Caregiver’s Kitchen: A Mixed Methods Feasibility Study of a Culinary Medicine and Coaching Intervention for Caregivers of Children With Cancer

Study Population and Recruitment

A caregiver-child dyad was eligible if the child was receiving treatment for cancer, was 4-21 years old, consumed ≥ 50% of nutrition orally, spoke English, and received approval from their oncologist and Registered Dietitian Nutritionist (RDN). The broad age range was selected to reflect the pediatric and adolescent/young adult oncology populations commonly treated within the participating clinics and to evaluate feasibility across a heterogenous group of diagnoses and developmental stages representative of real-world supportive care delivery. Children younger than 4 years old were excluded because feeding and nutrition needs in very young children differ developmentally and frequently require more specialized feeding approaches and caregiver support strategies. Children receiving bone marrow transplantation or active treatment for a brain tumor were excluded. Caregivers were eligible if they were a parent or legal guardian, English-speaking, ≥18 years old, had access to an internet-enabled device computer or internet-enabled device, and could provide informed consent. The study research coordinator and clinicians identified eligible families, medical records were reviewed, and interested caregivers provided consent for themselves and their child; child assent was obtained when appropriate. Although only caregivers participated in the intervention, both dyad members were considered participants because patient data were extracted from medical records.

This study was designed as an early-phase feasibility trial and was not powered to detect intervention efficacy. Ten caregiver-child dyads (N=20 participants) were enrolled using convenience sampling. The target same size was informed by methodological recommendations for pilot and feasibility studies, which suggest that approximately 10-12 participants per group can provide adequate precision for estimating key feasibility parameters needed to inform future trial design while minimizing unnecessary participant burden.^23-25 Consistent with guidance for feasibility research, sample size considerations were based on the ability to evaluate predefined progression criteria, assess intervention acceptability and implementation challenges, and obtain participant feedback to support refinement of the intervention and study procedures, rather than on formal hypothesis-testing calculations. ²⁴

Intervention

Caregiver’s Kitchen was an eight-week, remote culinary medicine intervention designed to enhance caregivers’ knowledge of whole-food nutrition, preparedness for caregiving, and self-efficacy in managing treatment-related side effects. ²¹ The intervention was formerly titled “Let’s Cook Together” in the published protocol; the program has since been renamed “Caregiver’s Kitchen” to better reflect its focus and target population. The program included four live, interactive culinary workshops co-led by Medical Chef Educator and a RDN via Microsoft Teams, and four bi-weekly RDN-led caregiver coaching calls.

Conceptually, Caregiver’s Kitchen was designed to strengthen caregivers’ practical capacity to respond to nutrition-related challenges that commonly arise during pediatric cancer treatment. The culinary workshops provided skill-based instruction, symptom-responsive recipes, and food safety guidance, while individualized RDN-led coaching calls supported goal setting, troubleshooting, and adaptation of strategies to the caregiver’s child and household context. The intervention was expected to most directly influence caregiver knowledge, cooking confidence, and self-efficacy for managing diet-related or diet-responsive side effects, including poor appetite, nausea, taste changes, constipation, diarrhea, and weight maintenance. We also explored whether improved confidence in this caregiving domain might contribute to broader caregiving preparedness and self-efficacy by increasing caregivers’ perceived ability to solve treatment-related challenges and reducing uncertainty around home-based nutrition care.

Workshops demonstrated low-cost, nutrient-dense recipes tailored to common treatment-related symptoms and nutrition concerns. Educational content emphasized practical approaches for increasing protein and caloric density during periods of poor appetite or weight loss risk, including meal fortification strategies, incorporation of protein-rich ingredients, smoothies and soft high-protein foods, and adapting textures and flavors to accommodate nausea and mucositis, for example. Additional guidance addressed food safety, food storage, and waste reduction. Supplementary Table 1 provides a condensed overview of workshop themes, educational content, and example recipes.

Thematic sessions addressed high-calorie cooking for appetite loss, strategies for managing nausea, and gastrointestinal health. Coaching calls (15–20 minutes) provided individualized goal setting, problem-solving, and support. Caregivers received educational materials, additional recipes in a cookbook developed by the integrative oncology team at CHOP, a culinary starter kit, and a $100 grocery gift card.

