National Institutes of Health Hematopoietic Cell Transplantation Late Effects Initiative: The Patient-Centered Outcomes Working Group Report

Abstract

In 2015 the National Institutes of Health (NIH) convened six working groups to address the research needs and best practices for patient-centered late effects of hematopoietic stem cell transplantation (HCT) survivors. The Patient-Centered Outcomes Working Group, charged with summarizing the HRQOL evidence base, used a scoping review approach to efficiently survey the large body of literature in adult and pediatric HCT survivors over 1 year after transplantation. The goals of this paper are to 1) summarize the current literature describing patient-centered outcomes (PCO) in survivors including the various dimensions of healthrelated quality of life affected by HCT and describe interventions tested to improve these outcomes; 2) highlight areas with sufficient evidence allowing for integration into standard practice; 3) address methodological issues that restrict progress in this field; 4) identify major gaps to guide future research; and 5) specify priority research recommendations. PCOs were summarized within physical, psychological, social and environmental domains, as well as for adherence to treatment, and health behaviors. Interventions to improve outcomes were evaluated for evidence of efficacy, although few interventions have been tested in long-term HCT survivors. Methodologic issues defined included lack of consistency in the selection of PCO measures, along with the absence of a standard for timing, frequency, and mode of administration. Recommendations for HCT survivorship care included: integration of annual screening of PCOs; use of evidence based practice guidelines; and provision of treatment summaries and survivorship care plans after HCT. Three priority research recommendations included: 1) Design and test risk-targeted interventions with dose intensity modulation matching the needs of HCT survivors with priority domains including sexual dysfunction, fatigue, sleep disruption, non-adherence to medications & recommended health care, health behaviors including physical inactivity and healthy eating, and psychological dysfunction, with particular consideration of novel technologies to reach HCT survivors distant from their transplant centers, 2) Design a consensus based methodologic framework for outcomes evaluation, and 3) Evaluate and compare existing practices for integrating PCOs screening and interventions across HCT survivorship programs.

INTRODUCTION

Advances in conditioning regimens, supportive care practices, and expanded donor sources and cellular therapies have led to expanded access and indications, as well as improved survival rates for HCT recipients (1). As new populations are exposed to HCT as a potentially curative therapy and survival expectancy lengthens, there is a growing recognition of the adverse impact of treatment on long-term survivorship, including health-related quality of life (HRQOL).

HRQOL is a subjective, multidimensional concept. Although a broad range of life domains can be represented, HRQOL almost universally includes psychological, social, physical (including symptoms), and contextual or environmental aspects of health and illness (2, 3). As a patient-reported outcome (PRO), HRQOL represents an intersection of the individual, their culture, values, and preferences relevant to health. As the field has evolved, examining outcomes beyond HRQOL that are patient-centered, though not always patient-reported, has provided a more comprehensive perspective on the impact of illness on an individual’s life (4). Examples of such outcomes are adherence to medication (5) and health behaviors such as tobacco use or physical activity (6), which may affect treatment success as well as improve HRQOL.

The impact of HCT on the global HRQOL of survivors has been described in adults (1, 7–13) and children (14–19). Overall, the majority of survivors experience improvement over time, with return to pre-HCT levels of HRQOL by 1 year (10, 12, 20–22). Risk factors for poor HRQOL following transplant include poorer pre-HCT physical health, younger age, depression, female gender, low education level, low social support, physical symptoms, unrelated donor (for childhood HCT), and active chronic graft-versus-host-disease (cGVHD) (22–28). Despite the overall positive perception of global HRQOL recovery, many survivors report ongoing residual deficits such as psychological distress, fatigue, sexual dysfunction, cognitive dysfunction, financial toxicity, and cGVHD consequences (22, 24, 26, 29).

Our knowledge regarding the impact of HCT on HRQOL and other patient-centered outcomes is largely derived from studies of the early effects post-HCT, rather than long-term or late effects occurring in the years after HCT, and is often limited to those with hematologic malignancies. Since the majority of HCT patients leave the transplant center and return to their primary providers for long-term care, residual problems can be difficult to follow. This dispersion of survivors also disrupts integrating appropriate services into practice due to challenges in dissemination and implementation of evidence-based care. Moreover, the population of survivors with nonmalignant conditions only recently reached a critical mass whereby the impact of HCT can be assessed in that set of survivors. The potential chronicity of a non-malignant condition, the absence of prior cytotoxic therapy, and mild or no graft-versus-host disease may contribute to a different experience compared to those with malignant diseases.

In 2015 the national Cancer Institute and National Heart, Lung, and Blood Institute convened six working groups to address the research needs and best practices for long-term and late effects of hematopoietic stem cell transplantation (HCT) survivors (Battiwalla et al 2017 BBMT). The Patient-Centered Outcomes Working Group was charged with summarizing the HRQOL evidence base and its integration into practice, as well as identifying key research gaps and opportunities within the HCT survivorship community. The goals of this paper are to: 1) summarize the current literature describing patient-centered outcomes in survivors over one year post-transplant, including the various dimensions of HRQOL affected by HCT and describe interventions tested to improve these outcomes; 2) highlight areas with sufficient evidence allowing for integration into standard practice; 3) address methodological issues that restrict progress in this field; 4) identify major gaps to guide future research; and 5) specify priority research recommendations.

