Reliability of Insurance Claims to Categorize Obesity in Patients with Obstructive Sleep Apnea
Author(s)
R Benjamin Dexter, MS1, Caleb Woodford, BMath1, Naomi Alpert, MS2, Will McConnell, PhD2, Kate Cole, MS2, Kimberly L. Sterling, MS, PharmD2.
1ResMed Science Centre, Halifax, NS, Canada, 2ResMed Science Center, San Diego, CA, USA.
1ResMed Science Centre, Halifax, NS, Canada, 2ResMed Science Center, San Diego, CA, USA.
OBJECTIVES: Obesity and obesity-related conditions are currently top of mind for researchers considering the recent advances in obesity medications. With real-world evidence, insurance claims are often used to identify people living with a particular condition, but this comes with limitations in reliability. Therefore, we sought to investigate the reliability of ICD-10 codes in insurance claims for classifying a person’s obesity status against their measured body mass index (BMI) as recorded in electronic medical records (EMRs) from a sample of people diagnosed with obstructive sleep apnea (OSA), an obesity-related condition.
METHODS: This retrospective study used a linked dataset comprising insurance claims and EMRs of adults newly diagnosed with OSA starting positive airway pressure therapy (index). The closest EMR BMI measurement and ICD-10 diagnoses codes (for weight status (overweight: E66.3, obesity: E66.9, severe obesity: E66.01) or BMI (Z68.1-Z68.4)) to the index were compared. ICD-10 E-codes and Z-codes were combined to classify claims-based BMI as normal weight (BMI <25kg/m2), overweight (BMI 25-30 kg/m2), obese (BMI 30-40 kg/m2), and severe obese (BMI >40 kg/m2). The same categories were applied to the measured BMI from the EMR.
RESULTS: Of 54,497 patients with EMR BMI measurements, the majority (74%) also had an ICD-10 claim. Within ICD-10 claims-based categories, the EMR BMI breakdown (% normal weight/overweight/obese/severe obese) was: normal weight (79%/17%/3%/1%), overweight (4%/70%/23%/3%), obese (1%/6%/78%/16%), severe obese (1%/1%/23%/75%), and uncategorized (13%/33%/45%/9%).
CONCLUSIONS: These data show that ICD-10 codes in insurance claims do a fair job at identifying patients with OSA that are at least overweight. All EMR BMI categories were present amongst those without an ICD-10 code and therefore a lack of an ICD-10 code should not lead to a categorization of normal weight or non-obese.
METHODS: This retrospective study used a linked dataset comprising insurance claims and EMRs of adults newly diagnosed with OSA starting positive airway pressure therapy (index). The closest EMR BMI measurement and ICD-10 diagnoses codes (for weight status (overweight: E66.3, obesity: E66.9, severe obesity: E66.01) or BMI (Z68.1-Z68.4)) to the index were compared. ICD-10 E-codes and Z-codes were combined to classify claims-based BMI as normal weight (BMI <25kg/m2), overweight (BMI 25-30 kg/m2), obese (BMI 30-40 kg/m2), and severe obese (BMI >40 kg/m2). The same categories were applied to the measured BMI from the EMR.
RESULTS: Of 54,497 patients with EMR BMI measurements, the majority (74%) also had an ICD-10 claim. Within ICD-10 claims-based categories, the EMR BMI breakdown (% normal weight/overweight/obese/severe obese) was: normal weight (79%/17%/3%/1%), overweight (4%/70%/23%/3%), obese (1%/6%/78%/16%), severe obese (1%/1%/23%/75%), and uncategorized (13%/33%/45%/9%).
CONCLUSIONS: These data show that ICD-10 codes in insurance claims do a fair job at identifying patients with OSA that are at least overweight. All EMR BMI categories were present amongst those without an ICD-10 code and therefore a lack of an ICD-10 code should not lead to a categorization of normal weight or non-obese.
Conference/Value in Health Info
2025-05, ISPOR 2025, Montréal, Quebec, CA
Value in Health, Volume 28, Issue S1
Code
MSR91
Topic
Methodological & Statistical Research
Disease
SDC: Diabetes/Endocrine/Metabolic Disorders (including obesity), SDC: Respiratory-Related Disorders (Allergy, Asthma, Smoking, Other Respiratory)