Specialty Care for Adult Survivors of Childhood Cancer

The field of pediatric oncology has made enormous advances over the past decades, as evidenced in countless scholarly publications and the lay press.^1-3 Among the primary beneficiaries of these successes are the more than 400,000 adult survivors of childhood cancer currently living in the United States.⁴ Unfortunately, late effects of childhood cancer therapy are common, serious, and potentially life-threatening.^5,6 Recommended screening and surveillance for late effects are carefully outlined in the Children's Oncology Group guidelines,^7,8 but the pool of available oncologists is shrinking,⁹ and primary care providers express uncertainty regarding how best to care for survivors.^10-12 Furthermore, a fragmented approach to health care delivery in the United States results in a haphazard and nonsystematic approach to care for this high-risk population.

In this issue of Cancer, Sutradhar et al¹³ from Canada have taken advantage of the visionary 1999 decision by the Ontario Ministry of Health and Long-Term Care to initiate a provincial network of specialty clinics for adult survivors of childhood cancer. Using multiple linked administrative health databases, the authors created a population-based cohort of 3912 adult survivors of childhood cancer diagnosed between 1986 and 2005 in Ontario, Canada. Approximately 43% of this cohort attended at least 1 survivor clinic visit during the study period, and 14,187 emergency department (ED) visits were captured. An Andersen-Gill recurrent event multivariate regression model was used, and the results demonstrated that 1 or more prior visits to a survivor clinic was associated with a 19% lower rate of ED visits in comparison with survivors who had not visited a specialty survivor clinic. When a visit to a survivor clinic was considered a cumulative count covariate, each additional visit to a survivor clinic decreased the subsequent ED visit rate by 5%. In further analysis, these results were independent of whether or not the survivor was also seen by a primary care physician, and importantly, both low-acuity and high-acuity ED visits were reduced. On the basis of the strength of these findings, it is reasonable to state that attendance at a specialized survivor clinic results in fewer ED visits; whether underlying morbidity is also affected is an intriguing possibility.

The policy, economic, and clinical care implications of these findings are significant. The Ontario Ministry of Health and Long-Term Care made the prescient decision to support the creation of 5 specialty clinics for adult survivors of childhood cancer in 1999. In the United States, a small number of cancer center–based programs for adult survivors of childhood cancer exist, but their reach is limited, and reimbursement for services can be difficult.¹⁴ Reducing ED visits for survivors could be achieved with an expansion of specialty clinics in the United States, perhaps via a Medicare approval process, as is already in place for intensive cardiac rehabilitation programs and other facities.¹⁵ though it could be less problematical, establishing survivorship clinics in countries with single-payer health care systems will also require stalwart leadership.¹⁶ Results such as those presented here by Sutradhar et al¹³ will go a long way in supporting this process.

The economic implications of a potential reduction in ED visits among adult survivors of childhood cancer cannot be ignored. One hypothesis is that observed ED visits were reduced because survivors with a history of a survivorship clinic visit sought care in an outpatient setting rather than the ED. However, the analysis was adjusted for primary care services (annual physical examinations) and continuity of care with a single provider. In addition, it does not seem likely that high-acuity visits or ED-initiated hospitalizations, which were also reduced among those with a survivorship clinic visit, could be replaced in this way. Finally, regardless of whether underlying morbidity was reduced, this transfer of services could also be expected to reduce health care costs.¹⁷

Health care delivery researchers should also take note of Sutradhar et al's article¹³ with respect to the study design. Investigators used multiple linked administrative health databases to take advantage of a local policy decision: the institution of specialty survivorship clinics by the Ontario Ministry of Health and Long-Term Care. Because of limitations of the available data, these investigators could not use a before/after study design. Nonetheless, should a similar initiative take place elsewhere, the effects could be examined as a natural experiment.¹⁸

Finally, the clinical care implications of Sutradhar et al's findings¹³ must be acknowledged. Although underlying morbidity could not be examined, a substantial and significant reduction in subsequent ED visits after 1 or more survivorship clinic visits was reported. The analysis included a multivariate adjustment for cancer diagnosis, relapse, treatment characteristics (including treatment intensity), and demographic and socioeconomic variables. Therefore, although it is tempting to hypothesize that the observed associations are due to bias (sicker patients, who are more likely to visit the ED, did not attend survivor clinics), the strength of the findings after the multivariate analysis argues against this possibility. Therefore, we must ask whether care for adult survivors of childhood cancer is so specialized that it requires the contributions of a dedicated provider. As noted by the authors, such specialized clinics are intended not to replace care provided by a primary care clinician but rather to complement such care through a shared-care model.¹⁹ The report from Sutradhar et al in this issue of Cancer makes a robust argument for this strategy.

So, what are the next steps? First, further study is necessary to determine whether specialized survivorship clinics reduce hospitalization rates and indeed whether they lead to a reduction in serious morbidity and an increase in long-term survival. The design of such studies is complex; we encourage the use of natural experiments described previously and similar registry data-linkage studies. We also posit that many adult survivors of childhood cancer do not need specialized services because the risk of serious late effects associated with some cancer therapies is rather modest. Edgar et al,²⁰ extending previous work by Wallace et al,²¹ developed a simple method for predicting long-term outcomes and stratifying risk into 3 levels based on treatment exposures. Use of this tool should help to distinguish who might benefit from care in specialized survivorship clinics from those survivors who would simply be overmedicalized. Lastly, 1 population of cancer survivors in Sutradhar et al's study¹³ appeared to have a disproportionately low percentage of visits to the survivor clinics: central nervous system tumor survivors. In comparison with other cancer groups, survivors of a childhood central nervous system tumor have a substantially higher risk of having cognitive dysfunction and problems with sight, hearing, and balance and thus often become socially isolated.²² Thus, it is imperative that methods be developed and tested with the intent of minimizing the loss to follow-up of this particularly high-risk population.