METHODS

A scoping review approach was applied to efficiently survey the large body of literature in adult and pediatric HCT survivors over one year post-transplant (30). Given the breadth of the topic, reviews rather than single studies were preferred. A search of the PubMed database was completed using the following terms: quality of life (HRQOL), psychological [depression, suicide, anxiety, distress, uncertainty, worry, post-traumatic stress disorder (PTSD), anger, emotional, mood, and fear of recurrence]; social (relationships, isolation, caregiver burden, development, learning, family conflict, dyadic adjustment, support, parents, caregiver, peer); symptoms (pain, neuropathy, fatigue, musculoskeletal, sleep, nausea); function (physical, social, sexual, cognitive, role, work, school); energy (muscle, cardiopulmonary capacity); behaviors (physical activity, adherence, tobacco use, stress management, nutrition); spiritual (existential, purpose, meaning, religion, post-traumatic growth, benefit finding, resilience); financial toxicity; and interventions [pharmacological, non-pharmacological (psycho-educational, cognitive behavioral, stress management, self-management, mindfulness, activity based, rehabilitation, exercise, yoga, social, web/internet-based, media, complementary and alternative)]. Filters were applied to limit the search to review papers, written in English, and published between 1 January, 1990 and 1 November, 2015. Quantitative and qualitative studies including adults or pediatric patients undergoing allogeneic or autologous HCT (stem cell transplant, blood and marrow, BMT, bone marrow) for any indication (malignant or non-malignant) were included. Secondary search strategies were applied by the working group when review papers were old or not found on a topic. These secondary strategies restricted the literature to survivors one or more years post-transplant and expanded the literature to individual studies or studies in other patient populations, e.g. general oncology when necessary.

FINDINGS

The patient-centered outcomes in this report are presented according to the dimensions of HRQOL (physical, psychological, social, and environmental) and the additional patient-centered outcomes of medical adherence and health behaviors. The estimated prevalence of common outcomes, known risk factors, and comments are summarized in Table 1.

Table 1.

Common patient-centered outcomes in HCT survivors: Estimated prevalence and documented risk factors

Outcome	Prevalence*	Risk Factors	Comments
Global HRQOL	The majority return to pre-HCT levels within 1st year	Sociodemographic: younger age, female, low education level, low social support Clinical: poor pre-HCT physical health, active cGVHD, physical symptoms, depression	Well described for adult hematologic malignancy survivors, not as well for pediatric survivors or those transplanted for non-malignancies
Physical
Pain	21–25%	Sociodemographic: unknown Clinical: cGVHD, avascular necrosis, arthralgia, myalgia, muscle cramps	Syndromes not well defined but largely musculoskeletal
Fatigue	35–42%	Sociodemographic: female, younger age Clinical: chronic pain, cGVHD, inactivity	Widely measured, mechanisms unclear
Sleep Disturbance	14–51%	Sociodemographic: older age, female, divorced, unemployed Clinical: autologous HCT, depression, distress	Poorly defined, especially for pediatric survivors
Sexual dysfunction	Males: 6–46% Females: 33–80%	Sociodemographic: older age at HCT, female Clinical: depression, cGVHD, TBI (for men)	Clearly defined as prevalent though rates vary and risk factors not fully defined
Physical dysfunction	25% or greater	Sociodemographic: none consistent Clinical: active cGVHD, depression	Well defined as recovering overall by one year post-HCT for most survivors, not well defined for pediatric survivors
Psychological
Emotional Distress	22–43%	Sociodemographic: lower income, higher education, low social support; Clinical: physical limitations, active cGVHD, exposure to prednisone, neurocognitive problems, chronic pain, more aggressive disease or treatment, greater perceived impact of treatment	Undefined in pediatric survivors
Depression	12–30%	Sociodemographic: younger age, female, poor social support, family history of mental illness†, lower socioeconomic status† Clinical: history of relapse, chronic pain, active cGVHD, transplant type†	Rates in adults vary widely because of methodologic variability
Post-traumatic stress	5–13% for full diagnosis	Sociodemographic: unclear Clinical: unclear	Few data are available from full diagnostic screening for PTSD in long-term survivors; many more have symptoms that do not meet diagnostic criteria
Cognitive dysfunction	10–40% if measured with neuropsychological testing 40–60% if self- reported	Sociodemographic: lower social economic status, age under 5 years†, female† Clinical: pre-HCT memory loss, cranial irradiation, intrathecal chemotherapy, active cGHVD, TBI†, immunosuppressive therapy, history of delirium, length of hospital stay, accumulation of risk factors, unrelated donor HCT†, cognitive decline in the first year after HCT†	Prevalence and predictors depend on methodology and time points assessed, variability in findings leaves many questions
Social
Return to work	15–40% do not return to previous employment	Sociodemographic: lower income, female Clinical: cGVHD, physical impairment, more symptoms, more hospitalizations	Fairly well defined
Environmental
Financial burden	Up to 73%	Sociodemographic: younger age, low income, health insurance not protective of long-term impact Clinical: poorer physical and mental functioning	Few studies, requires prospective, longitudinal investigation
Adherence to Treatment
Medication non- adherence	Adults up to 66% Adolescents: 40– 91%	Sociodemographic: adolescents† Clinical: unknown	Minimal data on prevalence, risk factors & patterns of non- adherence unknown; impact on survival and other outcomes requires investigation
Non-adherence to health care recommendations	50% adhere to <75% of guidelines	Sociodemographic: younger age, male, non-white, concerns about medical costs Clinical: autologous HCT, lower physical functioning, absence of cGVHD, longer time since HCT, poor knowledge of recommended tests	Few studies, but consistent prevalence and risk factors; needs clarification of impact of non-adherence
Health Behaviors
Risk behaviors	7–15% Tobacco use (≤ norms) 10%; excess alcohol (use < norms) 10% use of illicit drugs (comparable to norms) 5% lack of sun protection (< norms)	Sociodemographic: adolescents and younger aged adults, less education, lower income, lack of health insurance Clinical: depression, family conflict	Few studies so rates and risk factors are uncertain
Physical inactivity	27–85%	Sociodemographic: unknown Clinical: unknown	Inactivity is well documented, highly prevalent, risk factors undefined
Unhealthy diet	95%	Sociodemographic: unknown Clinical: unknown	Few studies so rates and risk factors are uncertain

cGVHD indicates chronic graft-versus-host disease; HCT, hematopoi etic stem cell transplant ation; TBI, total body irradiation, PTSD, post-traumatic stress disorder

^*As defined from available research for survivors > 1 year following HCT

^†Pediatrics

Physical

Physical symptoms are a major contributor to impairments in HRQOL, including physical and psychological function in HCT survivors (23, 31). Prevalent symptoms reported include pain (particularly musculoskeletal), lack of energy/fatigue, and difficulty with sleep (23). Although a variety of symptoms have been reported, few have been studied in depth, with little attention to patterns of symptom clusters that are common in cancer patients (32).

