no overt sequelae of chemotherapy, surgery, or radiation.[20] Most survivors are gradually lost to follow-up at their treating institutions.[2022] Further, most adult survivors of childhood cancer do not have a summary of their cancer and therapy, can describe their treatment only in broad terms, and are unaware of their risk of sequelae.[20, 23]

The risk and the severity of late effects are potentially modifiable by interventions that encourage survivor participation in specialized surveillance, screening, and risk management.[24] However, only 19% of 9434 adult survivors reported having had a cancer center evaluation within the preceding 2 years; 42% reported having had a cancer-related medical evaluation by a community or cancer center physician during that interval.[21] A recent study[11] documented that while 88% of 8,522 childhood cancer survivors had received some form of medical care during the preceding 2 years, only 32% had received survivor-focused care and only 18% reported survivor-focused care that included advice about screening for and reduction of late-effects risks. Many survivors who do not receive follow-up care report that their primary physician sees no need for specialized follow-up; others perceive themselves to be in good health and not in need of cancer-related care.[25]

To integrate the multiple factors that influence survivor behavior, we selected the Interaction Model of Client Health Behavior (IMCHB)[2630] (Figure 1). This model recognizes background variables, including demographics, environmental resources (e.g., income, health care access), social influences (e.g., social support, subjective norms), and physical limitations (e.g., pain, functional limitations), as influential moderators or direct predictors of health behavior outcomes. Dynamic variables, such as cognitive appraisal (e.g., health beliefs, knowledge), motivation, and affective response (e.g., fear, anxiety) are identified as important covariates, direct predictors, or mediators of these outcomes. Briefly, the model comprises three elements: client singularity (the unique intrapersonal and contextual configuration of the individual), client-professional interaction, and health outcomes. The model's working hypothesis is that the potential for positive health outcomes increases as the intervention is tailored to the uniqueness (singularity) of each individual (i.e., background, affective, cognitive, and motivation characteristics). The background variables are considered to be more static and to potentially moderate the influence of the more modifiable dynamic mediating variables – affective response, cognitive appraisal, and motivation. The model has been used successfully to identify predictors, moderators, and mediators of breast self-examination[31] physical activity,[3233] mammography,[34] bone densitometry, and echocardiography screening[35] in childhood cancer survivors.

Figure 1
The Interaction Model of Client Health Behavior

We used the modifiable constructs within the IMCHB to generate a latent class (LC) model of childhood cancer survivors. The primary objective was to determine whether unique groups of childhood cancer survivors could be defined on the basis of their members' configuration on known robust mediators (cognitive appraisal, affective response, motivation) of health behavior. Secondary objectives were to examine to what extent group membership covaried with more static moderating variables (e.g., demographics, disease, and treatment) and predicted subsequent medical follow-up outcomes.


Data Source

The CCSS is a multi-institutional retrospective cohort study initiated in 1994 to examine late effects in survivors of pediatric cancers diagnosed and treated between 1970 and 1986. Survivors completed a baseline questionnaire at study entry and respond to follow-up questionnaires at regular intervals. Participants consented to release their medical records. Questionnaires and sampling methods were detailed by Robison et al.[36] and are available for review at The CCSS has IRB approval from each of the 27 participating institutions.


The 20,346 survivors initially invited to participate in CCSS were those who had survived 5 or more years after treatment for a malignant disease diagnosed (before age 21 years). In an ancillary study, the Health Care Needs Survey (HCNS), 1600 of the survivors were randomly sampled; 978 (61%) completed and returned the survey. Non-respondents to the HCNS were predominantly male (59%), in a racial/ethnic minority group (37%), or did not complete high school (56%). Self-reported data from everyone who responded to the HCNS and Follow-up 2002–2003 surveys were used in the present study. All survivors in this study were age ≥ 18 years at the time of data collection.


Class indicators (fears, health concerns, intrinsic and extrinsic motivation): All of the items comprising the measures of interest within the survey data were subjected to both exploratory and confirmatory factor analyses. Only those items that were significant in these preliminary analyses were used in the latent class analysis. Affective response: Three summed items defined survivors' worries/fears about: future health, recurrence of cancer, and discovery of a problem at a check-up visit (1=not at all; 5=extremely). Cognitive appraisal: Three items were summed to define survivors' concerns about : general health, the risk of getting sick or developing problems related to previous cancer, the importance of a check-up (1=not at all; 5=extremely). Intrinsic motivation: Five summed items from the Multidimensional Health Locus of Control Scale[37] (Abbreviated as: In control of my health; My behavior determines my health; What I myself do is most important; I can avoid illness by self-care; ; My right actions keep me healthy) (1=moderately/strongly disagree; 4=moderately/strongly agree). Extrinsic motivation: Four summed items from the Multidimensional Health Locus of Control Scale[37] (Abbreviated as: Providers control my health; Must have regular doctor visits to stay well; If unwell, should consult a provider; I can only do what doctor tells me) (1=moderately/strongly disagree; 4=moderately/strongly agree).


