Validation of co–Symptom Screening in Pediatrics Tool: a novel dyadic approach to symptom screening in pediatric patients receiving cancer treatment

Abstract

Background

Co-Symptom Screening in Pediatrics Tool (co-SSPedi) is a dyadic (child-guardian) approach to symptom assessment. Objectives were to evaluate the reliability and validity of co-SSPedi for pediatric patients receiving cancer treatments.

Methods

This multicenter study included dyads of patients aged 4-18 years of age with cancer or undergoing hematopoietic cell transplant and their guardians. Two groups were enrolled. The more symptomatic group included those receiving active treatment for cancer or undergoing hematopoietic cell transplant where patients were in hospital or clinic for 4 consecutive days. The less symptomatic group included those receiving maintenance therapy for acute lymphoblastic leukemia or who had completed cancer treatments. At baseline, all dyads completed co-SSPedi, and guardians completed measures of mucositis, nausea, pain, quality of life, and overall symptoms. In the more symptomatic group, dyads completed co-SSPedi and a global symptom change scale on day 4.

Results

There were 501 dyads included: 301 in the more symptomatic group and 200 in the less symptomatic group. Median time to complete co-SSPedi was less than 3 minutes in both groups. Test-retest reliability intraclass correlation coefficient was 0.85 (95% confidence interval [CI] = 0.77 to 0.90). For internal consistency, total co-SSPedi Cronbach alpha was 0.81 (95% CI = 0.78 to 0.83). For known groups validation, mean difference in total co-SSPedi scores between the more symptomatic and less symptomatic groups was 7.8 (95% CI = 6.7 to 8.8; P < .0001). For convergent validation and responsiveness, all hypothesized relationships were demonstrated.

Conclusions

Co-SSPedi is a novel approach to dyadic symptom assessment that is reliable, valid, and responsive in pediatric patients aged 4-18 years.

Active symptom screening and monitoring are important to optimize symptom control and quality of life (QoL) for pediatric patients receiving cancer treatments (1-4). To enable symptom assessment for pediatric patients, we created and validated the Symptom Screening in Pediatrics Tool (SSPedi), mini-SSPedi, and proxy-SSPedi (5-7). All instruments were designed for pediatric patients with cancer and hematopoietic cell transplant recipients. We ask respondents to self-report (SSPedi and mini-SSPedi) or proxy report (proxy-SSPedi) how much 15 symptoms bothered them or their child day before or day of (SSPedi and proxy-SSPedi) or today (mini-SSPedi) on a 5-point or 3-point Likert scale (8-10). SSPedi was designed for children and adolescents aged 8-18 years of age, mini-SSPedi was designed for children aged 4-7 years of age, and proxy-SSPedi was designed for children and adolescents aged 2-18 years of age.

Thus, symptom assessment using SSPedi had previously focused on self-report or proxy report similar to other patient-reported outcomes in pediatric patients (8,9). With growing recognition that each reporter type has a valid perspective, it has been suggested that obtaining both self-report and proxy report could be ideal. However, routine collection of both reporter types may not be feasible in clinical care. Further, there will be situations in which pediatric patients cannot or will not be willing to self-report symptoms. For example, patients who are acutely unwell with severe pain or fever are unlikely to be willing to complete a symptom assessment tool. However, it is in these scenarios where symptom reporting is particularly salient.

To address these challenges, we previously developed a dyadic approach to symptom screening to facilitate implementation into routine practice. This approach involves the child and guardian completing SSPedi together (11). Our previous study developed a structured approach to the dyadic reporting of symptoms that ensured the children verbalized their impression of their symptoms first but allowed guardians to facilitate and mediate symptom reporting, particularly when children did not accurately remember their experience. We named this instrument co-SSPedi and found that when a dyadic approach was used, more co-SSPedi scores agreed with the original child self-report scores compared with original guardian proxy-report scores for discordant symptoms (11). Most dyads preferred to complete SSPedi together rather than separately.

The next step was to evaluate the psychometric properties of co-SSPedi. Objectives were to determine whether co-SSPedi is reliable and valid in pediatric patients receiving cancer treatments or hematopoietic cell transplant.