Pain

Pain syndromes have been documented in approximately a quarter of adult HCT survivors (13, 33–36). Although pain syndromes are not well described, evidence suggests that most pain is correlated with active cGVHD, which is known to be associated with several chronic pain syndromes (24, 37–40) as well as avascular necrosis, arthralgias, myalgias, and muscle cramps (13, 41). Adult survivors of childhood HCT report significantly worse bodily pain than healthy matched controls, with 21% reporting persistent pain (38, 42).

Fatigue

Fatigue is among the most common concerns of HCT survivors (7). Comparisons with age- and gender-matched controls and general population norms have consistently demonstrated greater fatigue in both allogeneic and autologous adult HCT survivors (26, 43). One study found that 35% of allogeneic and autologous patients were experiencing severe fatigue at one year post-transplant (44). Another recent study found moderate to severe fatigue in 42% of a cross-sectional cohort of survivors who were a mean of 13 years post-transplant (33). Factors associated with fatigue included younger age, and chronic pain for all HCT recipients and more specifically female gender and cGVHD for allogeneic recipients only (33). Our understanding of fatigue in long-term pediatric survivors is in its infancy (45).

Sleep

Sleep problems have been reported by 14% to 51% of HCT survivors (46), but findings are mixed regarding whether adult HCT survivors report more sleep problems compared to population norms and non-cancer comparison groups. These rates of insomnia do not appear to change over time (46). Preliminary evidence suggests that risk factors for sleep problems include older age, female gender, divorce, unemployment, depression, distress and receipt of an autologous transplant (46). The prevalence of sleep problems and fatigue in long-term childhood HCT survivors is unknown. As with fatigue, sleep difficulties in long-term pediatric survivors has not been much studied (45). However, reduced sleep quality and quantity contribute to poor HRQOL in pediatric oncology patients through their adulthood (47, 48).

Sexual Function

The few longitudinal studies of sexual health after HCT consistently indicate that sexual dysfunction and sexual dissatisfaction are common (49). The incidence of long-term sexual dysfunction in male transplant survivors ranges from 6–46% and in female survivors from 33–80% (29, 50–52). Sexual dysfunction may be permanent for many women and some men (50). Risk factors include pre-transplant depression, female gender (29, 50, 53), older age, active cGVHD and in men, the use of TBI (53). Few studies have examined the psychosexual impact of reproductive risks, infertility and sexuality in pediatric or adult survivors of HCT (54, 55).

Physical functioning

Deficits in mobility, endurance and other aspects of physical function may adversely impact a survivors’ ability to carry out activities of daily living. Patients report the lowest levels of physical functioning during hospitalization and scores generally improve to pre-transplant levels within a year post-transplant (22, 56, 57). During late survivorship, active cGVHD is the strongest predictor of poor physical functioning (57–59) along with depression (22). Similarly, adult survivors of childhood HCT assessed a mean of 16 years post-transplant, report more physical limitations than age and sex-matched controls (42). Childhood leukemia survivors who are post-transplant have not differed from those treated with chemotherapy alone (60).

Psychological

Emotional Distress

Emotional distress in HCT survivors often does not reach the level of clinical depression, anxiety, or post-traumatic stress disorder, but may nevertheless be disruptive to HRQOL. Clinically meaningful distress (particularly difficulty living with uncertainty, fear of recurrence, loneliness, memory concern, and somatic preoccupation), is elevated in 22–43% of long-term survivors (61, 62). Risk factors for persistent distress do not vary consistently by gender, age, or time since transplant, but are greater for those exposed to prednisone or with active cGVHD, lower household income, more physical limitations, a higher education, or lower social support (22, 61–63).

Depression

Depressive symptomatology in long-term HCT survivors ranges from 12% at 10 years post-transplant (13) to 30% in a cross-sectional sample (43). Two studies have found levels similar to age- and sex-matched norms in autologous survivors after 3–5 years or allogeneic survivors10 years after HCT (13, 26), while another study found that levels of depressive symptomatology were significantly higher in HCT recipients compared to a matched, healthy comparison group (43). Regarding risk factors, some demographic variables (e.g., younger age, female gender) and clinical variables (e.g., relapse) associated with depression in HCT survivors are fixed (10, 33) while others such as lower levels of pre-transplant social support or current chronic pain or active cGVHD may be modifiable (22, 33).

Post-Traumatic Stress

Few studies have examined prevalence and risk factors for post-traumatic stress disorder (PTSD) further than one year after HCT (61, 64). While in adults PTSD as a clinical diagnosis may persist for 5–13% (61), post-traumatic stress symptoms are likely underdiagnosed, and overlooked or diagnosed as depression or anxiety after HCT. Risk factors specific to HCT long-term survivors are not yet clear.

Psychological Functioning in Pediatric Survivors

In pediatric survivors, studies of psychosocial outcomes post-transplant are sparse and mixed, but suggest that a subset of survivors are at risk for the development of long-term emotional problems, including anxiety, depression, and PTSD (42, 65–72). Risk factors associated with psychological distress include female gender and lower socioeconomic status, like adult survivors, in addition to younger age at diagnosis, physical limitations (especially perceived poor physical health and chronic pain) (73), neurocognitive problems and family history of mental illness (42, 74).

