(C-192) Is There a Lack of Dietary Data Collection in Real-World Pediatric Rare Disease Studies? Literature Review and Future Considerations

Background: For many pediatric rare diseases, medical nutrition is a life-saving therapy. While real-world dietary data are essential for understanding disease management and treatment outcomes, its collection is often infrequent due to patient/caregiver burden and operational/geographic complexities.

Objectives: To characterize dietary data collection methods in identified observational pediatric rare disease studies, and to provide considerations to enhance data collection in future research.

Methods: A targeted literature review was performed in Embase and MEDLINE, with abstracts screened to describe dietary data collection methods in observational pediatric rare disease studies. Considerations to mitigate challenges in dietary collection and improve data quality are presented.

Results: 99 total studies were found that collected dietary data across 27 diseases/disease categories, with the highest proportion of studies in phenylketonuria (PKU; 33 [33.3%]), cystic fibrosis (CF; 26 [26.3%]), glycogen storage diseases (GSD; 5 [5.1%]) and spinal muscular atrophy (5 [5.1%]). Dietary data were collected via 3 to 7-day food diaries (31 [31.3%]), other food records/diaries (35 [35.4%]), food frequency questionnaires (15 [15.2%]) or other methods (18 [18.2]%). 81 (81.8%) were natural history studies. Only 8 (8.1%) studies collected dietary data when assessing drug treatment outcomes.
For future work, study-specific solutions should lessen burden and data quality challenges. For patients with a disease requiring nutritional monitoring, dietary data are expected to be in medical records. Where data are not in records, choosing recalls vs. diaries should consider objectives, care setting(s), ages, impact of recall bias, most appropriate person to enter the data, etc. Additionally, the support of specialized dietitians working with patients at medical centers, is key for successful collection of detailed data.
Nutritional data collection platforms with guided menus can ease data entry, reduce missingness and should consider differences in geography/culture. Burgeoning food-image recognition platforms (using artificial intelligence) may reduce burden and allow data entry from proxy reporters (e.g., caregiver). Lastly, caregivers of breast/formula-fed or enteral/parental-fed populations could have structured questionnaires (verified with prescription and using up-to-date technology) in place of diaries.

Conclusions: Among the observational pediatric rare disease studies, methods of dietary data collection were highly varied. Few studies collected dietary data when understanding drug treatment outcomes, identifying a key knowledge gap. Future research should consider tailored solutions to ease burden and improve quality of data entry.