Data Collection

After consent, research coordinators enrolled dyads and extracted medical record data (race/ethnicity, age, gender, diagnosis, treatment, date of diagnosis, relapse history, and residential ZIP code). Feasibility and acceptability metrics were tracked continuously and collected at post-intervention. Exploratory outcomes were measured at baseline, week 8, and 3-month follow-up. The RDN(s) scheduled individualized coaching calls with caregivers through the Epic patient portal. The research coordinator sent reminders to caregivers for pre-scheduled culinary sessions and provided access to session recordings for caregivers who were unable to attend. Caregivers completed electronic surveys (via REDCap) and dietary assessments.

Primary Outcomes: Feasibility and Acceptability

Sample characteristics included caregiver and child demographics, caregiving context, and treatment information. Baseline open-ended questions captured caregivers’ motivations and goals. Feasibility and acceptability were assessed using predefined progression criteria aligned with Mellor et al. ²⁶

• Recruitment feasibility was defined as enrolling ≥5 caregiver-child dyads per month, a benchmark informed by prior recruitment experiences at participating clinical sites and intended to reflect realistic staff capacity and caregiver burden within pediatric oncology settings. Approximately 225 children receiving cancer treatment met the broad clinical inclusion criteria during the study period, providing a sufficient recruitment pool to support the target enrollment rate. Achieving this threshold was considered indicative of sufficient reach to support future scale-up.

Overall feasibility was defined as meeting at least four of five predefined criteria, supporting progression to a future definitive trial. To further explain the quantitative feasibility and acceptability data, open-ended questions in the post-intervention survey invited caregivers to describe the most helpful aspects of the intervention, whether anything changed in the family’s eating as a result of participating, and how the intervention could be improved. It is important to note that while the protocol specified in-depth caregiver interviews for the collection of qualitative data, we employed open-ended survey questions after finding interviews infeasible for caregivers managing active treatment demands.

Exploratory measures assessed caregiving preparedness (Preparedness for Caregiving Scale), ²⁹ caregiver self-efficacy for managing treatment side effects,^30,31 feeding behaviors (Behavioral Pediatrics Feeding Assessment Scale)^32-36 and dietary intake via ASA-24. These measures were selected to explore whether gains in a targeted caregiving domain, nutrition-related symptom management, might be reflected in broader constructs of caregiving preparedness and self-efficacy. Because validated pediatric oncology instruments specific to caregiver self-efficacy for nutrition-related symptom management are limited, these outcomes were interpreted cautiously and used primarily to inform measure selection for future trials.

ASA-24 dietary recalls were used to evaluate the feasibility of collecting self-administered dietary data in this population. Dietary intake was assessed at each time point using the National Cancer Institute’s Automated Self-Administered 24-hour Dietary Assessment Tool (ASA24), an electronic dietary recall instrument designed to estimate food intake, dietary patterns, macronutrient and micronutrient consumption, and total caloric intake. ³⁷ Caregivers completed recalls for themselves and served as proxy respondents for their child. Proxy reporting was used because younger children and children undergoing active cancer treatment may have difficulty independently recalling dietary intake, while caregivers are often closely involved in meal preparation, food selection, and monitoring intake. Dietary data collected via proxy should therefore be interpreted as an estimate of intake rather than a direct self-report measure. Because the study was not powered to detect dietary changes and recall completion declined over time, dietary data were examined only for feasibility indicators rather than for nutritional analysis. ³⁷ Full descriptions of all instruments and procedures are available in the published protocol paper. ²¹

Data Analysis

Consistent with the protocol, analyses focused on feasibility and acceptability. ²¹ Quantitative outcomes were summarized using descriptive statistics and interpreted against progression criteria. Qualitative data from open-ended responses were analyzed using rapid content analysis, involving structured summaries, inductive identification of key concepts, and grouping of recurrent ideas into descriptive categories appropriate for a small dataset. ³⁸ Consistent with an embedded convergent design, qualitative and quantitative results were integrated in a joint-display table aligning feasibility criteria with illustrative caregiver quotes. ²² Exploratory outcomes were analyzed descriptively to inform refinement of measures and estimate variability for future trial planning.

