Senior Consultant Epidemiologist IQVIA Inc New York, United States
Background: Real-world data for cancer observational studies is essential for understanding the clinical trajectory of the patients and the safety and effectiveness of interventions. Weight loss and anorexia are cardinal manifestations of cancer cachexia, that is a multifactorial syndrome occurring in more than 80% of patients with cancer. Assessment and handling of cancer cachexia are major issues and challenges for the medical community.
Objectives: To assess the feasibility of conducting an observational study by evaluating quality and completeness of key study variables of interest, including BMI, lab tests, adverse events (AEs), treatment regimen, and dose changes and interruptions, to support the unmet medical need of treating cachexia in advanced cancer patients.
Methods: The feasibility study was conducted in IQVIA Oncology EMR (OncEMR) and US Health System EMR (US HS EMR). The study population included patients ≥ 18 years of age diagnosed with non-small cell lung-, pancreatic-, gastrointestinal- and colorectal cancer, and covered the period of 01Jan16-31Jul21. Results for quality and completeness of study variables were provided as descriptive statistics.
Results: In OncEMR and US HS EMR, among cancer patients, ~30% had a recorded/derived stage at initial diagnosis and ≥ 2 BMI measurements (N= 58,100 and 22,705, respectively). US HS EMR had a higher mean for BMI recorded per patient (22.8) as compared to OncEMR (18.3). Both data sources were comparable in the types of AEs and first line treatment therapies captured. The US HS EMR had much more complete and clean lab data (~80% of tests had >30 measurements per patient whereas ~70% of tests had >10 measurements in OncEMR) and a higher relative percentage of AEs captured. A small proportion of patients had complete information on treatment dosing (~40% and ~10% of patients in OncEMR had missing units and dose, respectively). Dose interruptions had clear median values across the two databases.
Conclusions: There was a significant number of patients identified of which a reasonable proportion had BMI measurements, lab test results, AEs, and treatment regimens. The quality and completeness of data was comparable in both databases except for cleaner lab test results in the US HS EMR. It was encouraging to see overlap in many clinical values across the two databases, which were consistent with the NCCN guidelines and FDA labels. Given the presence of unstructured data in the US HS EMR, it would be especially helpful in registering BMI as well as treatment-related dosing information. There is still an unmet need of having more carefully assessed patients and reported data to infer conditions such as cachexia.
