Study of Complementary and Integrative Interventions in Pediatric Patients with Serious Illness

Abstract

Context

Pediatric patients report using complementary and integrative health interventions (CHI) for symptom management.

Objective

Feasibility and initial effect sizes of CHI with patients at two children’s hospitals.

Methods

Using convenience sampling at two academic centers and accepting the wide age range of patients traditionally treated in children’s hospitals, we examined feasibility of CHI and outcomes of quality of life (QOL) and symptoms with validated surveys and two physiologic measures. A priori feasibility thresholds were 90% accrual rate and 60% completion of at least two surveys and one CHI session.

Results

Over 18 months we accrued 100 participants (Site 1, n=34; Site 2, n=66) who completed 811 assessments. Participants were aged 2–29 years (M=13.5, SD=5.6), 65% female, 23% from underrepresented populations, 52% with cancer versus other serious illness. Accrual rate was 94%, completion rate was 87%, acceptability was 96%. Ninety-nine participants received 191 total CHI sessions: acupuncture (39%), aromatherapy (35%), creative arts (20%), massage therapy (5%) and hypnosis (1%). After CHI heart rate decreased and symptom scores improved for anxiety, fatigue, nausea, pain, and sadness (Cohen’s d effect sizes 0.22–0.99). Adjusted mixed-effects models suggested that the Faces Scale scores improved over time (b= −0.19, p<.01).

Conclusion

Prospective two-site data collection in relationship to CHI exceeded feasibility thresholds and was acceptable. When given the choice, CHI were popular and may have contributed to improved QOL immediately and longitudinally. These preliminary findings support further study of CHI for targeted symptoms in distinct populations with rigor.

Children with serious illness are at risk of symptoms that are difficult to manage and report turning to complementary health interventions (CHI) (1–4). Americans spend $34 billion on CHI to improve well-being, and 60% of children with serious illness and/or their parent/caregiver report looking to CHI for symptom management (1, 5). Evidence for safety with CHI is reported in pediatrics (6) and CHI have been associated with improved symptom control and quality of life (QOL) in children with serious illness (1, 7–9).

The feasibility and efficacy of CHI interventions in pediatrics have been demonstrated for acupuncture (10–13), aromatherapy (14, 15), hypnosis (16, 17), massage therapy (18–20) and creative arts therapy (21–23). Although the evidence in support of CHI is mounting, these options are not routinely accessible to patients treated at pediatric centers, nor are they universally reimbursed, regardless of payor. These gaps can be partly attributed to the lack of rigor and reproducibility in the extant literature (3, 24–29). Thus, it is necessary to investigate the feasibility of CHI in pediatric patients using standardized outcome measures and objective endpoints.

Two patient reported outcome (PRO) measures in pediatrics are endorsed for rigor by the National Institutes of Health (NIH): the Patient Reported Outcomes Measurement Information System (PROMIS) global health measure for QOL (30, 31) and the Pediatric Patient Reported Outcomes Common Terminology Criteria for Adverse Events (Ped PRO-CTCAE) for symptom assessment (32). To augment these measures with a tool that has an immediate recall time period and may better match the developmental stages of childhood, we have explored the use of the McGrath Faces Scale (Figure 1) as an expression of well-being (33). We found significant correlations between the Faces Scale and QOL cross-sectionally and longitudinally (22, 34). In addition, objective outcome measures such as physical posture (thoracic kyphosis) and heart rate responses could enhance the assessment of efficacy of CHI (16, 22, 34–36).

Figure 1.

Faces Scale

Choose the face that shows how you feel right now. Choose the face that shows how you feel deep down inside--not just how your face looks, but how you really feel inside. (McGrath, 1996)

Therefore, we aimed to determine the feasibility and acceptability of administering available CHI interventions while implementing NIH endorsed PRO measures and novel physiologic measures among patients at two children’s hospitals. The theoretical basis for this study included the Obesity-Related Behavioral Intervention Trials (ORBIT) consortium model (37), the Symptom Management Theory (SMT) (38, 39), and the concept of emotional embodiment (QOL is reflected by the physical body) (40–42). The study was informed by our preliminary work implementing CHI and using rigorous outcome measures in children and adolescents with serious illness (22, 34, 43–45). To assess feasibility for the wide range of patients treated at children’s hospitals, we purposefully included participants with any disease and the diverse developmental stages of infants, children, adolescents, and young adults. The primary aim was feasibility of an 90% recruitment rate and 60% completion of at least two outcome measures and one CHI session. Secondary aims were acceptability and exploration of effect sizes and outcome changes over time.