Methods

This multicenter prospective observational study was designed to evaluate the reliability, validity, and responsiveness of co-SSPedi for pediatric patients receiving cancer treatments or hematopoietic cell transplant recipients.

Participants

Participants were dyads consisting of English-speaking pediatric patients aged 4-18 years with cancer or hematopoietic cell transplant recipients and their English-speaking guardians. We excluded dyads where either the pediatric patient or guardian did not understand English or had cognitive disability or visual impairment that precluded completion of co-SSPedi as determined by the patient’s primary health-care team.

As in our other SSPedi validation studies, 2 groups were enrolled for the purpose of construct validation. The more symptomatic group included dyads where the pediatric patient was receiving active treatment for cancer or undergoing hematopoeitic stem cell transplant (HSCT) and was either an inpatient or outpatient attending the hospital for 4 consecutive days. The less symptomatic group included dyads where the pediatric patient was at least 6 months into maintenance therapy for nonrelapsed acute lymphoblastic leukemia or had completed any cancer treatment excluding hematopoietic cell transplant at least 3 months prior to enrollment and were clinically well (according to the primary team) with no planned procedures on the day of enrollment.

Procedures

Participating sites were The Hospital for Sick Children (Toronto, Ontario, Canada), British Columbia Children’s Hospital (Vancouver, British Columbia, Canada), and IWK Health Centre (Halifax, Nova Scotia, Canada). The research ethics boards of all participating sites approved this study. Guardians provided informed consent, and pediatric patients provided consent or assent as appropriate. Potential dyads were screened in the inpatient or outpatient setting. For participating dyads, those in the more symptomatic group completed measures on days 1 and 4 (± 1 day), and those in the less symptomatic group completed measures only on day 1.

Demographic information was obtained from guardians and from the patient’s health records; data included child sex, age, and race and guardian sex, relationship to the patient, marital status, education, English as a first language, and household income. Completion of co-SSPedi was performed in a structured fashion (see below). On day 1, dyads completed co-SSPedi and guardians completed the following proxy-reported measures for the purpose of construct validation: Children’s International Mucositis Evaluation Scale (ChIMES), the Pediatric Nausea Assessment Tool (PeNAT), Faces Pain Scale-Revised (FPS-R), a global QoL scale, and an overall symptom visual categorical scale. For dyads in the more symptomatic group, co-SSPedi was completed a second time on day 4 (3 days later) along with a global symptom change scale. This change scale was reported by the dyad together, and it described the change in the patient’s global symptoms on day 4 compared with day 1. Dyads of pediatric patients aged 8-18 years completed a 5-point Likert scale consisting of much better, a little better, the same, a little worse, or much worse, and dyads of pediatric patients aged 4-7 years completed a 3-point Likert scale consisting of better, the same, or worse.

Instruments

Co-SSPedi consists of the following 15 symptom descriptions: feeling disappointed or sad, feeling scared or worried, feeling cranky or angry, problems with thinking or remembering things, changes in how your body or face look, feeling tired, mouth sores, headache, hurt or pain (other than headache), tingly or numb hands or feet, throwing up or feeling like you may throw up, feeling more or less hungry than you usually do, changes in taste, constipation, and diarrhea. For pediatric patients aged 8-18 years, symptoms as experienced by the pediatric patient day before or day of were rated by the dyad on a 5-point Likert scale consisting of not at all bothered, a little, medium, a lot, and extremely bothered. For pediatric patients aged 4-7 years, symptoms experienced by the pediatric patient today were rated by the dyad on a 3-point Likert scale consisting of not at all bothered, medium, and extremely bothered.

The approach to co-SSPedi administration was standardized. The dyad was told where to find the instructions to complete co-SSPedi once they selected “start.” The instructions directed the dyad as follows. They determined which of them would read each symptom description out loud. After each symptom description was read, the pediatric patient first proposed a rating. The guardian then stated whether they agreed or disagreed, and if they disagreed, they were to propose a rating and explain the rationale for the difference. The dyad would then discuss until they came to consensus. This process was repeated for each statement and was entered into the Supportive care Prioritization, Assessment and Recommendations for Kids (SPARK) website, which houses the SSPedi instruments. A synonym list is available if either member of the dyad had difficulty understanding the meaning of a symptom description. The same instructions were provided for the first and second (if applicable) co-SSPedi administration. For both administrations, prompting was not given once the dyad began to read the instructions. The time required for co-SSPedi completion started before the dyad began to read the instructions and extended to completion of the last question; thus, time included designation of the reader, proposal of scores, and discussion where required.