Demographic, disease, and treatment variables included: age at follow-up, age at cancer diagnosis, years since cancer diagnosis, gender, education, race, personal income, and current health insurance status. Perceived severity of late effects: Survivors indicated whether they had experienced chronic health problems lasting longer than six months and rated the severity of their main chronic health problem (1=mild – no medications, no effect on daily life; 4=life threatening). These responses were dichotomized to create a “late effects” variable (moderate, severe or life-threatening vs. mild or no chronic problems). Health status: Survivors rated their current health with a single-item measure as 1=Excellent to 5=Poor.

Medical Follow-up

Routine medical check-up: Survivors indicated their likelihood of having a routine check-up during the next 2 years (0=very unlikely or unlikely; 1=possible, likely, or very likely). Cancer-related check-up: Survivors indicated their likelihood of having a medical exam within the next 2 years to check for health problems caused by cancer treatment (0=very unlikely or unlikely; 1=possible, likely, or very likely).

Statistical Analysis

We used SAS 9.1 software (SAS Institute Inc, Cary, NC) to describe sample characteristics and MPlus Version 6.0 software (Muthén & Muthén, Los Angeles, CA) to develop the latent class models. We used survivors' responses to continuous indicators that represented the mediators from the theoretical model (affective response, cognitive appraisal, intrinsic and extrinsic motivation) to estimate latent class membership for each participant.[38]

As recommended by Nylund and colleagues[39] we used a combination of the Bayesian information criterion (BIC) and bootstrapped Lo-Mendell-Rubin parametric likelihood ratio test (BLMRT) to determine the number of latent classes. During the model-fitting process, classes are added to the model as long as the BIC continues to decrease and the added classes are substantively meaningful. The BIC balances model fit with model parsimony such that lower values represent better fit. The BLMRT compares progressive iterations of the more parsimonious models (k-1 classes) against models with a greater number of classes (k classes). Models that have the fewest distinctive classes that are substantively meaningful and a BLMRT P= ≤0.05 are accepted.

As part of the model-fitting process, covariates were added to further characterize the latent classes. Significant covariates with the latent classes (P < .05) were included in the best-fitting latent class model. Finally, binary outcomes (routine medical or cancer-related follow-up) were added to the model to determine class-specific likelihood of participation in follow-up care.


Descriptive Summary of the Sample

Respondents ranged in age from 18 to 52 years (median, 30 years) (Table 1). Slightly more than half (53%) were female and 73% were white. Less than half (43%) were college-educated, and 67% reported personal incomes less than $40,000 per year.

Table 1
Background variables by group classification (N=978)

Latent Class Models

We first identified survivor classes on the basis of our indicator variables without including covariates (Table 2). The 4-class and 5-class models had lower BIC values than the 3-class model, but the BLMRT for the 4- and 5- class models was not significant. The 3-class model was superior to the 2-class model on the basis of BIC (18,914.79) and BLMRT (≤0.001).

Table 2
Fit statistics of the latent class models

To ensure that the 3-class model's superior fit was not based on a small set of random start values, we repeated the analyses with 100 random start values and 10 optimizations for each of the 100 sets of start values; all 100 random start values resulted in the same solution, confirming that the 3-class model offered the best fit.

Our final step was to assess the substantive meaning of the 3-class model. Compared to the other classes, members of the smallest class (19%) reported poorer perceived health status, a higher percentage of moderate to life-threatening chronic illnesses, the greatest worries/fears and health concerns, and the lowest levels of intrinsic motivation in managing their health. Given the young age of the total sample, we would expect the group with the most health problems to be the smallest class. In the next largest class (26%), all class indicators were markedly opposite to those of the smallest class, and health was reported as good to excellent. Having two groups at these extremes on the class indicators and self-reported health variables would logically predict a more moderate class; indeed, the largest class (55%) showed intermediate scores on the class indicators and on self-reported health status. When covariates were included, a model that included sex, perceived health status, and perceived severity of late effects further reduced the 3-class BIC to 18,540.67 (BLMRT P=<0.001). For the final model with covariates, the average posterior probabilities (showing how well each participant fit the assigned class) ranged from 86% to 88%, indicating a high degree of confidence in class assignment and model fit.