Methods

Design, Setting, Population

This study was a two-site prospective feasibility clinical trial of CHI (ClinicalTrials.gov#: NCT05594693) with 100 participants over 18 months. The setting was two tertiary children’s hospitals at academic medical centers. Site 1 was a large free-standing children’s hospital with 600 beds and an average of 30 infusion center patients/day. Site 2 was a medium-sized children’s hospital with 80 beds within a 576-bed university hospital and about 15 infusion patients/day. Site 1 served as the IRB of record for both sites. Each site had experience providing CHI and clinical space to implement CHI embedded in usual care.

Pediatric patients were referred for CHI by their primary medical team and enrolled in the study using convenience sampling. Inclusion criteria for these patients were ages 31 days to 30 years and literate in the languages available for the study surveys (currently English). For children unable to read, a caregiver proxy (e.g., parent or legal guardian) was used. Exclusion criteria were physical or cognitive inability to complete the surveys, and non-consenting participants/caregivers or providers. After the study team confirmed eligibility, informed consent/assent was discussed and obtained.

Intervention

The CHI available at Site 1 were acupuncture, aromatherapy, and creative arts therapy. At Site 2, the CHI offered were acupuncture, aromatherapy, massage therapy, hypnosis, and creative arts interventions. Anecdotally, CHI were more commonly used in the oncology and chronic pain settings. The accessible CHI varied by day and time, so study participants had differing choices of CHI. CHI were provided by certified, credentialed practitioners and session duration was 10–60 minutes (Table 1). Patients could receive the CHI regardless of their participation in the study. A clinical acupuncture procedure consent was completed separately by the acupuncturist. Study participants were asked to complete at least one session of CHI and up to five sessions were allowed during the data collection period.

Table 1.

Interventions

	Site Available	Provider	Length	Location/Process	Equipment
Acupuncture	1 and 2	Licensed, credentialed acupuncturist or anesthesiologist	5–60 minutes	Hospital, clinic room, infusion center	Approved medical acupuncture needles
Hypnosis	1 and 2	Practitioners trained in hypnosis for symptoms	20 minutes	Infusion center	Reading a prepared script
Massage Therapy	2	Licensed, credentialed massage therapists, trained in pediatric medical massage	30–60 minutes	Hospital, infusion center	Gloves, lotion
Aromatherapy	1 and 2	Allied health professional trained in Soothing Scents® administration	10–20 minutes	Hospital, infusion center	Hospital approved aromatherapy (Soothing Scents®) 5 aromas with information to guide matching aromas to symptoms
Creative Arts	1 and 2	Licensed clinical therapist certified in art, music, dance/movement, or yoga therapies (site 1) Trained hospital approved art facilitators (site 2)	60 minutes	Hospital, infusion center	Art supplies, musical instrument, yoga mat, etc.

Outcome Measures

Study measures were administered electronically by QR code or email to the electronic devices of the participant or proxy using the Research Electronic Data Capture (REDCap) system (46, 47). At baseline, participants completed a demographics survey. To decrease the burden on participants and to streamline data analysis, we clustered the diagnosis categories on the demographic form by the two clinical areas we predicted most CHI would take place (Cancer and Chronic Pain) with an option for Other; followed by free text). The form also included biologic variables (age, sex), race/ethnicity, and highest level of parental education (as a surrogate of socioeconomic status). The longitudinal measures of QOL and posture were administered at baseline and monthly for up to 6 months. The short-term measures of symptom assessment and heart rate were administered before and after each CHI session (Figure 2). The Faces Scale was administered at all time points. Data were entered directly into the secure REDCap database. Participants received a $10 electronic gift card after completion of each time point.

Figure 2.

Study Flow Diagram

Feasibility and Acceptability

Number and type of CHI sessions were collected for each participant. Field notes were categorized by study design, technology, or clinical circumstances. Acceptability was measured by a five-point Likert scale questionnaire at each time point for satisfaction with the intervention and burden of completing measures.

