Contributors and guarantor: Morag Heirs, Sara Suekarran, Russell Slack and Kate Light worked on this review and prepared the final report. R.P. conceived the original idea in discussion with F.G., A.G. and M.H. and prepared the initial funding application. M.H. took responsibility for developing the protocol, M.H., S.S. and R.S. carried out screening, data extraction, quality assessment and synthesis. K.L. designed and carried out database and other literature searches (see Methods Section for full details). The project was managed by R.P. who took overall responsibility for the project. M.H. prepared this paper and all authors contributed to drafts and commented on the article. Clinical advisors were involved throughout the project and contributed to the paper: Faith Gibson, Adam Glaser and Mike Hawkins. We would like to acknowledge the input from our two patient representatives and colleagues from CRD, Sian Thomas and Alison Eastwood.
All the authors critically revised the manuscript and approved the final version. The views expressed here are those of the authors and do not necessarily reflect the views of MacMillan Cancer Support
Ethics: No ethical approval was required for this study.
Role of study sponsors: The funders had no role in the study design; data collection, analysis, and interpretation; writing the manuscript; or the decision to submit the research for publication.
Childhood cancer is rare. In developed countries approximately 0.5% of all cancers occur in children of 15 years or younger. In the UK, cancer accounts for approximately 20% of all deaths between the ages of 1 and 14 years and each year approximately 1,400 new cases of childhood cancer are diagnosed 1, 2. Improvements in the precision of diagnosis, therapy and supportive care have led to increasing survival rates since the 1960s, resulting in greatly increased numbers of adult survivors of childhood cancer, for example in 2000 there were almost 15,000 adult survivors, 45% of whom were aged 30 years or over 1, 2.
Children who survive childhood cancer are at risk of late-effects as a result of intensive multi-modality therapeutic strategies, for example surgery, radiotherapy, chemotherapy and combinations of these 3, 4. These can vary from physical health effects, such as second primary tumours (present in around 4% of survivors), secondary leukaemia (around 0.2% of survivors), cardiovascular disease as an effect of radiation and chemotherapeutic agents (or as an indirect effect) and effects on fertility and growth, educational, behavioural and social problems 5. Relatively little is known about the long-term consequences of therapy and the need for long-term surveillance has been identified in order to better characterise late-effects specific to this group of people 6. Guidelines recommending long-term follow-up of survivors have been produced by a number of influential organisations, however there remains a lack of consensus regarding the optimal setting and strategy 5, 7, 8. The cost of providing long-term follow-up programmes is unknown at present.
This systematic review was commissioned in partnership by MacMillan and the Department of Health as part of the National Cancer Survivorship Initiative (NCSI) within the Children and Young People (CYP) work stream. The NCSI-CYP programme also includes the establishment and monitoring of test sites which are exploring issues and models of aftercare. This review aimed to support this by comparing different models of follow-up care for survivors of childhood cancer including alternative modes of communication, nurse versus physician led care and hospital versus primary care settings.
SYSTEMATIC REVIEW METHODS
The search strategy comprised the searching of electronic databases and appropriate electronic resources, contact with experts in the field and scrutiny of bibliographies of included studies and existing reviews. A date limit was used to retrieve studies published between 1980 and January/February 2010 to minimise the presence of irrelevant papers since follow-up studies first appeared in the mid 1980s. The final database selection was made to include representatives from general medicine, psychology and nursing as well as ongoing trials and other literature that may not have been formally published.
The following electronic databases were searched from 1980 to February 2010 without language restrictions: MEDLINE, Cochrane Database of Systematic Reviews (CDSR), Database of Abstracts of Reviews of Effects (DARE), NHS Economic Evaluation Database (NHS EED), Health Technology Assessment (HTA) Database (HTA), Cochrane Central Register of Controlled Trials (CENTRAL), PsycINFO, Health Management Information Consortium (HMIC), CINAHL and Current Controlled Trials.
Internet resources that were searched included: American Society of Clinical Oncology, National Cancer Institute Clinical Trials and International Cancer Research Portfolio.
Survivors of cancers diagnosed by 18 years of age were eligible. Studies of follow-up or aftercare commencing after primary treatment had ceased were of relevance. The interventions were considered in three groups: a comparison of alternative communication modalities to face-to-face clinic visits (telephone, postal, email or SMS/text-based); the use of physician- versus nurse-led follow-up; the value of hospital staff versus primary care staff to provide clinical contacts.
