Secondary Logo

Journal Logo

Original Research

Using the Pediatric Patient-reported Outcomes–common Terminology Criteria for Adverse Events to Longitudinally Measure Symptom Adverse Events of Children With Advanced Cancer

Montgomery, Kathleen E. PhD, RN, PCNS-BC, CPHON; Raybin, Jennifer L. PhD, RN, CPNP-BC; Ward, Jessica PhD, MPH, RN; Grinde, Krista DO; Smith, Corey BSN, RN; Brown, Roger PhD,

Author Information
Cancer Care Research Online: April 2022 - Volume 2 - Issue 2 - p e020
doi: 10.1097/CR9.0000000000000020
  • Open



Despite advances in supportive care and progress in the development and accessibility of symptom management interventions, children with cancer still experience a spectrum of physical and psychological symptoms.1–3 Symptom experiences are dynamic and are influenced by a myriad of factors, including cancer type and status, treatment type and intensity, and sociodemographics.1 Children with advanced cancer, those who experience a disease relapse or progression, or have refractory disease, are of particular interest given their exposure to previous treatment and risk for cancer-related death.4,5 Studies have captured symptom experiences of children and adolescents with advanced cancer, providing evidence of the presence of physical and psychological symptoms and symptom distress.6–8 Furthermore, symptom prevalence and distress increase as children approach end-of-life.7 Symptom management, including assessing and managing symptoms and reducing symptom-related suffering, was recently identified as a care priority for end-of-life among childhood cancer stakeholders.9

Including the child’s perspective is essential to understanding the symptom experience and improving care during advanced cancer.6,10 However, <10% of pediatric oncology clinical trials investigating anti-cancer treatment use patient-reported outcomes (PROs) as endpoints (primary or secondary).11 Despite evidence of PRO integration within studies involving children receiving upfront treatment for cancer, PRO integration for studies of children with advanced cancer or at end-of-life still lags.12,13

The Food and Drug Administration’s Advisory Committee recently endorsed inclusion of PROs in pediatric cancer trials to amplify the patient’s voice and provide critical information to monitor for adverse effects of treatment.14 Clinical trials of all phases (I–IV) have led to increasing survival rates among childhood cancers.15 Eighty-five percent of the estimated 15 000 individuals diagnosed with cancer aged birth to 20 will achieve a 5-year relative overall survival.16 The improved survival rate of childhood cancer is due in part to increased intensity of treatment which may result in worse symptom experiences.17 In the absence of readily available and published statistics describing the number of children who experience relapsed, refractory, or progressive cancer annually, based on survival estimates approximately 15% will experience disease relapse or cancer-related death. Children with advanced cancer, especially those who have incurable cancer, may benefit from early-stage clinical trials. Yet, our understanding of children’s experiences and the impact of early-stage trial participation on child well-being is limited.18 Incorporation of PRO measures in clinical trials may enhance our understanding of symptom experiences, streamline adverse event (AE) reporting, and support decision-making regarding acceptability of new and existing therapeutic regimens.

Various symptom measures have been used to elicit patient- and caregiver-proxy reported symptom experiences during the advanced cancer trajectory, including the Memorial Symptom Assessment Scale (MSAS), SSPedi, and PediQUEST MSAS, resulting in variability in the number of symptom items measured, recall period, and frequency or time interval of measurement.19 This variability has made comparison of children’s symptom experiences across studies difficult, thus symptom science experts have been working to consolidate symptom measures. Recently, researchers developed the Pediatric Patient-Reported Outcome version of the National Cancer Institute’s Common Terminology Criteria for Adverse Events (Ped-PRO-CTCAE) measurement system, which measures symptom AEs. The addition of the Ped-PRO-CTCAE provides another validated instrument to measure symptomatology and treatment-related toxicity that can be readily integrated into therapeutic and nontherapeutic clinical trials. The Ped-PRO-CTCAE was published in 2021, after a series of studies establishing content, cognitive appraisal, and content validation.20–23 The instrument was developed and tested with children with cancer and their caregivers and implemented in the health care setting. The Ped-PRO-CTCAE has not yet been evaluated for use to capture patient-reported symptom experiences among the pediatric advanced cancer population and outside the health care setting.

Previous studies have demonstrated the feasibility of monthly6 and biweekly7 intervals for symptom assessment forms consisting of 7–11 items. However, the feasibility of a biweekly interval using a longer comprehensive symptom assessment measure in children with advanced cancer and inclusion of participants who are non-English speaking remains undocumented. The Ped-PRO-CTCAE instrument has undergone psychometric testing required with measurement development,23–25 but its use in children in the context of a longitudinal advanced cancer trajectory where symptom assessment may occur remotely, regardless of patient location and during end-of-life has not been established. The objectives of the study were designed to directly address the limitations of current research through evaluating administration of a newly validated instrument in an expanded sample (inclusion of children with advanced disease and those who are Spanish speaking) across care settings, including at home, and determining the ability to capture patient-reported symptom AEs during end-of-life.


The study objectives were to (1) determine the feasibility and acceptability of administering the newly validated Ped-PRO-CTCAE biweekly over a 6-month period to children with advanced cancer, and (2) describe patient-reported symptom AEs.


Participants and Settings

Children 2–18 of age diagnosed with any type of advanced cancer, defined as a 1-week or longer history of relapsed, refractory, or progressive cancer or a decision to not pursue curative focused treatment, able to read and understand English or Spanish and otherwise able to complete study procedures were eligible for the study. Patients who transfer care to a study site for cancer treatment that was anticipated to last less than the study period of 6 months or were unable to complete study procedures were ineligible. Eligible children who provided assent or written consent per institutional policies participated in the study. All participants, with exception of those 18 years of age required written consent of 1 caregiver. Three geographically diverse pediatric cancer centers representing the Midwest, Rocky Mountains, and Pacific Coastal regions participated as study sites. The size of participating pediatric cancer programs ranged from mid to large.