Affective, Cognitive, and Motivation Profiles of the Three Classes

Figure 2 and Table 3 show the class-specific means for the indicator variables (affective, cognitive and intrinsic/extrinsic motivation scales). We labeled the classes on the basis of survivors' responses to the class indicators (dynamic mediating variables): 1) Worried survivors (19% of the total) were the most worried about their health and endorsed the importance of routine check-ups most strongly. They were the least intrinsically and most extrinsically motivated to manage their health care. This group was predominantly female with the greatest percentage of black and Hispanic survivors, the fewest college graduates, and the greatest percentage of lower-income survivors and survivors with one or more chronic health problems (Table 1). 2) Self-controlling survivors (26%) were not greatly concerned about their cancer history and risk of sequelae, and they saw little value in medical check-ups. They were the least likely of the groups to report having read about childhood cancer or related topics. The self-controlling group reported the least worry or anxiety about their cancer and treatment and showed the least extrinsic motivation about their health. The group was predominantly male, with the lowest percentages of black and Hispanic survivors and the highest percentages of survivors who had completed college and who reported higher incomes. Most members of this group reported no late effects (84%); they also reported the least limitation of physical activity, and 95% reported their health as good to excellent. 3) Collaborative survivors (55%) were the most moderate in terms of their worries/fears and concerns about their health. While members of this group were highly intrinsically motivated to manage their health care, they showed significantly greater extrinsic motivation than the self-controlling group. Table 3 lists the estimated conditional probability of the groups' mean scores on the indicator scales. The estimated conditional probability divided by the standard error (EST/SE) indicates the strength of the relationship between the indicator and the latent class variable.

Figure 2
Estimated mean scores on the indicators by latent class
Table 3
Latent class indicators by group (N=978)

Survivor Participation in Medical Follow-up Care

We next compared the self-reported likelihood of a future routine or cancer-related medical check-up in the three groups. Controlling for the significant latent class covariates (sex, perceived health status, and perceived severity of late effects), we determined how well the model distinguished survivors who were likely to participate in follow-up care by testing the equality of the proportions of survivors likely to participate in routine or cancer-related medical follow-up across classes (Table 4). The greatest proportions of those likely to obtain a routine medical checkup or cancer-related check-up within the next 2 years were observed in the worried and collaborative groups, followed by the self-controlling group. A smaller proportion of the collaborative group than of the worried group was likely to obtain a cancer-related check-up. A similar proportion of survivors (90% vs. 88%) in the worried and collaborative groups were likely to obtain a routine medical check-up within the next 2 years.

Table 4
Comparison of the proportion of survivors planning to obtain a routine or cancer-related check-up by latent class


Three distinct groups of childhood cancer survivors (worried, self-controlling, and collaborative) were identified through latent class analysis on the basis of survivors' self-reported feelings and perceptions about their cancer and its treatment, their future health, and their motivation related to health. Sex, perceived health status, and the reported presence and severity of late effects were significant covariates of the classes, and class membership predicted the likelihood of participating in the recommended routine and/or cancer-related medical care. Latent class analysis supported the basic premises of the conceptual model[26]. Survivors could be profiled by the model's modifiable mediators; these heterogeneous profiles together with the model's background variables (moderators) predicted self-reported behaviorally-related outcomes.

Survivors in the self-controlling group perceived themselves to be in good health and as a result, they likely see no need for routine medical or cancer-related care.[25] This attitude may, however, reflect unawareness of their long-term risks of sequelae.[23] Given these survivors' strong intrinsic motivation for health, they may benefit from distance-based and autonomy-supportive intervention strategies delivered by experts in survivor care at major cancer treatment centers; such interventions could include print media and/or telephone interaction addressing the risks of late sequelae and recommending screening/follow-up. Because these survivors are the least worried about their future health, the information provided should be detailed and graphic in explaining the probability and nature of risks, given their treatment exposures. Increasing the affective response of survivors in this group may be a successful strategy to enhance their overall concern and interest in follow-up care.[9, 40]

The worried group was the class least likely to have easy access to risk-based health care, given their demographic and resources profile. A recent study[11] found that survivors who were African-American, older at interview, or uninsured were less likely to have received risk-based, survivor-focused care. Identifying these survivors through established long-term survivor cohorts (CCSS, British Childhood Cancer Survivor Study, and Childhood Cancer Research Network,[41]) treating institutions, and community solicitation (web or other media) and providing them with both distance-based (mail and/or telephone) and local care resources may facilitate their participation in medical follow-up care.