The other instruments were proxy reported by the guardian after the dyad completed co-SSPedi. ChIMES is a reliable, valid, and responsive measure of oral mucositis (12). ChIMES results in the ChIMES score, which ranges from 0 to 23, and the total ChIMES percent, which ranges from 0 to 100. For both scores, higher numbers indicate worse mucositis. PeNAT is a reliable and valid measure of present nausea severity that ranges from 1 (no nausea) to 4 (worst nausea possible) (13). The FPS-R is a reliable and valid measure of pain intensity that ranges from 0 to 10. Higher numbers indicate worse pain (14,15). The global QoL scale ranges from 1 (best possible QoL) to 5 (worst possible QoL). The overall symptom visual categorical scale ranges from 1 (none) to 4 (severe).

Statistics

As with SSPedi, mini-SSPedi, and proxy-SSPedi, the total co-SSPedi score is composed of the sum of each symptom description’s Likert scale, resulting in a total score that ranges from 0 (no bothersome symptoms) to 60 (worst bothersome symptoms).

For psychometric evaluation, we examined reliability, construct validity, and responsiveness. The threshold criteria were derived from previously established recommendations (16). We examined all dyads and, as a subgroup analysis, evaluated dyads where pediatric patients were aged younger than 8 years. The evaluation of test-retest reliability included dyads in the more symptomatic group who reported no change in symptoms between day 1 and day 4. We calculated the intraclass correlation coefficient (ICC) between the 2 co-SSPedi total scores (day 1 and day 4); we concluded reliability if the observed ICC was at least 0.75. We concluded internal consistency if the observed Cronbach alpha was more than 0.7 (17).

The evaluation of construct validity examined discriminative and convergent validity. To evaluate discriminative or known-groups validity, we hypothesized that mean day 1 total co-SSPedi scores would be statistically significantly higher for dyads in the more symptomatic group compared with the less symptomatic group. We compared the baseline scores of the more symptomatic group vs the less symptomatic group using the independent student t test. To evaluate convergent validity, we hypothesized that there would be fair correlation (Spearman r ≥ 0.25) between the following measures: the mouth sores co-SSPedi item and total ChIMES percent; the nausea and vomiting co-SSPedi item and PeNAT; the pain co-SSPedi item and FPS-R; the total co-SSPedi score and global QoL scale; and the total co-SSPedi score and overall symptom scale. The 95% confidence intervals (CIs) around the Spearman r values were obtained through bootstrapping 1000 iterations.

To evaluate the responsiveness of co-SSPedi, the day 1 and day 4 co-SSPedi scores in the more symptomatic group where dyads reported symptoms to be much worse or much better (pediatric patients aged 8-18 years, anchors of the 5-point scale) or worse or better (pediatric patients aged 4-7 years, anchors of the 3-point scale) were examined. We compared these scores using the paired student t test after accounting for the difference in direction on the global symptom change scale.

The sample size was based on the calculation for test-retest reliability. We assumed the ICC under the null hypothesis was 0.5 and under the alternative hypothesis was 0.75, an α of 0.05 and a β of 0.10, which provides a 95% confidence interval that excludes 0.75. With these assumptions, we needed 50 dyads who reported no change in symptoms between days 1 and 4 (18,19). Assuming that 15%-20% of dyads would provide a day 4 assessment and would report no change in symptoms, 300 dyads in the more symptomatic group were targeted. The number of dyads in the less symptomatic group was based on known-groups validation. Assuming a minimal clinically important difference of 5 points, standard deviation of 15, α of 0.05, and 10% missing values, enrollment of 300 dyads in the more symptomatic group and 200 dyads in the less symptomatic group would provide 93% power. Thus, the total targeted sample size was 300 dyads in the more symptomatic group and 200 dyads in the less symptomatic group, for 500 total dyads. Analyses were performed using R studio version 3.6.1, The R Foundation for Statistical Computing.