Neurocognitive Functioning

Cognitive alterations detected with neuropsychological testing have been reported in 40–60% of adult survivors while a recent meta-analysis reports no significant changes in cognitive functioning from pre- to post-transplant (58, 59, 75). Most adult survivors recover cognitive function within one year but mild impairments remain evident 5 years post-transplant for some survivors (59, 75, 76). Patient-reported problems are typically discrepant with and higher than detected with objective neuropsychological test results. Identified risk factors for cognitive dysfunction include pre-transplant memory problems, the use of total body irradiation (TBI), history of cranial irradiation, intrathecal chemotherapy, cGVHD, immunosuppressive therapy, a history of delirium during the acute transplant phase and lower socioeconomic status (59, 75–78). Length of hospital stay, perhaps reflective of greater complications, seems to confer a risk for sustained cognitive deficits after HCT (78).

Consensus has not been reached about the long-term neurocognitive sequelae of HCT in children. Most studies report small to no declines in global IQ and academic achievement among pediatric survivors (66, 79, 80), although subsets of children are at risk for significant decrements in IQ (81, 82) when assessed up to five years post-transplant (83). Like in adults, greater risk is associated with TBI or cranial radiation, intrathecal methotrexate, cGVHD and low socio-economic status, but also conferred by age less than five when transplanted, female gender, and a cognitive decline in the first year post-transplant (84, 85).

Studies examining micro-level cognitive domains in pediatric cohorts reveal subtle impairment in complex psychomotor skills, visual-motor integration, attention and concentration, visual-spatial processing and memory, executive functioning, and nonverbal reasoning (82, 86, 87). These declines in micro-level processes could possibly interfere with learning and the normal progress of cognitive development and consequently academic achievement in the developing brain beyond five years post-transplant (83). The limited data available on academic accomplishments post-transplant suggest that poorer performance is associated with young age at diagnosis, neurotoxic treatments, health status (e.g. cGVHD, fatigue, depression) and social or educational status of the parents (88, 89).

Positive Psychological Outcomes

Few studies have examined post-traumatic growth (PTG) or resilience in HCT survivors, and even fewer have considered spiritual or other existential evaluations. PTG is a positive psychological change experienced as a result of exposure to traumatic or highly challenging events, and is seen in HCT survivors who report emotional growth beyond their pre-transplant levels (90, 91). PTG is associated with enhanced appreciation of life, improved relationships, existential growth, increased self-confidence and self-esteem (90, 92). Factors that predict PTG include female gender, younger age, and higher levels of social support, spirituality, and resilience although the impact of these variables is not consistent across studies (63, 90, 92–94). Possible mediators of PTG include social support, proactive counseling by the health care team, and cognitive processing of the transplant experience (90).

Social

Social support

HCT recipients often encounter relationship, social role and communication changes following HCT (95). Living with a partner versus living alone has been associated with more positive outcomes for survivors, but spousal caregivers experience higher levels of anxiety and depression and increased social isolation at least through two years after HCT compared to HCT recipients (96–98). There is limited evidence that social support influences survival among patients following HCT (99).

In pediatric HCT survivors, the role of social support is also poorly understood. Social support is impacted by fatigue as well as separation from peers and age appropriate activities (45, 100). If this pattern of social withdrawal and isolation were to continue unchecked, skills central to making and keeping friends and social networks could potentially be jeopardized (101).

Work

Return to work is an important indicator of financial and role stability for survivors (102). Longitudinal studies indicate that the majority of survivors return to full time work in the first year post-transplant, with 60–85% returning by five years (21, 103–106); additional survivors return to part time work. Women and those with poorer physical health, more symptoms, and more hospitalizations are less likely to return to work (103–105). The income loss resulting from the inability to return to work has been associated with poor adherence to preventive care practices as well as higher rates of cGVHD (107). Interestingly, one study of 20-year survivors of HCT for β-thalassemia found no difference in employment for these adult survivors transplanted as children when compared with the general population (108).

Caregivers and Related Donors

The long-term effects of HCT on informal caregivers, the majority of whom are family members, have not been well studied. The majority of the studies exploring the effects of HCT evaluate the early period post-transplant (109–111). The individual studies that included caregivers at or beyond one-year suggest that caregiver distress generally improves over time, although caregivers still report long-term role adjustments and ongoing demands (112). Caregivers who are female, working full-time, and those who experienced more distress pre-transplant, are more likely to report problems with relationship adjustment and higher levels of distress (97, 113, 114). Emerging evidence suggests long-term effects can also include an adverse impact on caregiver health behaviors and physiological stress responses (115–117).

Parents may also report persistent symptoms years later, particularly when there are clinical complications (118, 119), especially when parents perceive their child to be much more vulnerable compared with healthy children (118, 120, 121). PTSD in parents has been identified post-transplant (121), with mothers reporting significantly worse emotional well-being compared with population norms (122).

Only a handful of studies have addressed HRQOL in sibling donors (123–126), each suggesting a critical need to better understand the donation-related experiences. A recent study in pediatrics found approximately 20% of donors had very poor HRQOL, significantly lower than that of norms at pre-donation and 4 weeks post-donation (127).

Environmental

Financial toxicity

The financial impact of HCT on recipients and their caregivers can be substantial (102, 128). One year or more after HCT up to 73% of survivors experience significant financial hardship in spite of health insurance benefits, with 3% declaring bankruptcy (102). During treatment, studies indicate a wide variability in out-of-pocket expenses for caregivers, depending in part on whether lost wages are included (129, 130). Greater financial and employment difficulties are associated with poorer HRQOL (131) and higher levels of perceived stress (132, 133). Low-income families report disproportionate transplantation-related income losses, while low household income is an independent predictor of worse HRQOL post-transplant after controlling for relevant treatment-related exposures (132). Younger age, poorer physical and mental functioning, and having to relocate closer to the transplant center also increase the risk for financial burden.