The study was approved by the institutional review boards at 2 participating sites. A third participating site executed an institutional review board reliance agreement with the primary site.

Theoretical Framework

The Dynamic Symptoms Model26 provided the theoretical framework for the study. The model guided the selection of variables, including demographic, clinical, and symptom frequency, severity, and interference for the current study. The model also highlights the importance of evaluating the symptom experience over time.


A multi-institutional prospective longitudinal cohort study design was used for this study. Patient-reported AEs using the Ped-PRO-CTCAE were electronically collected at baseline and biweekly for up to 6 months (maximum of 12-time points). The Ped-PRO-CTCAE was sent as a survey to participants via a unique text message or e-mail (based on participant preference) from a central database, Research and Data Capture (REDCap), hosted by the primary site.27 Participants were asked to complete the survey within 7 days and were provided electronic reminders every 2–3 days, and a final phone call reminder on the survey due date. Participants received a $10 electronic gift card for each completed survey. Demographics were collected at baseline and clinical characteristics were extracted from the medical record at baseline and with each completed Ped-PRO-CTCAE report. Feasibility was also assessed with each symptom AE report and an acceptability survey was administered for participants who were enrolled for the entire study duration regardless of number of surveys completed. Acceptability surveys were not sent to participants who voluntarily withdrew or to caregivers of children who experienced a cancer-related death during the study period to mitigate the risk for potential distress.


Demographics and Clinical Characteristics

Demographic variables of the participant, including age, race, ethnicity, and clinical variables, including cancer type were extracted from the patient’s medical record.


The Ped-PRO-CTCAE, which contains 15 core and 47 additional items, was used to capture patient-reported (participants 8–18 years of age) or caregiver proxy reported (participants 2–7 years of age) attributes of 24 cancer treatment symptom AEs.23 Items use a recall period of 7 days with 4 possible response options to symptom AE attributes (frequency, severity, and interference with daily activities).23 Individual attribute items are categorized as nonconditional or conditional symptom AEs. Nonconditional symptom AEs are answered by participants with each assessment and consist of largely frequency AEs, except for fatigue, numbness, itch, concentration impairment, dizziness, restlessness, and dysphagia, for which severity was nonconditional. Conditional symptom AEs, which are largely severity and/or interference AEs, are only assessed if a participant endorses the paired nonconditional symptom AE item. Responses to individual symptom AE items occur on a 0–3 scale with scores of 3 reflecting higher or worse frequency, severity, and interference. Self-report and caregiver proxy versions of the Ped-PRO-CTCAE have undergone content validation21 and were evaluated using cognitive interviewing with children as young as 8 years.22 Data have demonstrated acceptable internal consistency, validity, and responsiveness of the self-report measure when compared to the MSAS, and validity and responsiveness of the parent-proxy measure when compared to Patient-Reported Outcomes Measurement Information System measures.23 Twenty-four nonconditional items (15 core, 9 additional) were selected for inclusion in this feasibility study to represent the most commonly experienced symptoms in the target population. Participants may also have been asked to respond to 34 additional conditional severity/interference items based on their response to nonconditional items for a maximum of 58 items. The study team obtained permission to use select symptom AE items (English and Spanish versions) from the researchers who developed the Ped-PRO-CTCAE before its publication online (Pamela S. Hinds, PhD and Molly McFatrich, MPH, e-mail communication, September 2019). The fatigue frequency item was inadvertently not included, thus fatigue severity was treated as the nonconditional symptom AE item for fatigue. All translated items were reviewed again by a translation service at the primary site per institutional review board policies.

Feasibility and Acceptability

Feasibility data regarding the number of surveys administered, completed, or deemed missing were extracted from the REDCap database and tabulated. Questions regarding location and difficulty with survey completion were included at each time point. A brief 7-item acceptability survey was administered to participants who completed all 12-time points to gather information about the type of device used to complete surveys, technology issues encountered, overall difficulty completing surveys on a biweekly interval, the number of symptom AEs assessed, preferred survey interval, and preference for symptom AE reports to be seen by the health care team.


Descriptive statistics were used to analyze feasibility, acceptability, demographic, clinical, and symptom AE data. Given the absence of published recommendations for analyzing the Ped-PRO-CTCAE longitudinally, frequencies and mean scores were calculated across all participants and time points for each individual symptom AE item (0–3 scale).28 Continuous variables are reported using means and SDs, while categorical or binary data are reported using frequencies and percentages. Individual responses for most frequently occurring symptom AEs were tabulated and presented graphically as heat maps for the entire sample to illuminate high-level unique patterns of symptom AE data and provide information on individual patient experiences.29 Hierarchical cluster analysis using group averages and Manhattan distance was used to identify symptom and participant clusters de novo among nonconditional and conditional symptom AEs, respectively. Clusters were evaluated based on their conditional status based on the number of expected responses for each symptom AE item. Nonconditional symptom AE items are asked of participants with each time point, thus provide more data compared to conditional symptom items. All analyses were conducted using NCSS 12 Statistical Software (2018) (NCSS, LLC., Kaysville, Utah,


Demographics and Clinical Characteristics

Forty-nine children were included in the final sample (Figure 1). The mean age of participants was 11.1 years (range 2–18). The sample included diverse cancer diagnoses (leukemia/lymphoma 37%, solid tumor 45%, central nervous system [CNS] tumor 18%) and race was consistent with state-wide demographics of the represented sites (71% white). Sex was not represented equally with slightly more males (67%). Thirty-five percent of the sample identified as Hispanic or Latino. Sample characteristics are listed in Table 1.