The worried group was also distinguished by cancer-related fears and anxieties. These fears, together with the group's perceptions of poorer health and of a greater threat posed by chronic illness, may deter participation in screening/follow-up. Fear, worry, and anxiety exert both positive and negative influences on health-related behaviors[9, 42] and can support or inhibit motivation toward health-promoting behavior. Although early detection through screening may positively modify a disease course, the prospect of learning that one has a serious health condition can be profoundly frightening.[4243] Survivors may either resort to avoidance behavior (e.g., not going for routine evaluations) or adhere to follow-up recommendations to reduce the discomfort caused by fear and anxiety.[40, 42] Misconceptions and lack of specific information about risk and risk factors can exacerbate fear or contribute to denial of the possibility of significant health problems.[40, 42,-43] Survivors in the worried class would likely be receptive to individualized print summaries detailing their long-term risks and ways to reduce those risks. To avoid further escalating fears and concerns, the tone of the information offered should be non-threatening.

Survivors in the collaborative group were both receptive to professional intervention and highly motivated to assume a primary role in maintaining their health. They are likely to work effectively with their providers if they are made aware of their risks and the actions necessary to reduce those risks. Like many survivors, however, they may not fully understand their risks or realize that specific follow-up care is recommended to modify the risks. Previous studies have found substantial knowledge deficits and misperceptions among childhood cancer survivors about their cancer diagnosis, treatment, and cancer-related health risks.[4451] Although information about late effects is increasingly being provided by family and oncology clinicians, persistent knowledge deficits may limit adult survivors' participation in screening and other risk-reducing interventions.[51] Collaborators may be highly motivated to follow through on screening recommendations if they are: 1) provided with treatment summaries detailing their specific risks; 2) given specific recommendations for screening and follow-up; and 3) made aware of community resources for care. Cancer treatment centers can provide survivors and their providers with print summaries of individual risks and recommended follow-up at the end of therapy and periodically afterward (as reminders to seek follow-up care).

Tiered or staged follow-up approaches for childhood cancer survivors have recently been proposed.[5253] Levels of care range from providing the survivor with individualized print materials detailing treatment and recommended follow-up screening, to offering telephone consultation and support for specific risks, to care at specialized late effects clinics. These models have been based largely on survivors' diagnoses and treatment exposures. Other models of care are provider-based, with providers ranging from nurse specialists to a primary care physician and a panel of sub-specialists.[54] We offer yet another option to enhance emerging models of survivorship care – using survivor profiles as well as cancer treatment history and level of provider to inform tailored follow-up care.


The study sample reflects a subset of the overall CCSS population, that is, those who responded to the HCNS; therefore, survivors included in the present analysis may not be fully representative of the population from which they were derived. While the CCSS population represents a large and heterogeneous cohort of 5-year survivors, results may not be generalizable to all survivors of childhood cancer. Because the approach for determining the optimal number of classes in latent class analysis is less standardized than with other analytic strategies, it is possible that childhood cancer survivors comprise more than the three classes we identified. The BLMRT, however, did not support more classes. While we included a large number of covariates in the analysis, there may be additional variables that differ across classes but were not available in our data set.


We identified three distinct profiles of childhood cancer survivors on the basis of responses to affective, cognitive, and motivation items related to their disease and treatment. Significant class covariates included sex, self-reported health status, and the self-reported presence/severity of late effects. These profiles predicted whether survivors were likely to obtain routine medical check-ups and/or participate in cancer-related medical follow-up. Collectively, our findings suggest that behavior related to modification of late-effects risks is predicted by a constellation of factors that are uniquely configured in discrete subgroups of survivors. The findings further suggest that a “one size fits all” approach is not likely to be the optimal strategy for models of survivorship care. Items useful in profiling survivors' likely behavior patterns and their specific knowledge about treatment and long-term follow-up recommendations could be used for brief screening at the end of treatment and periodically throughout survivorship to identify those at risk and inform tailored intervention strategies. Current and future studies are focused on the utility of the class profiles in predicting actual participation in medical surveillance and behaviorally-related health outcomes.