Results

Between June 15, 2021, and April 26, 2023, 612 dyads were assessed for eligibility, and 501 were enrolled; 301 were in the more symptomatic group, and 200 were in the less symptomatic group. The less symptomatic group was comprised of 33 patients with acute lymphoblastic leukemia in maintenance and 167 patients who had completed cancer therapy. The flow diagram of dyad identification, enrollment, and study participation is shown as Figure 1. For dyads enrolled to the more symptomatic group, the proportion who completed the day 4 assessment within the prespecified window (± 1 day) was 282 of 301 (93.7%). Reasons for missed day 4 assessments included early discharge (n = 16) and child being too unwell to participate (n = 3). The same guardian participated in the day 1 and day 4 assessment in 256 of 282 (90.8%).

Figure 1.

Flow diagram of participant identification, enrollment, and study participation. The figure illustrates the flow diagram of participant identification, enrollment and study participation, and the reasons for exclusion.

Demographic characteristics of dyads stratified by the more symptomatic group and the less symptomatic group are shown in Table 1. Overall, the median patient age at enrollment was 10.6  (range = 4.0-18.8) years, and median time from diagnosis was 18 (range = 0-212) months. The number of participants who were aged 4-7 years was 165 (32.9%). The participating guardian was the mother in 363 (72.5%) of dyads.

Table 1.

Dyad demographic characteristics stratified by more and less symptomatic groups

Characteristic	Total, No. (%) (n = 501)	More symptomatic group, No. (%) (n = 301)	Less symptomatic group, No. (%) (n = 200)
Child characteristics
Sex
Female	218 (43.5)	134 (44.5)	84 (42.0)
Male	283 (56.5)	167 (55.5)	116 (58.0)
Age, y
4-7	165 (32.9)	110 (36.5)	55 (27.5)
8-10	96 (19.2)	44 (14.6)	52 (26.0)
11-14	140 (27.9)	90 (29.9)	50 (25.0)
15-18	100 (20.0)	57 (18.9)	43 (21.5)
Race
Aboriginal	5 (1.0)	5 (1.7)	0 (0.0)
Arab, West Asian	27 (5.4)	16 (5.3)	11 (5.5)
Asian	125 (25.0)	69 (22.9)	56 (28.0)
Black	27 (5.4)	21 (7.0)	6 (3.0)
Latin American	12 (2.4)	7 (2.3)	5 (2.5)
White	236 (47.1)	145 (48.2)	91 (45.5)
Other, mixed ethnicity	60 (12.0)	31 (10.3)	29 (14.5)
Unknown, missing	2 (0.4)	1 (0.3)	1 (0.5)
Prefer not to say	7 (1.4)	6 (2.0)	1 (0.5)
Diagnosis
Leukemia	260 (51.9)	124 (41.2)	136 (68.0)
Lymphoma	56 (11.2)	31 (10.3)	25 (12.5)
Solid tumor	113 (22.6)	74 (24.6)	39 (19.5)
Brain tumor	48 (9.6)	48 (15.9)	0 (0.0)
Other	24 (4.8)	24 (8.0)	0 (0.0)
Metastatic disease	78 (15.6)	52 (17.3)	26 (13.0)
Relapse	59 (11.8)	55 (18.3)	4 (2.0)
Active treatment	327 (65.3)	301 (100)	26 (13.0)
Treatments received
Chemotherapy	463 (92.4)	271 (90.0)	192 (96.0)
Surgery	123 (24.6)	90 (29.9)	33 (16.5)
Radiotherapy	80 (16.0)	55 (18.3)	25 (12.5)
Hematopoietic cell transplantation	39 (7.8)	39 (13.0)	0 (0.0)
Inpatient at interview	222 (44.3)	22 (73.8)	0 (0.0)
Reason for visit
Chemotherapy or transplant	220 (43.9)	217 (72.1)	3 (1.5)
Fever	24 (4.8)	24 (8.0)	0 (0.0)
In school	428 (85.4)	232 (77.1)	196 (98.0)
English as first language	418 (83.4)	254 (84.4)	164 (82.0)
Guardian characteristics
Sex
Female	374 (74.7)	227 (75.4)	147 (73.5)
Male	127 (25.3)	74 (24.6)	53 (26.5)
Relationship to patient
Father	125 (25.0)	72 (23.9)	53 (26.5)
Mother	363 (72.5)	219 (72.8)	144 (72.0)
Brother	2 (0.4)	2 (0.7)	0 (0.0)
Sister	2 (0.7)	1 (0.3)	1 (0.5)
Other	9 (1.8)	7 (2.3)	2 (1.0)
Married	427 (85.2)	252 (83.7)	175 (87.5)
College or university education	406 (81.0)	238 (79.1)	168 (84.0)
English as first language	297 (59.3)	181 (60.1)	116 (58.0)
House income >$60 000	277 (55.3)	156 (51.8)	121 (60.5)