Adherence to Treatment

Medication adherence, critical to the long-term success of the HCT, has not been widely examined in survivors. The sole study identified reported that among adult survivors taking immunosuppressants almost two-thirds were non-adherent (134). Medication adherence post pediatric HSCT is also poorly understood. In the only study of adolescent HSCT survivors (n=6), medication adherence rates per month were a mean of 73% (range 40–91%), but declined over 9 months (135). Potential serious implications of non-adherence are epitomized in a large ALL sample in which adherence rates of < 95% were associated with higher risk of relapse, after other confounding variables were controlled (136).

Adherence to recommended guidelines for health care practices in HCT long-term survivors has received sufficient investigation to indicate that important deficits exist for half of long-term survivors (107, 137). Predictors of non-adherence include lack of knowledge of recommended surveillance tests, male gender, non-white race, younger age, autologous HCT, absence of cGVHD, poorer physical functioning, longer time since HCT, and concern about medical costs.

Health Behaviors

The literature on risky behaviors such as the use of tobacco and alcohol, and sun exposure in HCT survivors is limited. In general, adult HCT survivors’ participation in unhealthy behaviors seems equal to or lower than that of other populations, with predictors of unhealthy behaviors including younger age, less education, and lack of health insurance (138). Smoking rates are similar to matched controls (139), but significantly lower than that of siblings (7% vs. 13%) (138, 139) and population-based controls (134). One study reported that 15% of HCT survivors restart smoking against the advice of their physicians (106). Alcohol use over the recommended 0–1 drinks per day is lower in survivors than matched controls, and lower than that of siblings (139), and weekly consumption is lower than population-based controls (134). Data on the use of sun protection are limited but reveal that survivors use sun protection more regularly than population-based controls (134).

The rates of risky behaviors in survivors of childhood HCT are not well documented although the evidence for childhood cancer survivors (CCS) more broadly is clearer. The rate of lifetime use of tobacco in adolescent CCS is comparable to that of healthy siblings (140, 141) although in another study of adult survivors of childhood cancer, smoking was found to be more common than among healthy controls (142). The rate of smoking initiation after HCT is higher in survivors who are younger at HCT (12.5 years vs 27.4 years) (138). In adolescent CCS, rate of alcohol use is comparable to that of healthy siblings and the general population (40, 142), and the rate of use for marijuana, heroin, cocaine, and methamphetamines is approximately 10% (40). Older children are noted to have increased sun exposure and decreased use of sun protection (143).

Health promoting behaviors such as physical activity and healthy eating have been the focus of several investigations. Adult survivors are more likely to be physically inactive compared to the general population (26.8% vs 12.5%) (134), and to exercise for shorter durations when active, but have a similar BMI compared to matched controls (139). Only 5% of survivors report eating a healthy diet that is low in fat and high in fruits and vegetables (134, 139). Recent data underscore the importance of physical activity, diet and weight control given the high rates of cardiovascular disease in long-term HCT survivors (144). Risk factors for poorer health behaviors are not well defined for adult survivors.

The literature focusing on healthy lifestyle practices in pediatric HCT survivors’ is sparse. Predictors of problematic health behaviors in adolescent survivors include older age at diagnosis, lower education or household income, and depression in the context of family conflict (138, 140). In contrast, predictors of problematic health behaviors in adult survivors of childhood HCT are younger age, lower education and lack of health insurance (138). Compared to healthy controls, pediatric allogeneic transplant survivors have lower lean body mass and higher fat mass (145). Further, being overweight pre-transplant is associated with poorer survival post-transplant (146). The limited data on exercise in pediatric survivors indicate that self-reported daily physical activity in long-term childhood HCT recipients is comparable to age-matched norms and somewhat higher in females than males who survived more than two years (147).

Interventions to Improve Outcomes: Evidence for efficacy

Data on successful interventions to improve or stabilize various patient-centered outcomes in HCT survivors beyond the first year are limited. Most intervention studies have focused on improving outcomes for HCT recipients acutely during HCT hospitalization and in the first year after transplant (148–151). One published randomized controlled trial has addressed psychosocial problems (post-traumatic stress, depression and distress) for long-term HCT survivors, delivering 10 sessions of cognitive behavioral treatment by phone (152). With a sample size of 89 randomized participants with elevated symptoms, from 408 screened, the study found sustained reduction in post-traumatic stress symptoms, depressive and distress symptoms over the year after intervention. Another randomized controlled trial examined an internet intervention with or without problem solving treatment (PST) telehealth calls compared with usual care (a 3-arm randomization) in 337 survivors 3–18 years after HCT who had elevated distress, depression and/or fatigue, from 1775 screened (153). Distress improved for those receiving the internet program with PST and marginally improved for those receiving the internet program alone. Both depression and distress improved with the internet-only program for those who used the intervention. Fatigue did not improve with the program. Other published reports of interventions with survivors more than one year after HCT have focused on feasibility of an intervention or had extremely small sample sizes (154–156).

Most interventions to improve patient-centered outcomes in HCT recipients have been tested during the early HCT period (Table 2). The most frequently tested interventions in randomized clinical trials include exercise (148, 157–161), cognitive behavioral therapy (149, 152, 162), and mind-body practices with stress management (161, 163–165). Multiple randomized trials of these interventions have shown small to moderate benefits in improving HRQOL, and reducing fatigue and psychological distress in HCT recipients. However, interventions vary substantially in intensity from self-directed programs to highly supervised one-on-one specialist involvement (148, 149, 157, 160, 163). Notably, higher-intensity interventions with substantial interventionist input yielded larger and more significant and enduring benefits (149, 163). It is unknown to what extent successful interventions have been implemented in clinical practice.