Sharon Naron, MPA, ELS for editorial assistance.

Jessica Phillips, Felicia Carson, and Camille Davis for assistance with graphics

This work was supported by grants RO3 NR009203 (CL Cox, PI) and U24 CA55727 (LL Robison, PI) from the U.S. Public Health Service, the Robert Wood Johnson Generalist Physician Faculty Scholar award (Oeffinger KC, PI), and the American Lebanese Syrian Associated Charities (ALSAC).


1. Pui CH, Relling MV, Downing JR. Acute lymphoblastic leukemia. N Engl J Med. 2004;350:1535–48. [PubMed]
2. Ries LAG, et al., editors. SEER Cancer Statistics Review, 1975–2004. Vol. based on November 2006 SEER data submission, posted to the SEER web site, 2007. National Cancer Institute; Bethesda, MD: 2007.
3. Hudson MM, Mertens AC, Yasui Y, Hobbie W, Chen H, Gurney JG, Yeazel M, Recklitis CJ, Marina N, Robison LL, Oeffinger KC. Health status of adult long-term survivors of childhood cancer: a report from the Childhood Cancer Survivor Study. JAMA. 2003;290:1583–1592. [PubMed]
4. Mattano LA, Sather HN, Trigg ME, Nachman JB. Osteonecrosis as a complication of treating acute lymphoblastic leukemia in children: a report from the children's cancer group. Clinical Oncology. 2000;18:3262–3272. [PubMed]
5. Friedman DL, Whitton J, Leisenring W, Mertens AC, Hammond S, Stovall M, Donaldson SS, Meadows AT, Robison LL, Neglia JP. Subsequent neoplasms in 5-year survivors of childhood cancer: the Childhood Cancer Survivor Study. J Natl Cancer Institute. 2010;102:1083–1095. [PMC free article] [PubMed]
6. Simbre VC, Duffy SA, Dadlani GH, Miller TL, Lipshultz SE. Cardiotoxicity of cancer chemotherapy: implications for children. Peadiatric Drugs. 2005;7:187–202. [PubMed]
7. Children's Oncology Group The Children's Oncology Group Long-Term Follow-up Guidelines for Survivors of Childhood, Adolescent, and Young Adult Cancers. 2006 [cited October 22, 2007]; Available from:
8. Bellizzi KM, Rowland JH, Jeffery DD, McNeel T. Health behaviors of cancer survivors: Examining opportunities for cancer control intervention. J Clin Oncol. 2005;23:8884–8893. [PubMed]
9. 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. [PubMed]
10. Klosky JL, Cash DK, Buscemi J, Lensing S, Garces-Webb DM, Zhao W, Wiard S, Hudson MM. Factors influencing long-term follow-up clinic attendance among survivors of childhood cancer. J Cancer Surviv. 2008;2:225–32. [PMC free article] [PubMed]
11. Nathan PC, Greenberg ML, Ness KK, Hudson MM, Mertens AC, Mahoney MC, Gurney JG, Donaldson SS, Leisenring WM, Robison LL, Oeffinger KC. Medical care in long-term survivors of childhood cancer: a report from the childhood cancer survivor study. J Clin Oncol. 2008;26:4401–9. [PMC free article] [PubMed]
12. Sheen V, Tucker MA, Abramson DH, Seddon JM, Kleinerman RA. Cancer screening practices of adult survivors of retinoblastoma at risk of second cancers. Cancer. 2008;113:434–441. [PMC free article] [PubMed]
13. 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. [PubMed]
14. Yeazel MW, Oeffinger KC, Gurney JG, Mertens AC, Hudson MM, Emmons KM, Chen H, Robison LL. The cancer screening practices of adult survivors of childhood cancer. Cancer. 2004;100:631–640. [PubMed]
15. Bauman AE, Sallis JF, Dzewaltowski DA, Owen N. Toward a better understanding of the influences on physical activity: the role of determinants, correlates, causal variables, mediators, moderators, and confounders. Am J Prev Med. 2002;23:5–14. [PubMed]
16. Breslow L, Lloyd D, Shumaker SA. Disease Prevention Research at NIH: An Agenda for All. Workshop B: Health behaviors--predictors, mediators, and endpoints. Prev Med. 1994;23:552–3. [PubMed]
17. Jeffery RW. How can health behavior theory be made more useful for intervention research? Int J Behav Nutr Phys Act. 2004;1:10. [PMC free article] [PubMed]