Comparative data from retrospective or prospective groups were the intended focus of the review, but uncontrolled studies which included practice audit/evaluations were also documented. Patient experience, percentage response to the contact and patient satisfaction were eligible for inclusion, as were comparative rates of detection of morbidity (e.g., cardiac dysfunction), secondary malignant neoplasm and mortality rates.
Two reviewers independently screened all titles and abstracts. Full paper manuscripts that appeared relevant were obtained where possible. The relevance of each study was assessed independently by two reviewers according to the inclusion criteria described in the previous section. Discrepancies were resolved by discussion, or by referral to a third reviewer when necessary.
This review intended to include studies which presented comparative data from retrospective or prospective groups. Due to the paucity of research in this area uncontrolled studies which included practice audit/evaluations were also documented. Details of the planned data extraction, quality assessment and data synthesis for use with comparative studies are given in the full report 9, 10.
Uncontrolled single-arm observational studies were documented by one reviewer and checked by a second using a standard data extraction proforma including study details, patient characteristics, intervention, comparison, study design and relevant outcomes.
Data extracted from the studies that were retained, despite not meeting study design criteria, were tabulated and discussed in a narrative review. No formal quality assessment was carried out for these studies due to the difficulty in assessing single arm audit/observational studies; however, relevant elements were considered including sample characteristics, the possibility of bias in response rates and allocation. The narrative summary covers the nature of the follow-up care offered, how it was evaluated and the resultant outcomes. Studies were grouped according to the type of follow-up care and setting offered. Comparisons were drawn between, and within, categories where possible. Recommendations for further research have been made as a result of gaps in the evidence base and the existing data.
The literature search yielded 4,010 studies, of which 266 references were considered potentially relevant to the research question and full papers were ordered (Fig. 1). A total of 266 full papers were ordered, of which six papers were unavailable through the Internet or British library. Our attempts to include non-English language papers appeared to be relatively successful; of the five non-English papers eligible for assessment, only one paper could not be translated or assessed (Japanese).
We were unable to identify any comparative studies which evaluated methods of providing follow-up care for survivors of childhood cancer as set out in our inclusion criteria. Where observational, single arm or audit-style studies were identified that would otherwise have met our inclusion criteria, these references were retained and are described below. A single comparative study comparing two different hospital based clinics was also documented; these studies do not directly answer the review questions but they do illustrate the nature of the research currently available and highlight areas for future development.
A total of eight papers were retained despite not meeting the study design inclusion criteria with the intention of reflecting the evidence base as it currently stands (Table I) 11–17. All but one of these papers were available as full publications; the remaining study was published only as an abstract reporting provisional results 11. Mostly these studies were feasibility or audit type studies which aimed to clarify whether a particular service design was feasible, and if patients found it useful. The single comparative cohort study compared follow-up settings where care was delivered in either a traditional paediatric late effects clinic or a multi-disciplinary adult clinic, both of which were hospital based 18. This study was included to provide as comprehensive a picture of the evidence base in this area as possible, although we acknowledge it does not directly answer our research question.
Table I. Table of Documented Studies.
No. of participants
Summary of results
Problem-oriented and informal follow-up
2 audits of patients not attending FU or discharged (Canada and Sweden)
38% of discharged patients were dissatisfied with their care, nearly half reported a late-effect symptom and a further 17% were likely to have one. 80% of patients without regular follow-up might benefit from this
The new clinic identified new medical problems in up to 30% of patients and patient feedback suggested the treatment plans were better coordinated. The in-house nature of the clinic made it easier to schedule appointments
Late effects hospital based clinic
3 studies; two uncontrolled cohorts, one audit (all UK)
Overall satisfaction was high and 2/3 of patients attended FU although only 50% felt the visits were beneficial. 41% of patients felt they had no related problems and would prefer to see their GP in future
Clinical care rated more highly than supportive care overall. Supportive care rated more highly by women, survivors with more late-effects and poorer mental health. Consultant-led FU care rated most highly out of four options. Mean satisfaction score reported but unclear how this should be interpreted
Adult versus paediatric hospital clinic
1 comparative study looking at predictors of satisfaction (UK)
Found no differences in satisfaction between paediatric and adult clinics; key predictors of satisfaction were aspects of clinic organisation
Participants were largely convenience samples of patients already attending/recently referred to the clinics under evaluation 13–17, or in two cases patients who were not currently being followed-up for various reasons 11, 12. All participants were survivors of cancer diagnosed in childhood; where this definition was explicit the diagnosis had to have been made before the age of 16 or 17 years in most cases. In general all childhood cancers were included in the samples with only one study excluding central nervous system tumours 13. The sample size of survivors were small to moderate (42–245 patients) in size and response rates to both invitations to attend and return of questionnaires varied considerably.
