The Dutch LATER physical outcomes set for self-reported data in survivors of childhood cancer

The vast majority of children diagnosed with cancer nowadays will achieve long-term survival [1, 2]. Those childhood cancer survivors (CCS) are a growing, vulnerable group of individuals who are at risk of developing long-term morbidity due to previous treatment for cancer in early stages of life. Knowledge on the burden of long-term morbidity in CCS, its underlying types of health conditions and its risk factors, has been presented in various studies during the past decades [3‐5].

In long-term morbidity research in CCS, a broad variety of outcome assessment methods is used. Long-term morbidity outcomes can be assessed by self-reporting via questionnaires [6‐24], by medical evaluation during outpatient clinic visits [25‐34] or by linkage with existing registries such as national hospital discharge registries [35‐39]. Authors often include different types and different numbers of organ systems in their calculations of physical long-term morbidity [6‐39]. Also, incidence or prevalence estimates are often reported without describing which health conditions or organ systems were included in these calculations. Definitions of long-term morbidity outcomes also vary, for example, authors reporting on cardiovascular conditions generally report on heart failure, myocardial infarction, and hypertension, but some also include stroke as a cardiovascular condition [10, 14, 17, 18, 36]. While many authors do not grade the severity of the reported long-term morbidity in CCS, others use the Common Terminology Criteria for Adverse Events (CTCAE) [40], either in its original form or an adapted version incorporating specific additional outcomes that authors considered missing [41‐43]. This lack of uniformity in types of outcomes, outcome definitions, and outcome grading—even among studies that use similar data ascertainment methods—limits interpretation, comparability, and generalizability of studies investigating the burden of long-term morbidity in CCS. Furthermore, the described outcomes in current studies include asymptomatic and symptomatic outcomes with or without treatment. To get a better insight in the overall burden for survivors, the Dutch LATER questionnaire study would like to evaluate only outcomes that are symptomatic and/or requiring medical treatment.

The aim of this study is to develop a set of self-reported long-term physical outcomes that are clinically relevant for CCS, defined as morbidities with clinical symptoms and/or requiring medical treatment, to investigate the burden of morbidity in the Dutch LATER questionnaire study.

We present a proposal for a core set of 74 self-reported long-term physical outcomes of clinical relevance in survivors of childhood cancer. By comparison of existing survivorship questionnaires and by reviewing every specific morbidity reported by CCS in the open text fields in our Dutch nationwide questionnaire study, we followed an innovative method which focuses on outcomes that are clinically relevant for the survivor, due to the fact that its presence influences daily life. Our outcome set will be used for investigating the burden of long-term morbidity in the Dutch LATER questionnaire study. This set can also be used for international harmonization of a uniform core outcome set for long-term morbidity in CCS, to facilitate worldwide collaboration in late effects research.

Compared with other grading scales used for long-term morbidity research in CCS, the newly developed Dutch LATER core outcome set differs on three important key points. First, this core outcome set was designed with the single purpose of investigating self-reported long-term morbidity in childhood cancer survivors, by combining existing questionnaires and outcomes reported by survivors. Second, we selected outcomes describing morbidity with clinical symptoms or requiring medical treatment, the so-called clinically relevant outcomes. Third we included outcomes where the treatment for childhood cancer caused direct damage that had persistent impact for the survivor also in later life, for example, limb amputation which results in a lifelong disability or removal of an eye which results in lifelong complications. Because the CTCAE criteria were originally designed for grading acute adverse events during adult cancer trials [54], the current CTCAE version 4.03 [40] does not cover the complete spectrum of long-term morbidity that CCS might encounter [42]. Several authors have already stated that relevant outcomes were missing for CCS and use adapted versions [41‐43]. Comparison of our core set of long-term self-reported physical outcomes to the commonly used CTCAE showed that 36% of the outcomes were not present in the CTCAE. Moreover, CTCAE does not incorporate self-reported data to assess long-term morbidity [42]. For nine out of the 48 conditions that were present in the CTCAE, we could not perform severity grading because detailed additional clinical information was needed for appropriate grading, which was not available from current questionnaires and is often too complicated to directly ask patients in a questionnaire. Although often only health conditions grade 3 and higher are included when studying severe physical long-term morbidity in CCS, our results show that many grade 2 conditions will have consequences for a survivor because of symptoms or needed treatment. From our core outcome set, up to 27 clinically relevant outcomes corresponded with CTCAE grade 2, for example, several endocrine deficiencies that require chronic medication use, and would have been missed in such studies. Comparison to the SJLIFE adapted CTCAE for grading of clinically ascertained data showed that more of our core outcomes were missing and that 24 clinically relevant conditions corresponded to grade 2 or even grade 1. Hence, our results support previous authors, concluding that the CTCAE in its current form is not optimal to grade severity of (self-reported) long-term physical morbidity outcomes for CCS [41‐43]. To our knowledge, this is the first comprehensive proposal to define a core outcome set for self-reported long-term physical outcomes in CCS. A strength of this study is that we focused on clinical relevance for CCSA and a limitation is that we were not yet able to incorporate the prioritization of outcomes by survivors. This can be the focus of future research. Also, because the purpose of this core outcome set was facilitating the investigation of physical long-term morbidity in the Dutch LATER cohort, the proposed outcome definitions reflect the agreement among the Dutch LATER experts team only. To overcome any subjectivity in outcomes used by various childhood cancer survivorship research groups, we advocate international harmonization of a core outcome set for physical long-term morbidity in childhood cancer survivors. A uniform global core outcome set is highly needed to enable comparison of future long-term morbidity studies, to uniformly evaluate survivorship care and to facilitate collaboration within survivorship research. The International Guideline Harmonization Group [55] started an initiative to develop a harmonized outcome set by a Delphi method. This will facilitate international collaboration and data pooling.

In conclusion, we propose a Dutch LATER core set of self-reported long-term physical outcomes of clinical relevance for CCS that will be used to investigate the burden of long-term morbidity in childhood cancer survivors from the Dutch LATER questionnaire study. We advocate to start international discussion and research to harmonize long-term physical morbidity outcomes that are clinically relevant for CCS.