Table 1. - Sample Characteristics for the Sample
n Mean or Percent
Age 49 11.1 y
 Male 33 67.3%
 Female 16 32.7%
 Asian 3 6.1%
 Black or African American 0 0%
 Native Hawaiian or Other Pacific Islander 2 4.1%
 White 39 79.6%
 Other 11 22.5%
 Hispanic or Latino 17 34.7%
 Not Hispanic or Latino 32 65.3%
 Leukemia or lymphoma 18 36.7%
 Solid tumor 22 44.9%
 CNS tumor 9 18.4%
Abbreviation: CNS, central nervous system.
*Documentation allowed for selection of more than 1 race.

Figure 1.:
The Consolidated Standards of Reporting Trials diagram for 6 months of enrollment.

Feasibility and Acceptability

Participants completed 85% of administered symptom measures (n = 441/515) over an average of 16 weeks (range 2–24 weeks). Thirty-eight participants were enrolled for the full duration of the study. Seven children in the sample (6 English speaking and 1 Spanish speaking) experienced a cancer-related death while on the study, providing self-report or parent-proxy data for 83% of administered surveys (n = 20/24) within the last 12 weeks of life. Three participants received all study surveys in Spanish for a total of 32-time points with a completion rate of 91%. Most symptom measures were completed at home (71%). The Ped-PRO-CTCAE (child and parent-proxy versions) demonstrated acceptable reliability with a Cronbach alpha coefficient of 0.84. The acceptability survey was sent to 38 of 49 participants who were enrolled for the full duration of the study. Participants who did not receive the survey included those who experienced a cancer-related death (n = 7), who voluntarily withdrew before study completion (n = 3), or who did not receive the survey from the research team (n = 1). Thirty-two participants completed the acceptability survey. Most participants (88%) completed data collection using a smartphone (88%), almost never experienced technology problems (88%), and experienced little difficulty with completing symptom measures biweekly (88%). Half (52%) wanted to report symptoms every 2 weeks, followed by monthly (26%), and weekly (22%).

Almost half of participants expressed sometimes wanting their symptom reports to be communicated to their medical team (47%), followed by always or almost always (37%). Sixteen percent said they would never want their reports to be communicated. Frequencies of open-ended comments revealed participants felt it was important for providers to know what symptoms they are having, and that more communication was perceived as helpful (n = 17). Others felt communication about their symptoms was already adequate (n = 6). Few participants expressed difficulty in remembering what to tell their medical team at appointments (n = 2) and one felt their provider “did not care about their symptoms” (n = 1).

Descriptive Symptom Findings

Participants reported an average of 7 symptom AEs (of 24 total nonconditional symptom AEs) across 12-time points. The average number of symptom AEs was highest at T1 (n = 8), when a diagnosis of advanced cancer was recent, and steadily declined at T12 (n = 5).

Symptom Attribute Counts

Fatigue severity (58.7%), anxiety frequency (48.3%), pain frequency (47.6%), nausea frequency (45.5%), anorexia frequency (45.3%), and insomnia frequency (39.0%) were the most reported nonconditional symptom AEs across all participants and time points (Table 2). Dysphagia severity and frequency AEs for dyspnea, urinary urgency and urinary frequency were the least reported. Diagnostic groups varied in their most reported symptom AEs (Table 3). Children diagnosed with a solid tumor reported fatigue severity, anxiety, pain, and nausea frequency more than half of the time. Children diagnosed with a CNS tumor reported fatigue severity (53.3%) as the most frequently reported nonconditional symptom AE, while children diagnosed with hematologic malignancies reported anorexia frequency (54.7%) and fatigue severity (53.7%) as most occurring.

Table 2. - Symptom AE Counts for the Sample
Symptom AE* AE Reported AE Assessed Percent
Abdominal pain F* 170 440 38.6
Abdominal pain S 166 170 97.6
Abdominal pain I 89 170 52.3
Constipation F* 104 441 23.5
Constipation S 92 104 88.4
Constipation I 40 104 38.4
Diarrhea F* 156 441 35.3
Diarrhea I 48 156 30.7
Mucositis oral F* 73 441 16.5
Mucositis oral S 66 73 90.4
Mucositis oral I 30 72 41.6
Nausea F* 201 441 45.5
Nausea S 195 201 97.0
Nausea I 107 201 53.2
Vomit F* 115 439 26.2
Vomit I 50 115 43.4
Fatigue S* 259 441 58.7
Fatigue I 177 259 68.3
Pain F* 210 441 47.6
Pain S 204 209 97.6
Pain I 109 203 53.6
Anorexia F* 200 441 45.3
Headache F* 125 441 28.3
Headache S 119 125 95.2
Headache I 61 125 48.8
Numbness S* 64 441 14.5
Numbness I 24 63 38.1
Anxiety F* 213 441 48.3
Anxiety S 198 213 92.9
Anxiety I 73 213 34.2
Depression F* 164 440 37.2
Depression I 81 164 49.3
Insomnia F* 172 440 39.0
Insomnia S 159 170 93.5
Insomnia I 72 169 42.6
Cough F* 91 439 20.7
Cough S 84 90 93.3
Cough I 9 90 10.0
Dyspnea F* 51 440 11.5
Dyspnea S 45 51 88.2
Dyspnea I 25 51 49.0
Itch S* 108 440 24.5
Itch I 23 106 21.7
Urinary frequency F* 58 440 13.1
Urinary frequency I 11 55 20.0
Urinary urgency F* 49 441 11.1
Urinary urgency I 13 47 27.6
Concentration impairment S* 92 441 20.8
Concentration impairment I 58 89 65.1
Dizziness S* 64 440 14.5
Dizziness I 36 63 57.1
Hot flashes F* 89 441 20.1
Hot flashes S 74 84 88.1
Hot flashes I 20 83 24.1
Restlessness S* 72 441 16.3
Restlessness I 19 71 26.7
Dysphagia S* 48 441 10.8
Abbreviations: AE, adverse event; F, frequency; I, interference; S, severity.
*Nonconditional symptom AE.