The outcomes measured included knowledge about the purpose of follow-up, late-effects symptoms, satisfaction with the service (views of patients, families and doctors) and preferences for future follow-up care. Where details of the evaluation methods were reported these appeared to be internally compiled questionnaires which used a combination of symptom or side-effect lists and Likert rating scales. Only one of these measures was reported as having been validated or piloted 15, making it difficult to assess the reliability or accuracy of the findings.
Problem-Oriented and Informal Follow-Up
One audit from Sweden evaluated their unstructured, problem oriented follow-up programme and concluded that this was not a successful model on the basis of 38% of patients reporting some level of dissatisfaction with their follow-up care to date 12. These patients evaluated the follow-up care they had received to date, although it was not clear what this comprised. A second audit from Canada has reported preliminary results of inviting patients at high risk for late-effects based on previous treatments to attend a follow-up clinic 11. Attendance led to increased knowledge and in 50% of patients new problems were diagnosed, of which 70% required some kind of intervention.
The observational study by Blauuwbroek et al. outlines a shared care model combining hospital clinic-based provision with family doctor (primary care) in the Netherlands. The results suggest this is both feasible, and acceptable, to the majority of both patients and family doctors 13.
The multidisciplinary clinic in this study was established to meet an informally acknowledged need to better fulfil the medical requirements of patients with multiple late-effects and was monitored via an observational study 14. The annual visit or traditional late effects clinic previously provided to patients was reported to be insufficient based on the number of patients who failed to follow through on referrals (20%). This new clinic model involved same-day, same-clinic, access to psychological support, tests, information and medical information. It was felt to overcome difficulties in the original follow-up care programme, families were unanimously satisfied and reported they would use the clinic again, and benefits around scheduling appointments were identified. The authors report that it enhanced clinical efficiency, however no economic data were presented in terms of cost of the service.
Late Effects Hospital Based Clinics
The paper by Eiser et al. 15 described an audit of a long-term follow-up hospital-based clinic which aimed to encourage the transfer of care from the parent to the young adult, and generally provide information on long-term prognosis to survivors who had been off treatment for at least 5 years. The majority of patients were classed as satisfied with the care overall and in particular 41% of patients felt they had no health related problems and were more likely to prefer being seen by their GP for future follow-up. This study did not report any data around late-effects symptoms or new diagnoses, but relied on patient and family self-reports. The authors concluded that there may be a sub-group of patients for whom hospital-based follow-up is not in fact a desirable outcome based on the proportion of patients who preferred seeing their GP for future follow-up.
Kinahan et al. 16 reported on the perspectives of survivors and their parents who had participated in a survivorship clinic which was hospital-based and included nurse and physician specialists. A cohort comprising all registered survivors were invited to take part and 53% were interviewed. Of most relevance to this review were the results around perceived benefits of follow-up care. Almost all of the survivors reported at least one benefit, and the top three benefits were considered to be: late-effects care; personal relationship with the nurse; and health maintenance. These results applied only to survivors who opted to attend the clinic, and the mean time since diagnosis was 18 years indicating commitment by the attendees.
Finally, a late-effects hospital clinic offering consultant led support with input from a paediatric oncologist, endocrinologist and late effects special nurse was evaluated in a cohort of young adults 17. Generally, clinical care was rated as more important than supportive care. In this study the authors described clinical care as the checking for signs/symptoms of potential late effects, looking at development; whereas supportive care was more oriented towards lifestyle advice, reintegrating into education and so forth. Patients with more late-effect symptoms, higher future vulnerability and lower mental health related quality of life rated supportive care more highly. It was not possible to interpret the satisfaction results due to poor reporting of the outcomes in the original paper. When asked to rate follow-up care options, survivors preferred consultant-led care over nurse, GP or telephone/postal based follow-up.
These hospital-based late-effects clinics clearly varied in their aim and the services provided. All appeared to offer access to more than one healthcare professional (usually a specialist nurse plus a physician or consultant) and appeared to be a reduced version of the multidisciplinary clinic models.