Table 2 describes characteristics of co-SSPedi and QoL scores. The median day 1 co-SSPedi score in the more symptomatic group was 10, and the median day 1 co-SSPedi score in the less symptomatic group was 3. Among all dyads, the median time to complete co-SSPedi was less than 3 minutes. For dyads enrolled to the more symptomatic group, the day 4 global symptom change scale was the same (no change in symptoms) in 84 (27.9%) and much better or much worse (ages 8-18 years) or better or worse (ages 4-7) years in 104 (34.6%).

Table 2.

Characteristics of co-SSPedi scores and quality-of-life scores

Outcome measures	More symptomatic (n = 301)	Less symptomatic (n = 200)
Co-SSPedi
Median total co-SSPedi scores day 1 (IQR)	10 (6-16)	3 (1-6)
Mean total co-SSPedi scores day 1 (SD)	12.0 (7.8)	4.2 (4.2)
Median minutes completion day 1 (IQR)	2.9 (2.1-3.9)	2.3 (1.7-2.9)
Day 4 sample size, No.	282a	NA
Median total co-SSPedi scores day 4 (IQR)	10 (5-15)
Mean total co-SSPedi scores day 4 (SD)	10.7 (7.7)
Median minutes completion day 4 (IQR)	2.1 (1.5-2.9)
Children’s International Mucositis Evaluation Scale
Median ChIMES scores (IQR)	0 (0-1)	0 (0-0)
Median ChIMES percent (IQR)	0 (0-4.4)	0 (0-0)
Pediatric Nausea Assessment Tool, No. (%)
Nausea now
No nausea at all	212 (70.4)	193 (96.5)
A little bit nauseated	59 (19.6)	7 (3.5)
Even more nauseated	23 (7.6)	0 (0.0)
Nauseated a whole lot	7 (2.3)	0 (0.0)
Vomited yesterday or today	51 (16.9)	4 (2.0)
Faces Pain Scale-Revised, median (IQR)	0 (0-2)	0 (0-0)
Global Quality of Life Categorical Scale, median (IQR)	3 (2-3)	1 (1-2)
Overall symptom scale, median (IQR)	2 (2-3)	1 (1-1)
Global symptom change scale on day 4, No. (%)b		NA
Much better, better if aged 4-7 y	68 (22.6)	NA
A little better	53 (17.6)	NA
The same	84 (27.9)	NA
A little worse	41 (13.6)	NA
Much worse, worse if aged 4-7 y	36 (12.0)	NA
Missinga	19 (16.3)	NA

a

No day 4 assessment in 19 participants. CHIMES = Children’s International Mucositis Evaluation Scale; IQR = interquartile range; NA = not applicable; SSPedi = Symptom Screening in Pediatrics Tool.

b

Dyads where pediatric patients were aged 8-18 years completed a 5-point global symptom change scale. Dyads where pediatric patients were aged 4-7 years completed a 3-point global change scale.

Table 3 summarizes the evaluations of reliability, validity, and responsiveness. Among the entire cohort, test-retest reliability ICC was 0.85 (95% CI = 0.77 to 0.90), meeting the a priori established threshold of an ICC of at least 0.75. For internal consistency, total co-SSPedi Cronbach alpha on day 1 was 0.81 (95% CI = 0.78 to 0.83), which was at least 0.7 threshold.

Table 3.