18. Hudson MM, Mertens AC, Yasui Y, Hobbie W, Chen H, Gurney JG, Yeazel M, Recklitis CJ, Marina N, Robison LL, Oeffinger KC, Childhood Cancer Survivor Study Investigators Health status of adult long-term survivors of childhood cancer: a report from the Childhood Cancer Survivor Study. JAMA. 2003;290:1583–1592. [PubMed]
19. American Academy of Pediatrics Section on Hematology/Oncology Children's Oncology Group Long-term follow-up care for pediatric cancer survivors. Pediatrics. 2009;123:906–915. [PMC free article] [PubMed]
20. Oeffinger KC, Robison LL. Childhood cancer survivors, late effects, and a new model for understanding survivorship. JAMA. 2007;297:2762–2764. [PubMed]
21. Oeffinger KC, Mertens AC, Hudson MM, Gurney JG, Casillas J, Chen H, Whitton J, Yeazel M, Yasui Y, Robison LL. Health care of young adult survivors of childhood cancer: a report from the Childhood Cancer Survivor Study. Annals Family Medicine. 2004;2:61–70. [PubMed]
22. Oeffinger KC, Wallace WH. Barriers to follow-up care of survivors in the United States and the United Kingdom. Pediatric Blood Cancer. 2006;46:135–142. [PubMed]
23. Kadan-Lottick NS, Robison LL, Gurney JG, Neglia JP, Yasui Y, Hayashi R, Hudson M, Greenberg M, Mertens AC. Childhood cancer survivors' knowledge about their past diagnosis and treatment: Childhood Cancer Survivor Study. JAMA. 2002;287:1832–1839. [PubMed]
24. Nathan PC, Ford JS, Henderson TO, Hudson MM, Emmons KM, Casillas JN, Lown EA, Ness KK, Oeffinger KC. Health behaviors, medical care, and interventions to promote healthy living in the Childhood Cancer Survivor Study cohort. J Clin Oncol. 2009;27:2363–2373. [PMC free article] [PubMed]
25. Maeda N, Hortibe K, Kato K, Kojima S, Tsurusawa M. Survey of childhood cancer survivors who stopped follow-up physician visits. Pediatr Int. 2010 Apr 22; (Epub ahead of print) [PubMed]
26. Cox CL. Online exclusive: a model of health behavior to guide studies of childhood cancer survivors. Oncol Nurs Forum. 2003;30:E92–9. [PubMed]
27. 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. [PubMed]
28. Cox C. An interaction model of client health behavior: Theoretical prescription for research. Adv Nurs Sci. 1982;5:41–56. [PubMed]
29. Cox C. Response to “Patient-centered care: Understanding its interpretation and implementation in health care” Sch Inq Nurs Pract: An International Journal. 2000;14:183–187. [PubMed]
30. Cox CL. The individual as client. In: Sullivan J, editor. In Directions for Community Health Nursing. Blackwell Scientific; Boston: 1984. pp. 129–172.
31. 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. [PubMed]
32. Cox CL, Montgomery M, Oeffinger KC, Leisenring W, Zeltzer L, Whitton JA, Mertens AC, Hudson MM, Robison LL. Promoting physical activity in childhood cancer survivors: Results from the Childhood Cancer Survivor Study. Cancer. 2009;115:642–54. [PMC free article] [PubMed]
33. Finnegan L, Wilkie DJ, Wilbur J, Campbell RT, Zong S, Katula S. Correlates of physical activity in young adult survivors of childhood cancers. Oncol Nurs Forum. 2007;34:E60–9. [PubMed]
34. Cox CL, Oeffinger KC, Montgomery M, Hudson MM, Mertens AC, Whitton J, Robison LL. Determinants of mammography screening participation in adult childhood cancer survivors: Results from the Childhood Cancer Survivor Study. Oncol Nurs Forum. 2009;36:335–344. [PMC free article] [PubMed]
35. Cox CL, Hudson MM, Mertens AC, Oeffinger KC, Whitton J, Montgomery M, Robison LL. Medical screening participation in the childhood cancer survivor study. Arch Intern Med. 2009;169:454–62. [PMC free article] [PubMed]