Table 2.

Scope of evidence for interventions tested with randomized controlled trials in HCT survivors (< 1year post HCT)

Type of Interventions	Intervention Type/Intensity	Intervention Timing	Outcomes	Limitations
Physical exercise	Mostly aerobic, some combined with strength training Varied from unsupervised programs to intensive 1:1 supervision by a specialist	Multiple studies during hospitalization for HCT Longest follow-up to 6 months	Positive effect on HRQOL, fatigue, physical functioning Moderate effect sizes seen with more intense interventions Earlier interventions have more positive effect on outcomes	Adherence to exercise programs highly variable Attrition is an issue Heterogeneity of exercise programs (strength training, aerobic exercise), unclear what is most beneficial
Mind-body practices and stress management interventions	Relaxation including guided imagery, breathing exercises, progressive muscle relaxation Meditation Hypnosis Partially self-guided interventions	Studies during hospitalization for HCT Short duration of follow-up	Small effects on anxiety, depression, and fatigue One study found small effect on pain during HCT HRQOL improvement seen in more comprehensive coping programs that integrate relaxation exercises Small effect sizes compared to cognitive behavioral therapy	Poor adherence in self-directed interventions Short-follow-up periods Some studies combined mind-body practices with exercise programs without integrating process measures to understand mechanism of improvement in outcomes
Cognitive behavioral therapy	Many studies incorporate CBT methods Some incorporate relaxation methods as part of CBT Mostly face-to-face interventions, with a few utilizing phone or online interventions 4–15 sessions, 20 minutes to 60 minutes	Most target distress during or immediately following HCT; a few target distress, depression, post-traumatic stress in long-term survivors	Some benefits with lower distress (including depression & anxiety), improved emotional functioning, and less post-traumatic stress Five studies showing enduring benefits at 3 & 12 months post-HCT Major CBT components involving emotional processing methods with substantial interventionist input yielded larger and more significant effects	Studies mostly in white participants No integration of process measures Performance, detection, and common factor biases were poorly addressed Most studies did not target high-risk patients or patients with higher level of targeted outcome

HRQOL indicates health-related quality of life; HCT, hematopoietic cell transplantation; CBT, cognitive behavioral therapy

There are limited studies assessing the efficacy of interventions addressing the needs of both pediatric and adult survivors and their family caregivers (110, 166). Two psychological interventions for family caregivers of patients undergoing adult allogeneic HCT show promising improvement in perceived stress early post-transplant: problem-solving training (167) and stress management (116).

Summary and Knowledge Gaps in Outcomes and Intervention Research

Several conclusions can be drawn despite limits in the evidence on patient-centered outcomes in HCT survivors. Years after treatment, survivors have high rates of sexual dysfunction, fatigue, sleep disruption, pain, cognitive dysfunction, emotional distress, financial toxicity, and non-adherence to medications and health care guidelines. They maintain low levels of physical activity and poor diets that can adversely affect their cardiometabolic health (168). Those at higher risk for multiple poor outcomes include: adolescents and young adults, those with lower income, females (except for medical adherence where males have poorer adherence) and survivors with active cGVHD. Psychosocial factors such as depression, emotional distress and low social function are cited as frequently impaired outcomes and are themselves associated with other outcomes which negatively impact HRQOL including fatigue, sleep disruption, sexual dysfunction, cognitive concerns, financial stress, and high risk behaviors.

Other deficit areas remain unclear. Although there is sufficient evidence to conclude that active cGVHD predicts poorer outcomes, it is not yet convincing whether those with resolved cGVHD also have poorer outcomes. Moreover, like many late effects such as subsequent neoplasm or lower bone mineral density (169), more investigation is needed to determine whether cGVHD itself or its treatment is the cause of poor outcomes. Additionally, longitudinal cohort studies will clarify which patient-centered outcomes are persisting long-term complications that begin during HCT and continue into the years of survivorship or are late effects that have onset in the years after HCT, or as seen with PTSD, whether some outcomes may have either onset pattern. This distinction may be important for determining the point of intervention or prevention approaches. Furthermore, evidence in underrepresented populations such as those with non-malignant conditions, ethnic minorities, and pediatric populations is still lacking although these patients are likely to experience unique sets of problems post-transplant. Additionally, studies are needed to assess the impact of family caregiver and dyadic interventions on long-term quality of life for HCT survivors and their loved ones (170). Finally, financial toxicity is a significant concern that needs elaboration (171 et al., 2016).

Evidence is sufficient to warrant development and delivery of interventions that address the specialized needs of HCT survivors for outcomes such as health care non-adherence, and cGVHD symptoms. Less clear is whether interventions that already exist for the management of more common persistent effects of cancer (e.g., distress, depression, sexual dysfunction, fatigue) are as effective in HSCT survivors. At the same time, notable challenges exist that impede progress in designing interventions to improve adverse outcomes that have high prevalence and can influence other aspects of HRQOL including: medication non-adherence, sleep disturbance, cognitive dysfunction (particularly for pediatric HCT survivors), pain syndromes, PTSD, and social dysfunction. The presence of co-occurring symptoms and how this phenomenon might affect interventions is not well elucidated. Guidelines for standard care of cancer survivors across diagnoses and treatments are well established (172) but have not been tested in HCT survivors. Guidelines for managing HCT survivors’ supportive care needs exist but have not been tested for efficacy, nor do they fully address patient-centered outcomes (21, 173, 174).

