Existence of late-effects instruments for cancer survivors: A systematic review

Introduction The number of cancer survivors is projected to increase to 22.1 million by 2030. Late effects incorporate the full domains of cancer survivorship (e.g., physiologic, psychosocial, economic). They are numerous, complex, and potentially alter the life trajectories of cancer survivors. Currently, research is missing on the impact of late effects (e.g., cardiomyopathy, fertility, lymphedema, anxiety) on cancer survivors. Objective The goal of this study is to present a systematic review of existing instruments for identifying, diagnosing, and managing late effects within cancer survivors. Methods Using PRISMA guidelines, a systematic search was conducted using the electronic databases of PubMed and Web of Science to identify relevant papers. Articles considered eligible for this review met the following criteria: 1) written in English, 2) published until September 30, 2019, and 3) containing instruments with questions on late effects. Hypothesis, study design, study sample, questionnaire domains, details of late effects, results, conclusions, and advantages/disadvantages of each article were assessed using a modified version of the NHLBI quality assessment tool. Results An exhaustive literature review revealed 576 publications in PubMed, 628 in Web of Science, and 260 from additional sources. After removing duplicates, articles without late-effects questionnaires, and publications using identical questionnaires, 11 studies fulfilled the eligibility criteria. Study quality assessment was measured on a scale of 0–6 (0 = poor quality; 6 = highest quality). Only one study was rated with a score of 5 (Rocque). Conclusions Taken in totality, none of the studies adequately addressed the prevalence, etiology, characteristics, management, and prevention of late effects. There is currently no comprehensive questionnaire that captures all of the relevant details of late effects across the cancer survivorship continuum nor that tracks the interrelatedness of multiple late effects. Hence, it is difficult to identify, diagnose, manage, and ultimately prevent late effects.


Introduction
There are 16.9 million cancer survivors in the United States, representing 5% of the population [1]. The number of cancer survivors is projected to increase to 22.1 million, by 2030 [1]. Nearly two-thirds of these individuals are 65 years or older [1]. As the population of cancer survivors in the US grows, it becomes essential to optimize health care delivery and long-term outcomes among survivors [2]. People with a history of cancer have unique medical and psychosocial needs that require proactive assessment and management by their oncologists and primary care providers [1]. Cancer survivors deal with both long-term effects and late effects.
Long-term effects refer to any side effects or complications of treatment for which a cancer patient must compensate; they begin during treatment and continue beyond the end of treatment [3] (heart, lung, kidney, or gastrointestinal tract problems; pain, numbness, tingling, loss of feeling, or heat or cold sensitivity in the hands or feet; fatigue; metabolic syndrome; bone loss; hearing loss; cataracts; dry eyes or dry mouth; and financial toxicity) [4,5]. Late effects refer specifically to unrecognized toxicities that are absent or sub-clinical at the end of therapy and manifest later [3]. These effects appear months to years after the completion of treatment [5] and their risk increases over time [6,7]. Currently, three out of five survivors develop late effects [8]. While most late effects are not life-threatening, they may cause serious problems that affect health and quality of life (physical, psychological, social, and spiritual wellbeing) [3,7]. Late effects are rarely examined in isolation; rather they incorporate the full domains of cancer survivorship (e.g., physiologic, psychosocial, economic) [6]. Adverse late effects can be numerous and complex and potentially alter the life trajectories of cancer survivors [9]. Hence, survivors can struggle daily with: a) the physical sequelae of late effects (e.g., cardiomyopathy, central nervous system problems [thinking, learning, memory, fatigue], sexual health and fertility, and lymphedema) as well as risks of cancer recurrences, b) psychosocial sequelae of late effects such as anxiety, depression, relationship complications, body image disturbances, and poor self-esteem, and c) financial consequences of late effects including unemployment, medical insurance, and finances [6,[10][11][12]. Research is urgently needed into the impacts of late effects on the physical and psychological health and quality of life of cancer survivors.

