Missing BMI Data

Comparisons

Evidence quality

Other

Highlighted claims

• Valid BMI measurements were available for 4432 symptomatic colorectal cancer patients, representing 82% of the symptomatic cohort. — Variations in symptoms, endoscopy use and emergency diagnosis of colorectal cancer by body mass index: a retrospective cohort study using linked electronic health records in England
• Patients with missing BMI were younger, less deprived, and had fewer comorbidities. — Variations in symptoms, endoscopy use and emergency diagnosis of colorectal cancer by body mass index: a retrospective cohort study using linked electronic health records in England
• In fully adjusted regression, missing BMI was associated with increased odds of emergency presentation for colon cancer but not for rectal cancer. — Variations in symptoms, endoscopy use and emergency diagnosis of colorectal cancer by body mass index: a retrospective cohort study using linked electronic health records in England
• Sensitivity analyses showed little change when individual comorbidities replaced the Charlson Comorbidity Index or when BMI was measured within three rather than six years before diagnosis. — Variations in symptoms, endoscopy use and emergency diagnosis of colorectal cancer by body mass index: a retrospective cohort study using linked electronic health records in England
• Among colon cancer patients, missing BMI was associated with lower endoscopy percentage and higher emergency presentation percentage. — Variations in symptoms, endoscopy use and emergency diagnosis of colorectal cancer by body mass index: a retrospective cohort study using linked electronic health records in England
• Standard multiple imputation was not considered appropriate because BMI data were likely missing not at random. — Variations in symptoms, endoscopy use and emergency diagnosis of colorectal cancer by body mass index: a retrospective cohort study using linked electronic health records in England