Quality of Life (QOL)

QOL was measured with the Pediatric PROMIS-Global Health-7. The PROMIS Pediatric Measures have been validated longitudinally for multiple pediatric chronic and acute disorders with acceptable reliability (30, 31, 48, 49). The Global Health-7 tool has seven questions with a one-month recall, valid for ≥ 8 years for self-report. For children aged between 1–7 years, the Early Child caregiver proxy is used (31, 48, 50). For participants ≥ 18 years, the Adult PROMIS Global Health-10 was shortened to the 7 items corresponding with the Pediatric tool items. This subset of 7 items from Adult PROMIS were summed before using the appropriate conversion table for scoring (51). The reported minimally important difference (MID) to reflect a change in QOL is 2 points (52).

Symptom Assessment

Symptoms were assessed before and after each CHI session with the Ped PRO-CTCAE, a 4-point patient-reported symptom measurement system valid for children ages ≥7 years (caregiver-proxy for children ≥ 7 years unable to read). It has strong convergent and discriminant validity, as well as responsiveness over time (32, 53), with translation for Spanish in progress. The core CTCAE terms used for this study were based on the reported priority symptoms of pain, nausea, fatigue, anxiety, and depression (54, 55). Although the recall for this instrument is 7 days, we educated the participants to use a shorter recall (one hour) for the post-intervention time point. We omitted this tool for children ≤ 6 years of age.

McGrath Faces Scale

The Faces Scale developed by McGrath is comprised of nine faces in order of happy to upset with lower numbers associated with a happier face. It is validated for children aged 3–17 years, with consistent rating by children over age 5 years for acute, recurrent, and chronic pain, regardless of age, gender, or health status (33, 56). We, and others, have used this tool to evaluate emotional response (43, 57), and have documented a correlation with QOL cross-sectionally and longitudinally (22, 34). The MID on the Faces Scale has been estimated as one face (58, 59).

Physiologic Measures

The two objective physiologic measures used were heart rate and posture. Heart rate has shown meaningful clinical changes with CHI (36) and was obtained from the electronic medical record or palpation pre/post CHI sessions. Posture has been correlated with QOL with an immediate and prolonged effect in relation to CHI (22, 34). The thoracic kyphosis angle was measured with the non-invasive gravity-dependent inclinometer (Isomed Inc.) comparing readings between spinous processes of thoracic vertebrae 1 and 12 (60). This instrument has demonstrated acceptable intra-rater reliability and validity with measurement error of 1.7 degrees (61–64). Posture was measured at baseline and the monthly timepoints. A larger thoracic kyphosis angle indicates worse posture. The mean thoracic kyphosis angle for the general population in pediatrics is about 36 degrees (65).

Analysis

We planned to estimate a feasible sample size based on the primary aim of actual accrual per CHI session per month. Power calculations revealed a sample size of 98 was needed to detect the a priori threshold of a 90% accrual rate (with a lower bound of the normal-based 95% confidence interval of 0.85). We defined accrual rate as the number of participants who signed consent after being approached about the study.

Demographic/clinical characteristics were summarized with descriptive statistics. Specifically for diagnosis category, we originally used the three levels of disease type (Cancer, Chronic Pain, Other), then collapsed to two levels (Cancer, Non-Cancer), then explored the free text. We calculated feasibility with recruitment and follow-up percentages for proportion of adherence. The a priori feasibility thresholds were a 90% accrual rate and a 60% completion rate of at least one CHI session and two data time points. The acceptability threshold was 80%. Amount of missing data was calculated by the percentage of measures completed at each time point. We also examined means, variances, covariances, and effect sizes (e.g. Cohen’s d) for pre/post CHI outcomes. Effect sizes were estimated using mixed-effects models including pre/post outcome measures which accounted for nesting within person (i.e., that the same patient could have completed multiple CHI). Cohen’s d was defined using the estimated pre/post change in the numerator and the sum of the random error terms in the denominator. Cohen’s d effect size ranges were classified as small (d  = 0.2–0.4), medium (d = 0.5–0.7), and large (d ≥ 0.8) (66). We used mixed-effects models to examine the change in QOL (PROMIS and Faces Scale) and posture over time. The first model included covariates: number of CHI sessions, gender, race (white/non-Hispanic versus non-White/Hispanic/mixed/missing), PROMIS scale type (as a surrogate for age: early child 1– 7 years, child 8–17 years, adult ≥18 years), diagnosis (cancer, chronic pain, or other serious illness), parental education level (college versus no college), and study site (Site 1 versus Site 2). The second model included the same covariates and included an interaction term between time and CHI sessions to test whether the magnitude of changes across time differed depending on the number of CHI sessions for each patient. Field notes kept by the principal investigator and research coordinator were summarized.