Psychometric properties of proxy-SSPedi and mini-SSPedi^a

Property	Hypothesis	All patients		Patients aged younger than 8 years
		No.	Results	No.	Results
Reliability
Test-retest reliability	ICC ≥ 0.75 when comparing total co-SSPedi scores between days 1 and 4 in those who report no change in symptoms	84	ICC = 0.85 95% CI = 0.77 to 0.90	40	ICC = 0.78 95% CI = 0.59 to 0.89
Internal consistency	Total co-SSPedi scores—Cronbach alpha ≥ 0.7 on day 1	501	alpha = 0.81 95% CI = 0.78 to 0.83	165	alpha = 0.77 95% CI = 0.71 to 0.82
Construct validity
Known groups validity	Total co-SSPedi score higher for more symptomatic vs less symptomatic groups	501	Mean diff = 7.8 95% CI = 6.7 to 8.8 P < .0001	165	Mean diff = 7.7 95% CI = 5.9 to 9.6 P < .0001
Convergent validity	Mouth soreness co-SSPedi item fairly correlated with total ChIMES percent, r = ≥0.25-0.50	501	Spearman r = 0.56 95% CI = 0.46 to 0.64	165	Spearman r = 0.49 95% CI = 0.29 to 0.64
Convergent validity	Nausea and vomiting co-SSPedi item fairly correlated with PeNAT, r = ≥0.25-0.50	501	Spearman r = 0.67 95% CI = 0.59 to 0.74	165	Spearman r = 0.75 95% CI = 0.62 to 0.86
Convergent validity	Pain co-SSPedi item fairly correlated with FPS-R, r = ≥0.25-0.50	501	Spearman r = 0.68 95% CI = 0.61 to 0.74	165	Spearman r = 0.71 95% CI = 0.58 to 0.81
Convergent validity	Total co-SSPedi score fairly correlated with global QoL scale, r = ≥0.25-0.50	501	Spearman r = 0.58 95% CI = 0.51 to 0.64	165	Spearman r = 0.53 95% CI = 0.40 to 0.63
Convergent validity	Total co-SSPedi score fairly correlated with overall symptom scale, r = ≥0.25-0.50	501	Spearman r = 0.69 95% CI = 0.64 to 0.74	165	Spearman r = 0.64 95% CI = 0.54 to 0.73
Responsiveness	Difference between day 1 and day 4 co-SSPedi scores among dyads who said the child was much better or much worse (aged 8-18 years) or better or worse (aged 4-7 years)	104	Mean diff = 4.5 95% CI = 3.1 to 5.8 P < .0001	63	Mean diff = 3.7 95% CI = 2.0 to 5.5 P < .0001

a

Statistical tests were Spearman correlation coefficients for convergent validation, independent Student t test for known groups construct validation, and paired student t test for responsiveness. CI = confidence interval; co-SSPedi = co-Symptom Screening in Pediatrics Tool; diff = difference; FPS-R = Faces Pain Scale–Revised; ICC = intraclass correlation coefficient; PeNAT = Pediatric Nausea Assessment Tool; QoL = quality of life.

For known groups validation, the mean difference in total co-SSPedi scores between the more symptomatic and less symptomatic groups was 7.8 (95% CI = 6.7 to 8.8; P < .0001). For convergent validation, all hypothesized relationships were demonstrated. For responsiveness, the mean difference between day 1 and day 4 co-SSPedi scores among dyads who said the child was much better or much worse (ages 8-18 years) or better or worse (ages 4-7 years) was 4.5 (95% CI = 3.1 to 5.8; P < .0001).

As a subgroup analysis, we repeated all psychometric evaluations among dyads where the pediatric patient was aged younger than 8 years. Results were generally similar to the overall cohort. Internal consistency, construct validity, and responsiveness met established thresholds for validation. In terms of test retest reliability, the ICC was 0.78 (95% CI = 0.59 to 0.89), whereby the lower limit of the 95% confidence interval on the ICC was less than 0.75.

Discussion

We found that co-SSPedi, a novel approach to dyadic symptom assessment that includes both pediatric patient and guardian perspectives, was feasible and quick to administer. Co-SSPedi demonstrated test-retest reliability, internal consistency, known-groups validity, convergent validity, and responsiveness. Consequently, co-SSPedi may now be used for research or clinical care.