36. Robison LL, Mertens AC, Boice JD, Breslow NE, Donaldson SS, Green DM, Li FP, Meadows AT, Mulvihill JJ, Neglia JP, Nesbit ME, Packer RJ, Potter JD, Sklar CA, Smith MA, Stovall M, Strong LC, Yasui Y, Zeltzer LK. Study design and cohort characteristics of the Childhood Cancer Survivor Study: A multi-institutional collaborative project. Med Pediatr Oncol. 2002;38:229–39. [PubMed]
37. Wallston KA, Wallston BS, DeVellis R. Development of the Multidimensional Health Locus of Control (MHLC) Health Educ Monogr. 1978;6:160–170. [PubMed]
38. Muthen B. Latent variable mixture modeling. In: Marcoulides GA, Schumacker RE, editors. In New Developments and Techniques in Structural Equation Modeling. Lawrence Erlbaum Associates; Florence, Kentucky: 2001. pp. 1–33.
39. Nylund KL, Asparouhov T, Muthen B. Deciding on the number of classes in latent class analysis and growth mixture modeling: A Monte Carlo simulation study. Structural Equation Modeling: A Multidisciplinary Journal. 2007;14:535–569.
40. Mullens AB, McCaul KD, Erickson SC, Sandgren AK. Coping after cancer: Risk perceptions, worry, and health behaviors among colorectal cancer survivors. Psycho-Oncology. 2004;13:367–376. [PubMed]
41. Steele JR, Wellemeyer AS, Hansen MJ, Reaman GH, Ross JA. Childhood cancer research network: A North American pediatric cancer registry. Cancer Epidemiol Biomarkers Prev. 2006;15:1241–2. [PubMed]
42. Cameron LD. Screening for cancer: Illness perceptions and illness worry. In: Petrie KJ, Weinman JA, editors. In Perceptions of Health and Illness: Current Research and Applications. Hardwood Academic; Amsterdam: 1997. ??
43. Millar MG, Millar K. Negative affective consequences of thinking about disease detection behaviors. Health Psychology. 1995;14:141–146. [PubMed]
44. Hopwood P. Breast cancer risk perception: what do we know and understand? Breast. Cancer Res. 2000;2:387–391. [PMC free article] [PubMed]
45. Mahdy N, Fatohy IM, Mounir GM, El-Deghedi BM. Assessment of students' knowledge, attitude, and practice concerning cancer and its prevention. Part I. J Egypt Public Health Assoc. 1998;73:399–431. [PubMed]
46. Pohls UG, Renner SP, Fasching PA, Lux MP, Kreis H, Ackermann S, Bender HG, Beckmann MW. Awareness of breast cancer incidence and risk factors among healthy women. Eur J Cancer Prev. 2004;13:249–256. [PubMed]
47. Bashore L. Childhood and adolescent cancer survivors' knowledge of their disease and effects of treatment. J Pediatr Oncol Nurs. 2004;21:98–102. [PubMed]
48. Byrne J, Lewis S, Halamek L, Connelly RR, Mulvihill JJ. Childhood cancer survivors' knowledge of their diagnosis and treatment. Ann Intern Med. 1989;110:400–3. [PubMed]
49. Hudson MM, Tyc VL, Jayawardene DA, Gattuso J, Quargnenti A, Greenwald C, Crom DB, Mason C, Srivastava DK, Hinds P. Feasibility of implementing health promotion interventions to improve health-related quality of life. Int J Cancer Suppl. 1999;12:138–42. [PubMed]
50. Kadan-Lottick NS, Robison LL, Gurney JG, Neglia JP, Yasui Y, Hayashi R, Hudson M, Greenberg M, Mertens AC. Childhood cancer survivors' knowledge about their past diagnosis and treatment: Childhood Cancer Survivor Study. JAMA. 2002;287:1832–9. [PubMed]
51. Caprino D, Wiley TJ, Massimo L. Childhood cancer survivors in the dark. J Clin Oncol. 2004;22:2748–50. [PubMed]
52. Freyer D. Transition of care for young adult survivors of childhood and adolescent cancer: Rationale and approaches. J Clin Oncol. 2010 Mar 29; Epub ahead of print. [PMC free article] [PubMed]
53. Oeffinger KC, Nathan PC, Kremer LCM. Challenges after curative treatment for childhood cancer and long-term follow up of survivors. Hematol Oncol Clin North Am. 2010;24:129–149. [PubMed]
54. Aslett H, Levitt G, Richardson A, Gibson F. A review of long-term follow-up for survivors' childhood cancer. Eur J Cancer. 2007;43:1781–1790. [PubMed]