Results

Feasibility/Acceptability

Over 18 months we accrued an average of five participants per month for a total of 100 participants (Site 1, n=34; Site 2, n=66) who completed 811 assessments. Patient participants were aged 2–29 years (M=13.5, SD=5.6), most identified as female (65%), and 23% reported as being from underrepresented populations (Table 2). Forty-one participants were ≤ 12 years, 42 were in the adolescent age group (13–17 years) and 17 were ≥ 18 years. More than half of the participants reported the primary diagnosis as cancer (54%), about a quarter reported chronic pain as the reason for treatment (21%), and the rest of the sample reported a mix of diverse subspecialty disorders (27%) (Table 3). Caregiver proxies (for children ≤ 7 years or those unable to read) completed 80 of the 811 (10%) surveys. Accrual rate (patients approached who signed consent) was 94%. Those who declined cited lack of interest (n= 5) or feeling too sick (n=1) (Figure 3). Completion rate of ≥2 surveys and ≥1 intervention session was 87%. Participants engaged in 191 CHI sessions in varied amounts: 41 participants completed one session, 32 completed two, 9 completed three, 6 completed four, and 7 completed five CHI sessions. One participant was enrolled, responded to surveys, but declined CHI sessions. The most common CHI sessions were acupuncture, aromatherapy, and creative arts (Figure 4). Acceptability of the CHI was 96% (at least “somewhat satisfied”) and no participants reported the surveys were difficult to complete. Field notes revealed challenges attributed to study design were the dependence on research coordinator staffing. The challenges regarding technology were noted to be electronic data capture errors and difficulty with REDCap texting to mobile phone carriers. Challenges with clinical circumstances were changes in treatment, loss of interest, and hospice enrollment or death.

Table 2.

Demographics

	Overall (N=100)
Site
Colorado	34
Oregon	66
Age
Mean (SD)	13.3 (5.54)
Median [Min, Max]	14.0 [2.00, 29.0]
Diagnosis
Cancer	52
Non-Cancer	48
Gender
Female	64
Male	31
Missing	5
Race
White/Caucasian	75
Black/African American	4
Asian/Native Hawaiian/Other Pacific Islander	0
Native American/Alaska Native	0
More than One Race	11
Other	5
Unknown/Not Reported/Missing	9
Ethnicity
Hispanic	20
Non-Hispanic	76
Unknown/Not Reported/Missing	4
Parent Education
No College	23
Some to Complete College	41
Graduate School	29
Decline to Answer or Missing	7

Table 3.

Patient Reported Disease Type

Disease Type	Diagnostic Category	Participant Free-Text Description	N
Cancer			51
	Hematologic Malignancy	Leukemia Lymphoma	15
	Solid Tumor	Neuroblastoma Osteosarcoma Langerhans Cell Histiocytosis Malignant Pleural Effusion	11
	Central Nervous System Tumor	Diffuse Midline Glioma Pilocytic Astrocytoma Intracranial Germ Cell Tumor	4
	Bone Marrow Transplant	Severe Aplastic Anemia NOS	2
	NOS		19
Chronic Pain
		Chronic fatigue Chronic migraine Herniated disc Joint pain NOS	21
Other			27
	Hematology	Beta Thalassemia Myelodysplastic Syndrome Aplastic anemia Hypogammaglobulinemia Low Iron	6
	Pulmonary	Pulmonary Hypertension	1
	Neurology	Postural Orthostatic Tachycardia Syndrome (POTS)* Pediatric Acute-onset Neuropsychiatric Syndrome (PANS)* Cerebral Palsy Ehlers Danlos Syndrome* Tuberous Sclerosis	7
	Gastrointestinal	Crohn’s Disease Ulcerative Colitis Inflammatory Bowel Disease Pancreatitis* Kidney Failure	10
	NOS		6

Note: NOS = not otherwise specified; N = number of participants

^*Reported within the Chronic Pain category but recategorized here by organ system.

Figure 3.

Consort Diagram

Figure 4.

Most of the participants participated in acupuncture, aromatherapy, and creative arts.

Missing Outcome Data

Missing data were most prevalent for the physiologic measures (heart rate 41% and posture 51%) and least prevalent for the Faces Scale at monthly time points (11% missing). Because the Faces Scale had the least missing data longitudinally, we used that measure to examine missingness for the pre/post CHI timepoints. At the post-CHI time point, the Faces Scale was missing for 26% of the sessions compared to 0% missing pre-CHI. We also had the following percentages of missingness in the optional free-text description that followed the three diagnosis categories: Cancer (35% missing), Chronic Pain (10% missing), Other (22% missing).