Table 3. - Symptom AE Counts for the Sample by Diagnostic Group
Hematologic Malignancy Solid Tumor CNS Tumor
Symptom AE* AE Reported AE Assessed Percent AE Reported AE Assessed Percent AE Reported AE Assessed Percent
Abdominal pain F* 50 148 33.8 87 195 44.6 33 97 34.0
Abdominal pain S 49 50 98.0 85 87 97.7 32 33 97.0
Abdominal pain I 32 50 64.0 41 87 47.1 16 33 48.5
Constipation F* 33 148 22.3 54 196 27.6 17 97 17.5
Constipation S 29 33 87.9 46 54 85.2 17 17 100
Constipation I 15 33 45.5 14 54 25.9 11 17 64.7
Diarrhea F* 63 148 42.6 78 196 39.8 15 97 15.5
Diarrhea I 20 63 31.7 24 78 30.8 4 15 26.7
Mucositis oral F* 29 148 19.6 33 196 16.8 11 97 11.3
Mucositis oral S 28 29 96.6 27 33 81.8 11 11 100
Mucositis oral I 17 28 60.7 7 33 21.2 6 11 54.5
Nausea F* 59 148 39.9 105 196 53.6 37 97 38.1
Nausea S 58 59 98.3 102 105 97.1 35 37 94.6
Nausea I 35 59 59.3 52 105 49.5 20 37 54.1
Vomit F* 46 147 31.3 44 196 22.4 25 96 26.0
Vomit I 18 46 39.1 19 44 43.2 13 25 52.0
Fatigue S* 78 148 52.7 129 196 65.8 52 97 53.6
Fatigue I 55 78 70.5 89 129 69.0 33 52 63.5
Pain F* 65 148 43.9 118 196 60.2 27 97 27.8
Pain S 64 64 100 114 118 96.6 26 27 96.3
Pain I 51 64 79.7 48 117 41.0 10 22 45.5
Anorexia F* 81 148 54.7 81 196 41.3 38 97 39.2
Headache F* 38 148 25.7 69 196 35.2 18 97 18.6
Headache S 37 38 97.4 64 69 92.8 18 18 100
Headache I 26 38 68.4 25 69 36.2 10 18 55.6
Numbness S* 11 148 7.4 47 196 24.0 6 97 6.2
Numbness I 6 11 54.5 18 47 38.3 0 5 0.0
Anxiety F* 56 148 37.8 120 196 61.2 37 97 38.1
Anxiety S 55 56 98.2 109 120 90.8 34 37 91.9
Anxiety I 24 56 42.9 39 120 32.5 10 37 27.0
Depression F* 58 148 39.2 77 195 39.5 29 97 29.9
Depression I 39 58 67.2 34 77 44.2 8 29 27.6
Insomnia F* 66 148 44.6 89 195 45.6 17 97 17.5
Insomnia S 59 65 90.8 86 88 97.7 14 17 82.4
Insomnia I 28 65 43.1 36 87 41.4 8 17 47.1
Cough F* 29 148 19.6 45 194 23.2 17 97 17.5
Cough S 28 29 96.6 40 44 90.9 16 17 94.1
Cough I 5 29 17.2 0. 44 0.0 4 17 23.5
Dyspnea F* 6 148 4.1 39 195 20.0 6 97 6.2
Dyspnea S 6 6 100 34 39 87.2 5 6 83.3
Dyspnea I 4 6 66.7 18 39 46.2 3 6 50.0
Itch S* 48 148 32.4 39 195 20.0 21 97 21.6
Itch I 10 46 21.7 7 39 17.9 6 21 28.6
Urinary frequency F* 19 148 12.8 22 195 11.3 17 97 17.5
Urinary frequency I 3 18 16.7 4 20 20.0 4 17 23.5
Urinary urgency F* 15 148 10.1 27 196 13.8 7 97 7.2
Urinary urgency I 1 14 7.1 8 26 30.8 4 7 57.1
Concentration impairment S* 20 148 13.5 56 196 28.6 16 97 16.5
Concentration impairment I 14 19 73.7 34 54 63.0 10 16 62.5
Dizziness S* 15 148 10.1 40 196 20.4 9 96 9.4
Dizziness I 12 15 80.0 16 39 41.0 8 9 88.9
Hot flashes F* 13 148 8.8 68 196 34.7 8 97 8.2
Hot flashes S 13 13 100 55 63 87.3 6 8 75.0
Hot flashes I 3 13 23.1 17 62 27.4 0 8 0
Restlessness S* 17 148 11.5 42 196 21.4 13 97 13.4
Restlessness I 4 16 25.0 11 42 26.2 4 13 30.8
Dysphagia S* 19 148 12.8 25 196 12.8 4 97 4.1
Abbreviations: AE, adverse event; CNS, central nervous system; F, frequency; I, interference; S, Severity.
*Nonconditional symptom AE.

Symptom Attribute Mean Scores

Participants reported the highest mean scores (scale 0–3) for symptom severity AEs of pain (M = 1.35, SD = 0.66), nausea (M = 1.31, SD = 0.43), headache (M = 1.26, SD = 0.57), abdominal pain (M = 1.24, SD = 0.44), constipation (M = 1.21, SD = 0.71), insomnia (M = 1.16, SD = 0.47), and anxiety (M = 1.13, SD = 0.42) (Figure 2). Participants reported that pain (M = 0.91, SD = 0.88), fatigue (M = 0.88, SD = 0.53), concentration impairment (M = 0.83, SD = 0.57), dizziness (M = 0.73, SD = 0.61), and nausea (M = 0.70, SD = 0.64) interfered the most with their daily lives. Pain (M = 0.72, SD = 0.68), anorexia (M = 0.72, SD = 0.70), insomnia (M = 0.68, SD = 0.71), nausea (M = 0.66, SD = 0.50), and anxiety (M = 0.63, SD = 0.54) were most frequently occurring. Diagnostic differences in mean symptom AE scores are visualized in Figure 3A through 3C. Similar scores for symptom frequency AEs were reported across disease groups (Figure 3A). Children diagnosed with a CNS tumor reported more symptoms with increased severity or interference compared to children diagnosed with a hematologic malignancy or solid tumor (Figure 3B and 3C). Differences in interference scores were most notable for fatigue, constipation, pain, numbness, insomnia, dyspnea, and dizziness.