Paediatric Versus Adult Clinic (Hospital Based)
The study by Absolom et al. 18 compared predictors of patient satisfaction in attendees of a traditional paediatric late-effects clinic and a multi-disciplinary adult setting clinic. Appointments were of longer duration in the paediatric clinic (30 minutes vs. 10 minutes) and the clinic was staffed by specialist paediatric oncologists among other professionals.
Overall, survivors were satisfied with the care they were receiving and there was no evidence that either group was more or less informed or felt at-risk to future problems. Survivors who understood that the purpose of follow-up care was for clinical support were more satisfied than those expecting psychological support. It was aspects of clinic organisation (e.g., waiting time, length of consultation) rather than the setting or clinic type which seemed to influence patient satisfaction.
Summary of Results
Given the heterogeneity of the evaluation tools and follow-up programmes, it was difficult to draw any overall conclusions as to clinical benefits or perceived patient needs. The outcomes reported were based almost entirely on perceived patient or parent/carer satisfaction, with little clinical data on the success of clinics in detecting late-effects that may require further treatment. Clinical care was valued highly by the majority of follow-up clinic attendees, as might be expected for these self-selected samples. Supportive care was perceived as more important by patients who required more clinical interventions and were experiencing more late-effects symptoms 14. There was a contrast between findings that patients who were not followed-up may have been receiving inadequate care both in terms of their perceived satisfaction and detection of late-effects which require treatment 11, 12, and the idea that there may be a sub-group of patients for whom long-term follow-up is not an essential part of care 15. These findings may represent different sub-groups of patients (potentially relating to risk stratification models), or they may be conflicting results from similar patient groups. As the editorial around the observational study of shared follow-up care combining primary and secondary care has emphasised, this study represents an important step in prospective evaluation of service provision and highlights the need for comparative studies in this area 13, 19. The multidisciplinary survivorship clinic was described in some detail and appears to provide a comprehensive integrated service which was valued by parents/carers and resulted in greater uptake of services 14.
A variety of models have been explored and evaluated, however the data have been largely observational in nature and, therefore, potentially vulnerable to the biases inherent in such research designs. The evaluation tools used in these studies may not have been reliably developed and validated, and lack of comparability makes it more difficult to draw together the available evidence.
The aim of this review was to identify studies that presented comparative data from retrospective or prospective groups that evaluated methods of providing follow-up care for survivors of childhood cancer. Despite a rigorous search for studies of this design we were unable to identify any comparative studies. The conclusion of a systematic review that there is too little, or no, evidence to answer the research question is not unusual, particularly in reviews of healthcare interventions 20, 21. Such findings, however, can inform future research priorities. The review has identified areas for focussing future research efforts based on the observational studies retained and it is envisaged that the absence of research findings will provide impetus for the development of rigorous comparative evaluations that seek to address these comparisons in relevant settings and for relevant samples of patients.
The main strengths of this review rest in the systematic attempts to identify and assess the literature. We carried out comprehensive literature searches including a thorough search of relevant electronic databases. Our study selection process was independently undertaken by two reviewers, with reference to a third reviewer where necessary, thus minimising the impact of error and bias. We were able to obtain and assess almost all of the potentially relevant papers. Attempts were made to interpret foreign-language papers, only one of which was not translated.
The coverage of the available literature achieved by the review suggests that we can be confident that important papers were not missed. It appears unlikely that the small number of unobtainable papers would have altered the results of the review.
A crucial part of the review process was to incorporate the views of young people, two of whom were contacted via our clinical advisors. The rapid nature of the review and lack of any substantial findings meant that we were unable to involve them as much as we had hoped within the review process; however their comments have been incorporated within the discussion and suggested areas for future research. Involving patients in systematic reviews can be difficult where the project is of short duration, and we intend to prioritise this in future work.
The studies discussed within this review were not formally quality assessed as the majority of the studies were single arm observational/audit type design and therefore difficult to assess for validity, therefore, caution is advised when considering these findings. We have highlighted where the populations were particularly heterogeneous and as a result have drawn only tentative conclusions while focusing on areas for further rigorous comparative research.
This review has demonstrated that there may be opportunities to undertake research in the following areas of primary research: there is a clear need for follow-up programs for survivors of childhood cancer to be evaluated using comparative study designs and cost-effectiveness should also be considered. Risk stratification may be particularly relevant, published guidelines advocate a stratified follow-up programme although there has been no evaluation of these recommendations to date.
The ongoing programme of test-site work commissioned by the NCSI-CYP will present models of care suitable for evaluation and inform decisions regarding funding and research into the provision of long-term follow-up services for survivors of childhood cancer.