Longitudinal Outcomes

In the fully adjusted mixed-effects models, Faces Scale scores significantly decreased over time (b = −0.19, p < 0.01), approximately 1 face over the 5-months (Figure 5). The minimally clinical important difference (MCID) for the Faces measure is 1 face. The interaction between time and CHI sessions was not statistically significant. Neither QOL (measured by PROMIS) or posture changed significantly over time and the interaction between CHI sessions and time was nonsignificant.

Figure 5.

The Faces Scale improved over time (lower number is a happier face) with approximately 1 face change (meeting the minimally important difference).

Adjusted for by covariates: site, age, diagnosis, gender, race/ethnicity, and parent education. Minimally clinical important difference (MCID): 1 face, p < 0.01

Immediate Outcomes after CHI Sessions

Because the number of participants varied among the interventions, we only examined specific interventions in which there were at least 20 CHI sessions received to have sufficient statistical precision (Table 4). After ≥ 20 sessions of acupuncture, we found small-to-medium effect sizes (0.25–0.47) for reduction in frequency of fatigue, pain, and sadness, and less interference of daily life for pain and sadness. Additionally, the Faces Scale showed more happiness with a medium effect size (0.69) and heart rate decreased with a large effect size (0.98). After ≥ 20 sessions of aromatherapy, small effect sizes (0.22–0.40) were seen for decreased frequency of anxiety and pain, and interference of daily life for nausea and sadness. Improvement with a medium effect size (0.61–0.71) after aromatherapy was seen both on the Faces Scale and with a decrease in heart rate. Finally, with ≥ 20 sessions of creative arts, medium-to-large effect sizes (0.70–1.28) occurred for improvement in frequency for anxiety, pain, and sadness, and more happiness on the Faces Scale. Although massage therapy was only received for 10 sessions, scores showed improvement in anxiety, fatigue, and nausea (data not shown). Heart rate decreased 10.07 beats per minute [95% CI (6.82, 13.32)] on average after CHI sessions with the effect most pronounced after acupuncture.

Table 4.

Outcome variables with Cohen’s d ≥ 0.2 with n=20+ sessions of CHI

Variable	Acupuncture (n=75)	Aromatherapy (n=67)	Creative Arts (n=38)
Anxiety
Frequency		−0.35 (n=41)	−0.86 (n=21)
Fatigue
Frequency	−0.30 (n=44)
Interference	−0.26 (n=43)
Nausea
Interference		−0.23 (n=20)
Pain
Frequency	−0.47 (n=46)		−0.99 (n=22)
Interference		−0.22 (n=30)
Sadness
Frequency	−0.40 (n=48)	−0.40 (n=41)	−0.70 (n=22)
Interference	−0.25 (n=33)
Faces Scale	−0.69 (n=54)	−0.61 (n=47)	−1.28 (n=29)
Heart Rate	−0.98 (n=46)	−0.71 (n=28)

Discussion

We exceeded our a priori feasibility thresholds among 100 participants at two children’s hospitals. We found that patients with pediatric serious illnesses can participate in CHI and complete measures both in response to the CHI and longitudinally. Our average accrual of five participants/month provides an estimate accrual rate for future studies. Our 94% accrual rate reflects the interest in CHI for patients at children’s hospitals. Our 87% completion rate of study measures shows that pediatric patients undergoing treatment at specialty centers can complete validated surveys with rigor. Our 96% acceptance rate exemplifies the popularity of the CHI and the lack of burden of the measures for patients at pediatric centers. While our study was not powered to measure changes in outcomes, preliminary analysis showed improvement in all the measured symptoms with varied effect sizes. Additionally, QOL measured by the Faces Scale improved significantly over time.

Because the American Academy of Pediatrics and others endorse the need to study CHI in all children with serious illness (4, 67, 68), we included any patient receiving treatment at a children’s hospital. Even with convenience sampling, we had an almost equal distribution between participants with cancer and those with other serious illnesses which supports a future two-group study design. Our experience at two sites is that pediatric infusion centers have a patient census that comprises about 40% of patients with diagnoses other than cancer. This number may be increasing as intravenous targeted genetic and molecular therapies for many chronic diseases of childhood are developed.