Figure 2.:
Symptom AEs are ordered by lowest mean score (top of figure) to highest mean score (bottom of figure) for all participants across all time points. Bars indicate the mean score for individual symptom AEs. Higher scores reflect greater frequency, increased severity, or higher levels of interference for symptom AEs. AEs indicate adverse events.
Figures 3.:
Mean Symptom AE Scores for Disease Groups. A–C, Radar plots display mean scores for frequency (A), severity (B), and interference (C) symptom AEs for all participants across all time points by their cancer disease group. Colored lines indicate each disease group (red = hematologic malignancy, green = solid tumor; blue = CNS tumor). Higher scores reflect greater frequency, increased severity, or higher levels of interference for symptom AEs. AEs indicates adverse events; CNS, central nervous system.

Symptom Clusters

Results of hierarchical cluster analyses for nonconditional and conditional symptom AEs are displayed in Figures 4 and 5, respectively. Heat maps display mean symptom AE scores across all time points for Ped-PRO-CTCAE items by individual participant (numbered 1–49). The blue and yellow shaded boxes for each figure indicate grouping of participants, while the red, yellow, and green shaded boxes indicate grouping of symptom AEs. Two distinct clusters of participants (Low Symptom and High Symptom) were identified for nonconditional symptom AEs, with most participants comprising the Low Symptom group experienced lower average symptom AE scores compared to the High Symptom group. One participant did not cluster to a either group. Three symptom AE clusters were identified among nonconditional AEs. Diarrhea and anorexia frequency AEs clustered together to create a Diarrhea-Anorexia cluster, while non-nausea, fatigue, pain, and insomnia clustered to form a Nausea-Fatigue-Pain-Insomnia cluster. The remaining symptom AEs clustered to comprise a Low Multi-symptom cluster.

Figure 4.:
Clustered heat map displaying mean scores for nonconditional symptom AEs by individual subject (numbered 1–49). Individual colored boxes indicate the mean symptom AE score on a 0–3 scale with blue indicating a score of 0 and red indicating a score of 3. Participant clusters are represented on the left side by the blue and yellow shaded boxes. Symptom clusters are represented on the topside by the red, yellow, and green shaded boxes. Mention gray indicates missing data or no response for conditional question. AEs indicate adverse events.
Figure 5.:
Clustered heat map displaying mean scores for nonconditional symptom AEs by individual subject (numbered 1–49). Individual colored boxes indicate the mean symptom AE score on a 0–3 scale with blue indicating a score of 0 and red indicating a score of 3. Gray boxes indicate no (due to conditional nature of item) or missing data. Participant clusters are represented on the left side by the blue and yellow shaded boxes. Symptom clusters are represented on the topside by the red, yellow, and green shaded boxes. AEs indicate adverse events.

Analysis of conditional symptom AEs yielded 2 participants clusters (Low Symptom and High Symptom). Most participants were included in Low Symptom and experienced lower symptom AE scores compared to the High Symptom group. The same individual participant (16) did not cluster to either group. Three symptom clusters were identified based on interference and severity symptom AEs and characterized as Severity, Low Interference, and High Interference. The Severity cluster consisted of all conditional severity symptom AEs. The Low Interference cluster included symptom AEs with lower average interference scores compared to the High Interference cluster.


A biweekly interval for assessing symptom AEs over a 6-month period was found to be feasible and acceptable. For the Spanish-speaking arms of the study, 3 participants received all surveys in Spanish, and our data suggested similar completion rates with English speaking participants. The assessment interval timing, 6-month time period, and Spanish speaking feasibility findings add to a growing body of literature supporting longitudinal symptom evaluation using different intervals and symptom measures with variable numbers of items and adds preliminary evidence of feasibility and acceptability among Spanish-speaking participants.

We have replicated our previous work which demonstrated just slightly higher completion rates with a symptom assessment form consisting of 7–11 items.7 Our current study showed evidence of good completion rates and minimal missing data.7 Similar data and person missingness rates results were found in a study evaluating the feasibility and acceptability of a standardized approach of measuring symptoms and function among children with incurable cancer receiving treatment in a phase 1/2 cancer trial.30 The existence of feasibility and acceptability data is encouraging for researchers who seek to characterize symptom experiences over time or wish to use PRO symptom data as clinical trial endpoints to support decision-making for standard of care therapeutic regimens. Furthermore, for some children diagnosed with cancer, longitudinal monitoring may support PRO symptom capture during the critical period of end-of-life.

Study findings add further evidence for the presence and persistence of priority symptoms in children and adolescents with advanced cancer,3,6,31 those without advanced disease,1 and those receiving pediatric palliative care,32 including fatigue, pain, anxiety, nausea, anorexia, and insomnia. Symptom AE occurrence in this sample was notably higher compared to other studies of children with advanced cancer. Compared to our previous study with a comparable sample size, the occurrence for fatigue, anxiety, and anorexia in this sample were more than doubled, while pain, nausea, and insomnia were 12%–18% higher.7 Furthermore, those increases were also reflected in symptom AE occurrence by diagnostic group with the biggest increases occurring in children diagnosed with solid tumors. These differences may be attributed to the use of different symptom measures (PediQUEST MSAS versus Ped-PRO-CTCAE), the expected variation between disease and treatment characteristics throughout the advanced cancer trajectory, and changes in treatment regimens and modalities over time. However, both symptom measures include similar symptoms and use of conditional questions.