We evaluated the psychometric properties of co-SSPedi among dyads where pediatric patients were aged younger than 8 years separately. All evaluation thresholds were met with the exception of test-retest reliability in which the point estimate was at least 0.75, but the lower 95% confidence interval bound fell below the threshold. It is possible that very young children have difficulty recalling their initial symptoms and that their input into the global symptom change scale worsened the measure of test-retest reliability. Using a guardian-reported global symptom change scale may have been a preferable approach for the younger age group.

Co-SSPedi appears to be a valid approach to measure symptom bother. However, the question of where co-SSPedi fits into the array of available SSPedi measures for clinical and research use remains. For children aged younger than 11 or 12 years, co-SSPedi likely would be an appropriate approach to symptom assessment recognizing that the guardian respondent may alternate depending on family structure. It was encouraging that our study maintained reliability even though some day 4 assessments used a different guardian compared with the day 1 assessments, suggesting that this approach may be appropriate. However, for older pediatric patients, particularly for those who are feeling well and are participating in typical activities, co-SSPedi may not be optimal and SSPedi (self-report) may be preferred. Nonetheless, in our previous study, we found some older patients preferred co-SSPedi, particularly when they were tired or were clinically unwell (11). Because pediatric patients undergoing cancer treatment may have rapidly changing acuity of illness, the ideal approach should provide flexibility in whether the patient would prefer co-SSPedi, SSPedi/mini-SSPedi, or proxy-SSPedi. Although these approaches would not be considered interchangeable, it would provide a pragmatic approach to symptom reporting in clinical practice although may not be ideal for research because perspectives may be systematically different. We anticipate that proxy-SSPedi will be less commonly used when co-SSPedi is available.

It is important to note that co-SSPedi and the other SSPedi instruments are symptom screening and assessment tools. They are meant to capture a broad array of symptoms that commonly affect pediatric patients receiving cancer treatments. They are not meant to be a measure of QoL such as the PedsQL instruments, which can measure generic or disease-specific QoL (20).

The strengths of this study were the multicenter design and the evaluation of a novel pragmatic approach to symptom reporting in pediatric patients. Another strength was the large sample size and heterogeneous patient characteristics. However, this study was limited by its conduct only in English.

In summary, co-SSPedi is a novel approach to dyadic symptom assessment that is reliable, valid, and responsive in pediatric patients aged 4-18 years. Co-SSPedi can be used for clinical and research purposes. Future work should focus on determining mechanisms that allow patients to choose their preferred modality of SSPedi completion dynamically over time.

Data availability

All data generated and analyzed is available on reasonable request.

Author contributions

Deborah Tomlinson, MN, RN (Conceptualization; Data curation; Funding acquisition; Methodology; Project administration; Writing—original draft; Writing—review & editing), L Lee Dupuis, PhD, RPh (Conceptualization; Investigation; Methodology; Validation; Visualization; Writing—review & editing), David Dix, MBChB (Project administration; Supervision; Writing—review & editing), Nicole Crellin-Parsons, BMSc (Conceptualization; Data curation; Project administration; Validation; Writing—review & editing), Sadie Cook, BA (Data curation; Software; Supervision; Writing—review & editing), Ketan Kulkami, MD (Supervision; Writing—review & editing), Tal Schechter, MD (Data curation; Visualization; Writing—review & editing), George Tomlinson, PhD (Formal analysis; Writing—review & editing), and Lillian Sung, MD, PhD (Conceptualization; Data curation; Formal analysis; Funding acquisition; Investigation; Methodology; Project administration; Supervision; Validation; Visualization; Writing—original draft; Writing—review & editing).

Funding

The authors have declared no conflicts of interest. The study was supported by a grant from the Garron Family Cancer Centre.

Conflicts of interest

None declared.

Acknowledgements

LS is supported by the Canada Research Chair in pediatric oncology supportive care. The funder had no role in the design of the study; the collection, analysis, or interpretation of the data; or the writing of the manuscript and decision to submit it for publication.

Informed consent was obtained from all individual participants included in the study.

The protocol was approved by The Hospital for Sick Children, Research Ethics Board, and the Research Ethics Boards of all participating centers. All methods were carried out in accordance with relevant guidelines and regulations.