Objectives
The goal of this study is to present a systematic review of existing instruments for identifying, diagnosing, and managing late effects within cancer survivors.

Methods
This review was conducted following the guidelines outlined in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The (PRISMA) Statement [13] (S1 Fig).

Search strategy
A systematic search was performed using the electronic database PubMed to discover all relevant articles. The search included the keywords: late effects, cancer survivors, instrument, questionnaire, and survey. A second search was completed using the same keywords in the Web of Science to ensure attaining all relevant articles. Additionally, articles located in the bibliography of the relevant papers were reviewed for significance. All reported studies that used the key words "late effects" as well as those that discussed late effects in their introduction and abstract were included.

Eligibility criteria
Eligible articles met the following criteria: i) written in English, ii) original articles published until September 30, 2019, and iii) containing instruments with questions on late effects.

Study selection
Titles, abstracts, and keywords were screened to confirm reliability with the eligibility criteria. All articles were examined to identify presence of late effects questions within the study questionnaires. Only those relevant articles were included in this review.

Data collection process and data items
The selection process of studies was performed independently by two reviewers (HKC and MP). The author, hypothesis, study design, study sample, sample size, questionnaire, domains of questionnaire, details of late effects, results, conclusions, and advantages and disadvantages of each article are presented in this systematic review.

Study quality assessment criteria
A thorough independent assessment of the methodological quality was also performed by two authors (HK and MP). The NHLBI Quality Assessment of Systematic Reviews and Meta-Analyses tool contains 8 questions with no total score. A modified version of this tool was used in this study which consisted of the following questions: i) Was the study question or objective clearly stated? Ii) Were the subjects comparable within each study? Iii) Were the outcome measures clearly defined, valid, reliable, and implemented consistently across all study participants? Iv) Were details regarding the exposure provided? V) Was the questionnaire relevant to late effects? Vi) Was the length of follow-up adequate? The original NHLBI Quality Assessment of Systematic Reviews and Meta-Analyses consists 8 questions with no total scores, rated as yes or no by two independent reviewers. The best evidence synthesis was performed rating qualities of the studies using dichotomous criteria (i.e., 0 = no, and 1 = yes).

Results
An exhaustive review of the literature revealed a total of 576 publications in PubMed and 628 in Web of Science. The reference list of related review articles resulted in an additional 260 relevant records. One-hundred eighty seven duplicates were deleted using EndNote bibliographic software. Upon elimination of duplicates, a thorough screening of titles, abstracts, and keywords was implemented yielding 1277 articles (Fig 1). A total of 125 full text articles were then reviewed against the eligibility criteria. One hundred and fourteen articles were eliminated because they: i) did not contain a questionnaire, ii) were one of multiple publications using the same questionnaire and data already captured, and iii) did not study late effects among cancer survivors. In the end, eleven studies fulfilled the requirements of the eligibility criteria [14][15][16][17][18][19][20][21][22][23][24]. All the studies included in this systematic review are summarized and presented in Table 1.