Symptom AEs that were most occurring are aligned with the 15 Ped-PRO-CTCAE core items20 and should be considered for routine monitoring. The occurrence of a spectrum of symptom AEs in the sample and over time suggest monitoring plans should be tailored to best match the child’s disease and treatment plan. In response to a diagnosis of advanced cancer, special consideration should also include the child’s symptom history and input or preference on priority symptoms.

The highest mean symptom AE scores for the sample were associated with symptom’s severity attribute. The absence of interference AEs in the top ten highest mean scores was somewhat surprising to our team. This may be due to several reasons. Children with advanced cancer likely have undergone treatment for a greater period of time compared to those without advanced disease, thus they may view their symptoms as less interfering with their daily activities (eg, norming their symptom experience). Alternatively, their daily activities may be hindered by their disease and/or treatment. Symptom severity AEs may be useful to clinicians for assessing the degree of toxicity associated with treatment. When comparing diagnostic groups, mean symptom AE scores were greater for several severity and interference AEs in children with CNS tumors. These findings may best be explained by disease location and treatment type. For example, children with CNS tumors often receive vinca alkaloids for treatment,33 leading to side effects of constipation, abdominal pain, headache, dizziness, and numbness and tingling.34

Symptom clusters have been previously described in several studies among children diagnosed with cancer.3 Findings from the hierarchical cluster analysis among nonconditional and conditional symptom AEs suggests preliminary signal towards clustering of both children with advanced cancer and certain symptom AEs, including nausea, fatigue, pain, and insomnia, which warrant further investigation. Different participants represented each distinct cluster group across nonconditional and conditional symptom AEs, suggesting participant-level clustering may vary based on symptom AE attribute (frequency, severity, and interference) or the presence of nonconditional symptom AEs. Our approach to cluster analysis based on the conditional nature of the item may have been most useful for exploring the presence of symptom AE clusters among nonconditional items. Nonconditional symptom AEs provide researchers far more data points compared to conditional symptom AEs, thus may be most appropriate to characterize symptom clusters when using the Ped-PRO-CTCAE measure. We found that conditional symptom AEs exclusively clustered based on their attribute (severity versus interference) and may be best explored separately. Reports of symptom clusters are highly dependent on the symptom measure, sample characteristics, and treatment, leading to challenges in generating reproducible results across studies.35 However, evidence supporting clustering of patients may provide important signal to support intentional investigation of symptom profiles and identification of children at risk for high symptom suffering. Additionally, the use of different visualization techniques to illustrate descriptive findings may be useful in illuminating participant-and sample-level trends.

Strengths and Limitations

Notable strengths of this study include the (1) multisite prospective longitudinal design, (2) heterogeneity of the sample with diverse representation of tumor types, (3) use of a long and comprehensive symptom assessment form, and (4) inclusion of Spanish-speaking participants. These design elements provide important feasibility data for investigators interested in evaluating symptoms and AEs over time in or outside the context of a clinical trial. The study strengths must also be balanced with limitations. The primary and expected limitation was the small sample size given the feasibility nature of the study, limiting our ability to detect differences between children with different cancer types, treatment characteristics, developmental/age groups, or primary language. Potential bias may be present in our acceptability data because the acceptability survey was only administered to participants who were enrolled on the study for the full 6 months, thus excluding important perspectives of participants who voluntarily withdrew before study completion. Furthermore, despite the large numbers of completed measures for symptom AEs, we may have missed identification of certain symptom AE- and patient-level clusters.

Implications for Clinical Practice

The persistent occurrence, severity of symptoms, and their interference with activities among children with advanced cancer highlights the urgency to develop scalable solutions. Systems that can capture patient-reported, or when appropriate caregiver proxy report, symptom data across the continuum of treatment for advanced cancer are necessary to promote timely symptom assessment, intervention, and evaluation. The number of items selected for routine or ad hoc symptom assessments should be determined using a holistic approach, considerate of the patient’s previous symptom experiences, physical and psychosocial needs, and cancer treatment medications or modalities and associated toxicities. If too few items are selected for inclusion, patients may be uncertain of how to express additional symptoms, risking clinicians being unaware of the presence of certain symptoms. This point is exemplified by the results of a recent study of polysymptomatology among pediatric patients receiving palliative care. When the number of symptoms assessed increased to 20 from 7 (used in a previous study), the average number of symptoms reported increased 2-fold.32 If too many items are selected, patients may experience potential burden associated with time required to complete long assessments and clinicians may be constrained by barriers, including insufficient staff, time, and resources to address multiple independent or co-occurring symptoms.36

Future Research

Researchers can leverage longitudinal symptom data generated from routine monitoring associated with or without a clinical trial to characterize the presence of symptom profiles among children with advanced cancer who express high symptom AE occurrence, severity, and/or interference. A recent study identified symptom profiles among children diagnosed with cancer to illuminate a subset of children with high symptom suffering.1 Future research can expand our understanding of symptom suffering profiles by identifying and characterizing high-risk symptom phenotypes. Discovery of symptom phenotypes, including their persistence or the ability for children with a high-risk phenotype to transition to a low-risk phenotype will support researchers in developing novel screening techniques to intervene on poorly controlled symptoms in a timely manner.