Other limitations in the evidence restrict recommendations for interventions to improve patient-centered outcomes. First, most studies include both autologous and allogeneic HCT recipients collectively. While the challenges facing autologous and allogeneic HCT recipients are similar during hospitalization for HCT, their quality of life and illness trajectory differ in some areas in the months and years following HCT (12, 175). These distinctions are not well addressed in the published literature. Second, most studies have short-follow-up periods without an assessment of longer-term benefits following the intervention. Third, the most effective interventions are highly resource and time intensive without a clear pathway for affordable, remotely delivered dissemination. Little research has been conducted utilizing novel online technologies, telehealth, and social media, which may offer highly disseminable and potentially sustainable interventions targeting adult and pediatric HCT survivors regardless of their proximity to HCT centers (156, 176). Finally, we have to recognize the wide-spread issue that those who most need services may be most reluctant or have greatest difficulty in accessing services for emotional, social, financial and other reasons (177).

Despite these research gaps, we also acknowledge that different transplant centers provide varying recommendations and offer a variety of services to their patients that could be construed as interventions targeting patient-centered outcomes, e.g., exercise, symptom management, distress screening, routine psychosocial assessment and intervention. A systematic evaluation of practice variation and research to evaluate efficacy and implementation would be valuable (178).

Methodological Issues in Patient-Centered Outcome Assessment

In order to advance research and clinical care related to HRQOL and patient-centered outcomes following HCT, a consistent measurement methodology is needed (179). While long-term and late effects after transplant have been explored in cohort studies (13, 22, 138, 180, 181), most study designs have focused on transplant as a specific entity, despite abundant literature reporting that patients entering transplant are often highly symptomatic with significant HRQOL deficits from prior exposures. Thus, there is a poor integration of the study of persistent and late occurring effects in HCT survivors within the overall context of disease and treatment trajectories. By consequence, there is a significant need to disentangle the contribution of specific pre-transplant factors to long-term HRQOL deficits (e.g., underlying malignant or non-malignant diseases, pre-existing comorbidities, pre-HCT treatment-related effects, sociodemographic vulnerabilities) and the contributions of unique peri- or post-transplant related factors (cGVHD, infection, post-transplant comorbidities, and relapse prevention, e.g. donor cell infusions) to patient-centered long-term and late effect outcomes.

Lack of consistency in the use of measures related to patient-centered outcomes after transplant severely impinges on our ability to consolidate findings across studies, and thereby use them in designing and targeting interventions to those with needs. In a recent review of 114 publications addressing PROs in transplant survivors, while three HRQOL measures were frequently used (FACT-BMT, EORTC QLQ-C30, and SF-36), an additional 25 PRO measures were used in more than one study, and another 50 measures were used in single studies (179). Further, the optimal time points and sampling frequency for these measurements are unclear and must be balanced against considerations of feasibility and respondent burden (179). Also unclear is, the optimal mode of data collection. Electronic methods of PRO collection are feasible and efficient but infrequently used in transplant studies. Whatever the mode, data collection must provide options for patients who do not access or use the internet, and match the options to the preferences of diverse patients (e.g. young and old, high and low literary) to ensure representativeness of the HCT survivor population.

A body of implementation research has developed around the use of PROs in other areas of oncology (182, 183), covering issues such as data display (184, 185), mode of data capture (186), adjudication of patient vs clinician PRO assessment (187, 188), the use of PROs for decision-making (189, 190), and the use of PROs as performance measures (191). This body of research should be applied to the use of patient generated health data in the setting of transplantation. Indeed, there is an increasing regulatory focus on patient-reported measures of physical function for use within clinical trials (192). For example, PROs collected post-transplant could be displayed in a dashboard-type format for viewing by providers and patients in order to provide real-time information about how patients are feeling and functioning. Real-time PROs could provide information on whether therapeutic or supportive care interventions are beneficial.

Stakeholders increasingly agree that survival alone is an insufficient metric to gauge the value of transplantation. Defining a core PRO measure set, measurement time points, and optimal methods of data capture are critical first steps (Table 3), which, given their complexity, are beyond the scope of this review. Patients, caregivers and families as well as patient advocates, nurses, psychologists and social workers should be engaged in this early work and throughout the research process, including interpretation and dissemination to ensure that resulting research is truly patient-centered (193). Given the growing diversity of the transplant patient population, researchers can rely on patient stakeholders to identify and/or develop culturally appropriate PRO measures.

Table 3.

Methodologic Recommendations

Domains for routine screening Symptoms: An inventory that includes common symptoms considering both disease-related and treatment-emergent symptoms such as fatigue, sleep disturbance, and pain Physical: Physical capacity, work, comorbidities Psychological: survivorship-related distress, depression, post-traumatic stress symptoms, neurocognitive function Social: Sexual function, social support, caregiver and donor outcomes (when available) Adherence: Medication, health care and lifestyle behavior recommendations Standard time points Pre-transplant Post-transplant: Day 100, 1 year and annually thereafter Methods Consistent patient and patient-centered outcomes expert stakeholder engagement to drive research Centralized registry for cost-effective, representative data collection Online measurement with hard copy or phone backup options Dashboard reports of summary patient-centered outcomes for providers and patients Establish a registry to support the prospective collection of patient-centered outcomes. Assure inclusion of underrepresented groups such as pediatric survivors, non-malignant HCT recipients, survivors of diverse races and ethnicity

An efficient infrastructure to collect these data from many if not most patients undergoing transplantation would be very beneficial. The Center for International Blood and Marrow Transplant Research (CIBMTR) has recently successfully piloted the collection of PRO data (194) as a demonstration that this type of activity is feasible. To demonstrate the comparative effectiveness of transplantation, for which patient-centered outcomes should be a key factor, it will be necessary to increase data sharing across transplant centers, payers and outcomes registries.

Limitations and Areas for Future Consideration

There are several challenges encountered in providing these recommendations. First, we did not conduct a systematic literature review given the broad scope of topics and research weaknesses in many areas. Our review of the various patient-centered outcome aspects was intended to provide context but not comprehensively review the field. Although the working group attempted to provide clear figures for prevalence and risk factors, the inconsistencies in measurement and follow-up time frames do not allow us to confidently offer indications of the magnitude of relative risks.