Study quality assessment
Study quality assessment for each study was measured on a scale of zero to six (i.e., 0 = poor quality; 6 = highest quality. Only one study was rated with a very good quality assessment score of 5 [23]. Six studies were rated as good with a score of 4 [15,[18][19][20][21][22]. Four studies were  rated as poor-quality assessment with a score of �2 [14,16,17,24]. The details of the study design and study quality are presented in Table 2. The scoring system was devised by the authors. The table is organized by total quality score from lowest to highest. There were five cross sectional studies, two studies combining cross-sectional and longitudinal designs, one randomized clinical study, one longitudinal study, one retrospective cohort, and one systematic review. Sample sizes ranged from small (range 16-30), to mid-size, (range 337-577) to large (range 1,668-14,054).
All of the research questions and objectives were clearly stated. Several studies [15,16,[19][20][21][22]24] intentionally included multiple cancer diagnoses to appeal to a broader audience. Hence, it was difficult to compare the cancer survivors within and across studies. The remaining studies classified late effects as treatment effects, concerns, long-term effects, and unmet needs. The instruments that best captured late effects were created by Rocque [23], St. Jude's [15], Rechis [22], Lavoie [20], Geller [18], and Ness [21]. Rocque's study encompassed diagnosis, treatment, chronic side effects, and late side effects, but only among breast cancer survivors [23]. The major limitation of this study arose from "the dictionary of late or chronic side effects" developed by breast oncology specialists at the U of Wisconsin Carbone Cancer Center. Chronic side effects from surgical treatments (e.g., lumpectomy, mastectomy, sentinel lymph node biopsy) were presented [23]; however, there were no late effects revealed. Additionally, late effects related to radiation and endocrine therapy were not disclosed, apart from tamoxifen being linked to uterine cancer. However, late effects as a result of chemotherapeutic drugs were specified. For example, Epirubicin listed infertility or menopause, heart problems, and leukemia; carboplatin was grouped with infertility or menopause and neuropathy; and doxorubicin was put together with infertility or menopause, heart problems, and leukemia. The St. Jude's Childhood Cancer Survivor Study (CCSS) [15] is a component of the Long-Term Follow-Up Study. The CCSS is a retrospective cohort of 35,923 childhood cancer survivors diagnosed between 1970 and 1999 [25]. The current review determined that the St. Jude's questionnaire was one of the higher quality questionnaires because it queried specific questions on late effects since May 1995 on medical conditions of the heart or lungs (e.g., blood clot), hearing/vision/nervous system, fatigue, sexual health, and cancer recurrences among pediatric cancer survivors. Additionally, there was one question cited in the breast cancer questionnaire, which pertained to the survivor developing a health problem several years later, which was related to their previous cancer treatment. Nevertheless, there were several organs omitted (e.g., gastrointestinal, genitourinary, central nervous system, musculoskeletal, thyroid, lymphatic, immune) from the CCSS questionnaire, the questions on late effects were cursory (i.e., present or absent), and the study sample only included childhood cancer survivors. Furthermore, emphasis was placed on long-term complications for one type of pediatric cancer (e.g., pediatric astrocytoma) [26]. The authors state that monitoring of late effects is important however, this was omitted from the results. Hence, the Childhood Cancer Survivor study did not provide a comprehensive list of questions pertaining to late effects of all organs; rather it only focused on four conditions within pediatric cancer survivors.
The LIVESTRONG survey by Rechis was created to better understand experiences after completion of cancer treatment [22]. It provided information on 1,719 survivors' physical concerns including 1) Concentration, 2) Energy and rest, 3) Heart problems, 4) Infertility, 5) Lungs and breathing (short of breath), 6) Lymphedema, 7) Neuropathy, 8) Oral (tooth decay), 9) Pain, 10) Sexual functioning and satisfaction, 11) Thyroid function, 12) Urinary incontinence, and 13) Hearing loss. These concerns appeared to be synonymous with commonly occurring late effects, particularly because the participants had the opportunity to seek help for their complaints. The survey also ascertained emotional concerns including: 1. Faith and spirituality, 2. Fear of recurrence, 3. Personal appearance, 4. Personal and social relationships, 5. Sadness and depression. The final category for the LIVESTRONG survey consisted of Practical concerns including 1) Debt, 2) Insurance, 3) Employment issues, and 4) School issues. LIVES-TRONG contained the premier questionnaire because it concentrated on numerous medical, psychosocial (emotional), and practical concerns. It also captured information regarding whether respondents received care for their complaints. However, the information collected on concerns was tabulated by participants selecting all pertinent options that apply (e.g., I have had trouble with my heart/lungs, I have been told by a doctor that I have heart problems, lung problems) with no additional details. Additionally, this study was cross-sectional.
The main objective of Lavoie's study was to pilot test a Web-based cancer survivor (CS-WEBS) needs assessment survey [20]. The CS-WEBS survey was more focused on measuring a comprehensive list of needs of cancer treatment than other scored instruments. "Needs" appeared to be synonymous with symptoms that were present but are not necessarily bothersome. For example, "I have numbness and tingling in my hands and feet," "I have this problem," "I have this problem but it is not bothering me," "I don't have this problem," and "I don't have this problem because I am getting help for it." The participants were predominantly white (98%) women (70%). One of the major limitations of this study was the lack of characterization of symptoms (e.g., tingling, anxious, depressed, tired, chest pain) of late effects.
Both Geller [18] and Ness [21] evaluated concerns rather than actual late effects experiences of cancer survivors. Ness's cross-sectional survey [21] was used to assess the most prevalent physical, social, emotional, and spiritual concerns of cancer survivors. The sample consisted of survivors actively seeking cancer-related information and documented whether they were concerned about specific symptoms [21]. However, the study did not assess whether survivors actually experienced specific late effects.
Geller's large community-based cancer survivor study, which used a modified version of the CaSun instrument [27], assessed 53 specific needs of cancer survivors [18]. This study team also determined whether the identified needs were met. The researchers did not assess which specific late effects the survivors were afflicted with at the time of the survey.
In contrast, Smith [24], and the Cancer Experience Registry Index study [14] did not include specific questions on late effects. Ganz [17] evaluated quality of life, nineteen comorbid conditions, and 42 everyday problems/symptoms; however, there were also no questions on late effects.