Institution IRB Type of Approval IRB #
University of Wisconsin-Madison Health Sciences IRB Full review and approval for Minimal Risk 2019-1068
Children’s Hospital Colorado Colorado Multiple Institutional Review Board Expedited approval for Minimal Risk 19-2198
Children’s Hospital Los Angeles CHLA IRB Ceded Review to UW-Madison

To accomplish this goal of characterizing symptom experiences, including symptom AE clusters, across the continuum of care, further guidance on analysis of longitudinal data generated from the Ped-PRO-CTCAE is required. This is especially important for children with advanced cancer and the evolution of precision medicine. Participant data may become increasingly heterogenous, even within single cancer types, due to diversity in treatment characteristics with tailored treatment regimens and individual genetic profiles. Single symptom attribute analysis provides simple and useful attribute-level data28 and reduces the risk of masking symptom suffering (Pamela S. Hinds, PhD, oral communication, June 2021) by honoring the complete experience of symptom frequency, severity, and interference. However, inability to capture total or subscale scores, may limit researchers in examining relationships among broader constructs, like symptom burden or symptom suffering and health-related quality of life. The PRO-CTCAE version used with adult populations recently developed a composite grading algorithm, allowing researchers to calculate and analyze a single numerical grade per symptom AE by combining frequency, severity, and interference scores.37

To advance the science in symptomatology in children with advanced cancer and increase patient-reported AE reporting, future research must be inclusive of experiences of non-English speaking participants.13 Certain racial and ethnic minorities, like Hispanic children and children of Spanish-speaking parents, have shown to be underrepresented in pediatric cancer research.38,39 This underrepresentation is attributed to several barriers at the structural, clinical, provider, and patient levels, and include culture and language barriers.13,38 Psychometrically testing the Ped-PRO-CTCAE in multiple languages will promote representation across studies investigating symptom experiences or treatment-related toxicities and support understanding the unique contribution of culture and language to symptom experiences.


In summary, this study provides evidence of feasibility for electronically administering the Ped-PRO-CTCAE biweekly across the continuum of care for children with advanced cancer, including at end-of-life. Children with relapsed, refractory, or progressive disease continue to experience a spectrum of symptoms, including pain, fatigue, nausea, anxiety, and anorexia. There is an early signal to suggest clustering may occur among symptom AEs and individual participants. The development of the Ped-PRO-CTCAE provides clinicians and researchers alike with a new tool to readily assess symptom AEs from the patient perspective that can be integrated in therapeutic clinical trials and within the standard of care.


The authors would like to acknowledge members of the Ped-PRO-CTCAE Research Team, including Drs. Pamela Hinds and Janice Withycombe and Molly McFatrich for their permission to use the Ped-PRO-CTCAE instrument and guidance on analysis, and Dr. Kris Kwekkeboom for her mentorship.

Author Contributions: Kathleen E. Montgomery was involved in conceptualization, investigation, supervision, writing—original draft, writing—review and editing, visualization, supervision, and funding acquisition. Jennifer L. Raybin was involved in conceptualization, investigation, supervision, writing—original draft, writing—review and editing. Jessica Ward was involved in conceptualization, investigation, supervision, writing—review and editing. Krista Grinde was involved in conceptualization, investigation, data curation, writing—review and editing. Corey Smith was involved in investigation, data curation, writing—review and editing. Roger Brown was involved in conceptualization, formal analysis, visualization, writing—review and editing.

This manuscript reflects a prospective longitudinal nurse-led multisite research project, involving human subjects research. Above contains institution-specific IRB information:


1. Hinds PS, Weaver MS, Withycombe JS, et al. Subjective toxicity profiles of children in treatment for cancer: a new guide to supportive care? J Pain Symptom Manage. 2021;61:1188–1195.e2.
2. Hooke MC, Linder LA. Symptoms in children receiving treatment for cancer-part I: fatigue, sleep disturbance, and Nausea/Vomiting. J Pediatr Oncol Nurs. 2019;36:244–261.
3. Linder LA, Hooke MC. Symptoms in children receiving treatment for cancer-part II: pain, sadness, and symptom clusters. J Pediatr Oncol Nurs. 2019;36:262–279.
4. Nguyen K, Devidas M, Cheng SC, et al.; Children’s Oncology Group. Factors influencing survival after relapse from acute lymphoblastic leukemia: a Children’s Oncology Group study. Leukemia. 2008;22:2142–2150.
5. Ceschel S, Casotto V, Valsecchi MG, et al. Survival after relapse in children with solid tumors: a follow-up study from the Italian off-therapy registry. Pediatr Blood Cancer. 2006;47:560–566.
6. Wolfe J, Orellana L, Ullrich C, et al. Symptoms and distress in children with advanced cancer: prospective patient-reported outcomes from the PediQUEST study. J Clin Oncol. 2015;33:1928–1935.
7. Montgomery KE, Raybin JL, Ward J, et al. Using patient-reported outcomes to measure symptoms in children with advanced cancer. Cancer Nurs. 2020;43:281–289.
8. Zhukovsky DS, Rozmus CL, Robert RS, et al. Symptom profiles in children with advanced cancer: patient, family caregiver, and oncologist ratings. Cancer. 2015;121:4080–4087.
9. Ananth P, Mun S, Reffat N, et al. Refining patient-centered measures of end-of-life care quality for children with cancer. JCO Oncol Pract. 2021;18:e372–e382.
10. Montgomery K, Sawin KJ, Hendricks-Ferguson VL. Experiences of pediatric oncology patients and their parents at end of life: a systematic review. J Pediatr Oncol Nurs. 2016;33:85–104.
11. Riedl D, Rothmund M, Darlington AS, et al.; EORTC Quality of Life Group. Rare use of patient-reported outcomes in childhood cancer clinical trials - a systematic review of clinical trial registries. Eur J Cancer. 2021;152:90–99.
12. Hinds PS, Brandon J, Allen C, et al. Patient-reported outcomes in end-of-life research in pediatric oncology. J Pediatr Psychol. 2007;32:1079–1088.
13. Tomlinson D, Yuan C, Cheng L, et al. Patient-Reported Outcomes in Pediatric Oncology: The Voice of the Child. In: Hinds PS, Linder L, eds. Pediatric Oncology Nursing: Defining Care Through Science. Springer International Publishing; 2020:107–129.
14. Sutter S. Patient-reported outcomes in pediatric cancer trials endorsed by US FDA Advisory Committee. Pink Sheet. Informa Pharma Intelligence. 2021. Accessed May 19, 2021.
15. O’Leary M, Krailo M, Anderson JR, et al.; Children’s Oncology Group. Progress in childhood cancer: 50 years of research collaboration, a report from the Children’s Oncology Group. Semin Oncol. 2008;35:484–493.
16. Surveillance Epidemiology and End Results (SEER) Program. SEER*Stat Database: Relative Survival - SEER Research Data, Nov 2020 Sub (2000-2018). SEER Research Data; 2020.
17. Tomlinson D, Baggott C, Dix D, et al. Severely bothersome fatigue in children and adolescents with cancer and hematopoietic stem cell transplant recipients. Support Care Cancer. 2019;27:2665–2671.
18. Crane S, Haase JE, Hickman SE. Well-being of child and family participants in phase 1 pediatric oncology clinical trials. Oncol Nurs Forum. 2018;45:E67–E97.
19. Leahy AB, Feudtner C, Basch E. Symptom monitoring in pediatric oncology using patient-reported outcomes: why, how, and where next. Patient. 2018;11:147–153.
20. Reeve BB, Withycombe JS, Baker JN, et al. The first step to integrating the child’s voice in adverse event reporting in oncology trials: a content validation study among pediatric oncology clinicians. Pediatr Blood Cancer. 2013;60:1231–1236.
21. Weaver MS, Reeve BB, Baker JN, et al. Concept-elicitation phase for the development of the Pediatric Patient-reported Outcome Version of the Common Terminology Criteria for Adverse Events. Cancer. 2016;122:141–148.
22. Reeve BB, McFatrich M, Pinheiro LC, et al. Eliciting the child’s voice in adverse event reporting in oncology trials: cognitive interview findings from the Pediatric Patient-Reported Outcomes version of the Common Terminology Criteria for Adverse Events initiative. Pediatr Blood Cancer. 2017;64:10.1002/pbc.26261.
23. Reeve BB, McFatrich M, Mack JW, et al. Validity and reliability of the pediatric patient-reported outcomes version of the common terminology criteria for adverse events. J Natl Cancer Inst. 2020;112:1143–1152.
24. Reeve BB, McFatrich M, Pinheiro LC, et al. Cognitive interview-based validation of the patient-reported outcomes version of the common terminology criteria for adverse events in adolescents with cancer. J Pain Symptom Manage. 2017;53:759–766.
25. Reeve BB, McFatrich M, Lin L, et al. Validation of the caregiver pediatric patient-reported outcomes version of the common terminology criteria for adverse events measure. Cancer. 2021;127:1483–1494.
26. Brant JM, Dudley WN, Beck S, et al. Evolution of the dynamic symptoms model. Oncol Nurs Forum. 2016;43:651–654.
27. Harris PA, Taylor R, Thielke R, et al. Research electronic data capture (REDCap)–a metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform. 2009;42:377–381.
28. Hinds PS, Pinheiro LC, McFatrich M, et al. Recommended scoring approach for the pediatric patient-reported outcomes version of the Common Terminology Criteria for Adverse Events. Pediatr Blood Cancer. 2021:e29452. doi: 10.1002/pbc.29452.
29. Thanarajasingam G, Atherton PJ, Novotny PJ, et al. Longitudinal adverse event assessment in oncology clinical trials: the Toxicity over Time (ToxT) analysis of Alliance trials NCCTG N9741 and 979254. Lancet Oncol. 2016;17:663–670.
30. Hinds PS, Wang J, Stern ED, et al. Voices of children and adolescents on phase 1 or phase 2 cancer trials: a new trial endpoint? Cancer. 2017;123:3799–3806.
31. Hendricks-Ferguson VL, Sawin KJ, Montgomery K, et al. Novice Nurses’ experiences with palliative and end-of-life communication. J Pediatr Oncol Nurs. 2015;32:240–252.
32. Feudtner C, Nye R, Hill DL, et al.; Pediatric Palliative Care Research Network Shared Data and Research (PPCRN SHARE) Project Group. Polysymptomatology in pediatric patients receiving palliative care based on parent-reported data. JAMA Netw Open. 2021;4:e2119730.
33. Bjornard KL, Gilchrist LS, Inaba H, et al. Peripheral neuropathy in children and adolescents treated for cancer. Lancet Child Adolesc Health. 2018;2:744–754.
34. Moudi M, Go R, Yien CY, et al. Vinca alkaloids. Int J Prev Med. 2013;4:1231–1235.
35. Rodgers C, Hooke MC, Ward J, et al. Symptom clusters in children and adolescents with cancer. Semin Oncol Nurs. 2016;32:394–404.
36. Schepers SA, Haverman L, Zadeh S, et al. Healthcare professionals’ preferences and perceived barriers for routine assessment of patient-reported outcomes in pediatric oncology practice: moving toward international processes of change. Pediatr Blood Cancer. 2016;63:2181–2188.
37. Basch E, Becker C, Rogak LJ, et al. Composite grading algorithm for the National Cancer Institute’s Patient-Reported Outcomes version of the Common Terminology Criteria for Adverse Events (PRO-CTCAE). Clin Trials. 2021;18:104–114.
38. Aristizabal P. Diverse populations and enrollment in pediatric cancer clinical trials: challenges and opportunities. Pediatr Blood Cancer. 2020;67:e28296.
39. Aristizabal P, Singer J, Cooper R, et al. Participation in pediatric oncology research protocols: racial/ethnic, language and age-based disparities. Pediatr Blood Cancer. 2015;62:1337–1344.

symptoms; cancer; pediatric; patient-reported outcomes

Copyright © 2022 The Authors. Published by Wolters Kluwer Health, Inc. on behalf of Cancer Care Research Online. All rights reserved.