Synthesis of Evidence with Priority Recommendations

The findings from this scoping review of the science related to patient-centered long-term and late effects in HCT survivors identify clear areas of knowledge, as well as gaps in our understanding, and point to numerous opportunities to improve research and practice. HCT survivors vary widely in their needs as a consequence of their underlying disease, HCT conditioning and prior treatments, as well as routine post-transplant management (e.g., immunosuppression) and cGVHD. In addition, survivors bring their individual coping style, social, cultural and environmental context, financial health, and many other personal factors to their HCT survivorship. Fortunately, pre-existing guidelines (Table 4) are available and serve as the foundation for our recommendations for immediate implementation as standard practice for HCT survivorship care (Table 5). Availability of HCT best practice recommendations for health surveillance provides an opportunity for extending knowledge and measuring adherence to these recommendations by both providers and survivors (174). These adult focused guidelines, along with pediatric guidelines (74, 142), can be used to structure programs for HCT survivors, while considering whether adaptations unique for HCT survivors are needed.

Table 4.

Guidelines for HCT survivorship standard care

Commission on Cancer: Cancer Program Standards: Ensuring Patient-Centered Care (195) ASCO Key Components of Survivorship Care (196) NCCN Guidelines for Survivorship Care (74, 197) American Cancer Society Prevention, Early Detection, and Survivorship Guidelines (198) Recommended Screening and Preventive Practices for Long-term Survivors after HCT (174) Psychosocial Follow-up in Survivorship as a Standard of Care (142) Late Effects Surveillance Recommendations Among Survivors of Childhood HCT (74)

Table 5.

Recommendation for HCT survivorship standard care

Integrate routine screening* for common and bothersome patient-centered outcomes at least annually starting one 1 year post-transplant. Increase the frequency of screening to every six months if there is a change in health status (e.g. relapse) or chronic late effect such as cGVHD. Integrate evidence based practice guidelines** for treating bothersome patient-centered outcomes that are identified during routine screening. Integrate treatment summaries and survivorship care plans into HCT survivorship care with special attention to critical transitions such as pediatric to adult care and transplant center care to community providers.

cGVHD indicates chronic graft-versus-host disease

^*Screening – Identifying individuals with unrecognized conditions(s) who should receive a comprehensive assessment and targeted intervention as prevention or management.

^**See text and references for details(74, 142, 172, 174, 195, 196)

Despite existing guidance for practice, there are several essential needs for addressing remaining gaps. The three priority recommendations (Table 6) for patient-centered outcomes research in HCT survivors are 1) methodological framework development, 2) risk-targeted intervention testing, and 3) existing practice evaluation. To accomplish these goals, methodological gaps must be addressed especially in critical areas of standardized measurement and time point assessment. Although the collection of patient-centered outcomes for hypothesis driven research is a key priority, their application as secondary outcomes in treatment studies and in screening HCT survivors for problems is essential. Of equal importance is the need to design and test interventions that are reproducible and effectively address long-term and late effects that impair patient-centered outcomes in HCT survivors and their families. The needs of survivors will differ and intervention planning needs to target those with, or likely to develop specific long-term and late effects. Improving our ability to target those with elevated risk or need for specific prevention or treatment is a high priority. Intervention research for HCT survivors needs to include manualized interventions and process measures to examine consistency of delivery and feasibility of the intervention. Finally, documenting best practices at HCT centers will allow for collaborative efforts to improve care for survivors, especially those who are at a distance from their transplant centers. Comparing effective interventions including routine PRO symptom and functional monitoring will support timely implementation of quality care to survivors.

Table 6.

Priority research recommendations for patient-centered effects in HCT survivors

Design and test risk-targeted interventions with dose intensity modulation matching the needs of HCT survivors including measures of a) health resource utilization and costs, b) process measures including feasibility and treatment fidelity, and c) sustainability and dissemination potential. Priority Domains: Sexual dysfunction, Fatigue / sleep disruption, Non-adherence to medications & recommended health care, Health behaviors including physical inactivity and healthy eating, Psychological dysfunction. Special attention should be given to novel technologies allowing for easy access and sustainable use regardless of proximity to HCT center, and to the inclusion of informal caregivers as well as patient/caregiver dyads. Consideration should be given to replicating interventions with documented efficacy in general oncology survivors. Convene stakeholders to design a consensus-based methodological framework for both primary (hypothesis driven) and secondary outcome evaluation including standardized measures and time points and longitudinal, prospective designs. Priority populations: Pediatric and older survivors, non-malignant diagnoses, new treatment regimens, non-myeloablative regimens, and populations with comorbidities pre-transplant, disease relapse, and representing diverse races and ethnicity. Evaluate and compare existing practices for integrating patient-centered outcome screening across HCT survivorship programs to identify best practices and barriers. Address barriers and promote facilitators associated with incorporating patient-centered outcome data into electronic medical records.

The consensus of the working group is that improved synergy is needed across patient-centered outcomes and interventions, as well as with performance-based assessments and biomedical endpoints. These efforts could optimize efficiency, build a normative reference data set, and reach distant survivors using electronic PRO, wearable performance assessments, and standardized, brief measures such as the PRO-CTCAE and PROMIS (194). Research on patient-centered outcomes needs to focus on identifying risk factors for poorer outcomes that are amenable to intervention. While we wish to move the focus to interventions, more sophisticated methodologic approaches, and standardized measures and metrics are needed to address the inconsistencies in outcomes. Although past inconsistencies restrict the synthesis of findings and slow the forward progress in improving outcomes for HCT survivors, an integrated approach would serve as an exemplar for interdisciplinary care as well as team science to ensure quality survivorship care and effective, translational research.