Discussion
This systematic review summarizes the existence of all cancer survivorship instruments that assess the development, progression, and treatment of late effects among cancer survivors.
Although this systematic review used the best available evidence, there were several limitations. The exposure definitions and outcome measurements were defined differently across the individual studies. The studies all used different self-report instruments. There was no validation with medical records to confirm whether the survivor reported the late effect to their oncologist or primary care physician. The instruments were only distributed once, and thus, the majority of studies were cross-sectional, except for the CCSS study [15], as well as Curcio [16], and Rocque's studies [23].
None of the instruments were customized by the type of cancer, age at treatment, and choice of treatment that would determine which specific late effects may be relevant. Furthermore, there was no information on timing, frequency, and duration of late effects in any of the questionnaires.
Other health problems of the cancer survivors were not taken into account, apart from CCSS [15] and LIVESTRONG [22] studies. Finally, there was no information on minority and medically underserved populations or non-English speaking cancer survivors.
One main overarching problem was the overlap of terms within all eleven papers, including treatment effects, late effects, long-term effects, unmet needs, and concerns (implying that the patient was worried but may not have experienced the late effect). Hence, it is evident that there is a lack of agreement about these important terms within the literature. A consensus should be developed so that each concept can be appropriately studied. Researchers should begin to tease out the differences between late effects and unmet needs. For example, the most frequently reported unmet needs of Australian cancer survivors were for help with psychosocial issues, including fear of cancer recurrence, uncertainty about the future, worry about partners, friends, and families, help to reduce stress, and sexual changes [28]. All of these factors relate to cancer survivors expressing a demand perceived by the patient that was not adequately met by the health care system [29]. A survivors' opinion/impression or viewpoint is different from them actually experiencing physical or mental late effects. Hence, in the future, the presence of late effects and/or unmet needs should be presented separately, and categorized based on the cancer diagnosis, and time since completion of treatment, since both could vary considerably based on the initial years (1-2 years) compared to later years (5 or more years).
Finally, there is currently no comprehensive questionnaire that captures all of the relevant details of late effects (e.g., type, frequency, onset, duration) across the cancer survivorship continuum nor that tracks the evolution, chain of events, or interrelatedness of multiple late effects. Because these details are missing it is difficult to differentiate between late and longterm effects as well as identify, diagnose, manage and ultimately prevent late effects in cancer survivors.