Participant Characteristics

Twenty participants (10 caregiver-child dyads) were enrolled. Patients included children across a broad developmental spectrum (ages 4-17 years, median age of 11 years), with representation from early childhood (<6 years, n=2), middle childhood (6-11 years, n=4), and adolescence (12-17 years, n=4). Six patients were female and most were white (90%). The sample also included a variety of cancer diagnoses and treatment experiences, reflecting the heterogeneous population commonly seen in pediatric oncology practice. Diagnoses included acute lymphoblastic leukemia (n=5), Hodgkin lymphoma (n=2), and one each with acute myeloid leukemia, rhabdomyosarcoma, and germ cell tumor. One dyad withdrew within two weeks due to scheduling challenges. Caregivers reported providing 5-24 hours of daily care. Two reported food insecurity and participation in federal nutrition assistance programs.

Feasibility and Acceptability

Counts, percentages and means for feasibility outcomes are summarize in Table 1. Three of five feasibility criteria were met. Recruitment occurred over four weeks, meeting the target rate with ten dyads. Recruitment was facilitated by close communication with oncology care teams, who helped identify families for whom study participation was clinically appropriate and unlikely to create excessive burden at that time. The total number of families discussed, approached, or declining participation was not systematically tracked. No harms or unintended effects were reported. Although intervention engagement was high, only 70% (or seven out of ten dyads) completed the post-intervention survey, just missing the progression criterion of ≥75%. Among retained participants, assessment completion rates varied by time point and assessment task, with 100% of caregivers completing baseline surveys and 89% completing dietary assessments; 78% (seven out of nine) completing post-intervention surveys and 44% completing dietary assessments; and 56% (or five out of nine) completing surveys at 3-months follow-up. Caregiver engagement was strong, with 80% attending at least three of four culinary workshops and 90% attending at least two of four individual coaching sessions. Program acceptability was high; mean scores across items assessing clarity, relevance, and usefulness was 4.35/5, exceeding the predefined threshold of ≥4.0. These findings support progression to a refined pilot study with reduced assessment burden and flexible delivery options.

OPEN IN VIEWER

Table 1.

Joint Display Integrating Feasibility and Acceptability Outcomes With Embedded Qualitative Caregiver Feedback

Outcome	Measure	Criteria	Caregiver’s kitchen	Qualitative insights (explanatory caregiver feedback)
Recruitment feasibility	Participants enrolled per month	1-2 dyads/week	2-3 dyads/week	Caregivers described strong motivation to participate due to challenges with symptom-related eating and desire for practical support: “Increased knowledge of ways to manage symptoms – specifically nausea, constipation, and diarrhea – utilizing food.”
Retention feasibility	% providing follow-up data	≥75%	70%	Illness severity, work schedules, and caregiving demands affected ability to complete later assessments: “It was tough to find the time to attend live sessions for me because of my job. The recorded sessions helped with that part though.”
				Length of each session… it was difficult to find a chunk of time to watch the entire recording if I missed a session due to scheduling conflicts.”
Assessment feasibility	% providing valid assessment data	≥95%	Baseline: 100%; Post-intervention: 78%; 3-month: 56%	Caregivers valued the surveys but noted time constraints: “When my daughter was too sick, I couldn’t put my full focus on the class.”
Intervention acceptability	Acceptability survey	Mean score ≥4	4.35/5	Coaching supported engagement: “Check-ins kept me focused and on track.” Recorded sessions supported attendance: “Recorded sessions helped when I couldn’t join live at noon.” Program improved family eating patterns through enhanced knowledge and awareness: “I have an increased understanding of health-related benefits of food, which carries over to my family’s eating habits.”
				“I added some diversity of ingredients to the family’s diet which is good as most of the kids are picky.”
Intervention engagement	Attendance: ≥3 culinary sessions; ≥2 coaching check-ins	≥75%	80% (culinary); 90% (coaching)	Program materials facilitated participation: “I loved learning more about ingredients that I have not used before and the benefits they can have for our bodies!”
				“The recipes, the ‘starter kit’ helped to make the process easy vs. having to add additional elements we may not have quickly at hand, especially with a picky child.”
				Coaching supported engagement: “Check-ins kept me focused and on track.”
				“I think they (coaching sessions) were great. Could be as long/short as needed and provided another outlet for questions you may have after the sessions and more specific to your child.
Progression criteria	--	Meet ≥4 of 5 criteria	Met 3 of 5	Suggestions for future refinement focused on session length/flexibility: “Shorter or segmented modules would help busy families;” “Modules that can be watched independently.”