Feasibility/Acceptability

Our findings are supported by other pilot and feasibility studies of CHI (36, 69–76). Although others have shown feasibility, some of the limitations in rigor are single site studies and small sample sizes. Our accrual of 100 participants compares favorably to a recent meta-analysis of randomized controlled trials of CHI in children with cancer in which only two of 18 identified studies had 100 or more participants (77). With smaller populations in pediatric diseases compared to adults, the results achieved with our two-site study design demonstrate a potential strategy for accrual of large sample sizes required for intervention studies. Another limitation in the extant literature of CHI is the variation in outcome measures rendering comparison of outcomes difficult. Our findings show that children can complete two NIH standardized outcome measures without excessive burden, enabling comparison across studies. The CHI available varied by day and site which influenced how many sessions of each CHI occurred. Further study is warranted to determine whether specific CHI may target certain symptoms or be favored by patients in particular age groups, cultures, or diagnosis types.

Quality of Life

Although changes in QOL measured by the PROMIS scale were not significant, QOL measured by the Faces Scale scores improved significantly for all participants over time. Other published evidence is mixed whether QOL improves due to normalization of the treatment center and resolution of disease and symptoms; or if QOL worsens over time with disease progression and the accumulation of treatment toxicities (78–81). More focused study designs could help to establish causal connections between specific types of CHI and QOL, particularly in the vulnerable and relatively small population of children with serious illness. Replicating our prior studies, the Faces Scale continues to be a simple, readily available measure for this population that is sensitive to change.

CHI and Symptoms

We found that the five priority symptoms of pain, fatigue, nausea, anxiety, and mood may improve in relation to CHI. The medium to large effect sizes for improvement in anxiety, pain, and sadness after creative arts supports our current NIH funded feasibility randomized controlled trial of creative arts in children with cancer (ClinicalTrials.gov #: NCT06342453). Nausea only had improvement associated with aromatherapy with a small effect size which may reflect the difficulty in treating nausea (82). Other researchers have shown improvement in nausea with aromatherapy and acupuncture (77). Improved fatigue scores were associated with acupuncture which has been shown in adults and shows promise for this intractable symptom in pediatrics (50).

Biologic Measures

Physiologic measures and biomarkers offer intriguing objective indicators of symptom burden and QOL (83, 84). We anticipated that heart rate would be a vital sign that would be readily attainable in our population. Indeed, we found a 10 beats per minute decrease after a CHI session. The large effect size for improvement in heart rate after acupuncture perhaps represents embodiment of this physical intervention. In our prior studies more upright posture was correlated with better QOL in the context of creative arts therapy (22, 34), a finding replicated in the current study with multiple CHI. The amount and type of missing data and field notes showed the difficulty in obtaining physical measures in real time. Although both posture and heart rate measures may be valid surrogates of symptoms and quality of life, further work is needed on reliability and controlling for confounding variables. Perhaps biomarkers in the serum and saliva are more promising for measuring QOL in this population as shown recently in a pilot study by Montgomery et al (83).

Limitations

Convenience sampling limits our findings in terms of generalizability and selection bias. The wide inclusion criteria allowed for a heterogenous population which limits specificity. We did not collect location of treatment (inpatient versus outpatient) which may have influenced QOL scores. We did collect broad categories of diagnosis (Cancer, Chronic Pain, Other), but the free-text box for diagnosis was optional, so we lack specificity regarding disease type. Because of the variability in the number of CHI sessions received, these data are influenced by level of exposure to CHI. Although only 10% of survey responses were given by caregiver proxies, we did not collect their demographic data. Finally, our approach to compare PROMIS QOL scores across age groups could limit the validity of the results for the young adult participants.

Conclusions

Prospective two-site data collection in relationship to CHI exceeded feasibility thresholds and was acceptable to participants with serious illness treated at children’s hospitals. When given the choice, CHI were popular and reduced the frequency and interference in daily life of pain, nausea, anxiety, fatigue, and sadness with varying effect sizes. QOL measured by the Faces Scale improved over time. Objective physiologic measures also reflected change related to CHI and hold promise as surrogates of symptoms and QOL. The convenience sampling highlighted the natural comparison groups of participants with and without cancer at children’s hospitals and pediatric infusion centers, which is ripe for further study. Our next steps are to measure changes in specific symptoms with prospective multi-site CHI intervention studies.

Key Message.

This article describes the feasibility of a prospective study of complementary and integrative health interventions with standardized outcome measures in patients at two children’s hospitals. Improvement in symptoms with varied effect sizes supports further study targeting interventions to symptoms.