Table 1 presents a mixed methods joint-display integrating quantitative feasibility and acceptability outcomes with embedded qualitative caregiver feedback to illustrate how participant experiences contextualized progression criteria. Qualitative feedback largely aligned with quantitative feasibility patterns. High engagement and acceptability scores were supported by caregivers’ descriptions of the usefulness of coaching, the clarity of nutrition and culinary education, and ease of implementing the strategies at home. Qualitative insights also helped explain feasibility challenges, particularly scheduling demands and the burden of lengthy recordings, which contributed to lower assessment completion at later time points.

Exploratory Outcomes

Exploratory outcomes were summarized descriptively to inform refinement of measures and to estimate variance for future sample size planning; however, the small sample size and missing data at follow-up preclude meaningful interpretation. Descriptive values for exploratory outcomes at each timepoint are provided in Supplementary Table 2 to support interpretation and future trial planning, but no inferences about intervention effects should be drawn. ASA-24 dietary recalls were collected as an exploratory outcome to assess the feasibility of administering 24-hour recalls within a remotely delivered culinary medicine intervention. At baseline, seven caregivers (78%) completed a recall and six (67%) completed a recall for their child; at post-intervention, two caregivers (22%) completed a recall and two (22%) completed a recall for their child; and at 3-month follow-up no participants completed dietary recalls. Given the exploratory intent of this measure, the small sample size, and limited completion at subsequent timepoints, no detailed dietary analyses were conducted. These data were used primarily to assess caregivers’ willingness and ability to complete ASA-24 recalls as part of a future fully powered trial.

Qualitative Findings

Caregiver motivations and goals. Caregiver motivations for participation included identifying appealing foods during taste changes or poor appetite, managing differing household preferences, and preparing meals their child could eat with the family. Several described fatigue around cooking, limited nutrition knowledge, and uncertainty about food safety. One caregiver summarized this need for practical guidance by explaining, “If I could know how a pancake, egg, and bacon breakfast could be made healthier vs. just being given a quinoa salad recipe, it would ground my understanding of how food works a bit more.”

Caregivers identified a range of goals they hoped to achieve through participation, including a desire to learn new recipes and become familiar with less familiar foods, expand their cooking skills, and rekindle interest in meal preparation. Caregivers sought guidance on creating meals that met the nutritional needs of both their child with cancer and other family members. Several emphasized wanting to better understand the role of food in health, including gut health, symptom management, and the nutritional contributions of specific foods. Additional goals included increasing family meal participation and enjoyment and addressing personal health objectives such as weight loss.

Program evaluation. Caregivers valued the clear nutrition education, practical symptom-focused guidance, and confidence gained using new ingredients. The practical orientation of the classes, including easy-to-follow instruction, accessible recipes, and curated “starter kits,” was viewed as highly supportive of their ability to cook at home. Coaching calls reinforced lessons and allowed troubleshooting. Several caregivers noted that these brief meetings helped them stay focused and offered a sense of support throughout the program.

When discussing aspects they liked least, caregivers expressed challenges related to time, including difficulty attending live sessions due to work demands or managing their child’s health, and finding long recordings hard to fit into their schedules. Some caregivers reported positive changes in family eating patterns, including increased awareness of healthier options when shopping, incorporating program recipes into weekly routines, diversifying ingredients, and gaining a deeper understanding of the health benefits of food. Recommendations for strengthening future sessions included offering modular or shorter recordings with optional Q&A review sessions. Suggestions for coaching improvements were fewer, reflecting broad satisfaction; one caregiver recommended having a more structured set of discussion prompts.