Follow me on

Diagnosing Obesity



sharma-obesity-scale3I am currently attending the 74th Scientific Session of the American Diabetes Association in San Francisco, where obesity is certainly a topic that permeates its way through much of the program.

However, despite all this talk, obesity continues to not be “formally” recognized as a “diagnosis” when it comes to patient care.

Thus, a paper by Canadian Obesity Network boot camper Bliie-Jean Martin and colleagues from the University of Calgary, published in BMC Health Services Research, the coding for obesity in administrative data bases and hospital discharge data is rather sketchy.

For their study, Martin and colleagues used a large coronary catheterization registry and a hospital discharge abstract database, which together consisted of more than 17,000 patients.

Based on how often the ICD-10 codes for obesity (E65-68) appeared in these datasets, it is evident that obesity was poorly coded for in the discharge database – in fact, only 2.4% of the discharge abstracts had this diagnosis (in contrast to about 20% of patients in the cardiac registry – which is likely to be more accurate).

Assuming the actual prevalence of obesity to be at least as high in patients discharged from hospital, as it is in the cardiac registry, the sensitivity of identifying obese patients based on the coding of the diagnosis is only about 8% – this means the vast majority of cases of obesity would be missed.

On the other hand, in the few cases where obesity codes were included in the discharge data set, this label was indeed correct (99% specificity).

As the authors conclude, given this state of affairs, hospital discharge databases are highly unreliable when it comes to determining obesity prevalence or burden of disease.

While there may certainly be other conditions that are “under diagnosed” and do not find themselves well reflected in such databases, nowhere is the discrepancy between prevalence and coding likely to be as great as for obesity.

This rather cavalier attitude towards coding for obesity must change if we hope to better understand the importance of obesity related morbidity in the health care system.

@DrSharma
San Francisco, CA

ResearchBlogging.orgMartin BJ, Chen G, Graham M, & Quan H (2014). Coding of obesity in administrative hospital discharge abstract data: accuracy and impact for future research studies. BMC health services research, 14 PMID: 24524687

.

2 Comments

  1. Obesity is a physical state. As is leanness. Is “leanness” a diagnosis applied to a professional marathoner? Is “marked muscularity” something with which bodybuilders can be “diagnosed”? Is “slightly flabby” a diagnosis?

    Post a Reply
  2. I got on Dr. Freedhof’s case for the same picture you used for your blog today: The picture you chose for your article shows 125 lb. or 56 kg. and this isn’t anywhere close to obesity unless one is EXTREMELY short. Like about 137 cm. or 54 inches, which would be 4’6″. And I think it’s misleading to include this kind of picture when talking about obesity — it contributes to a very distorted view of body size which no one needs to have foisted upon them. But otherwise, I continue to enjoy your columns! 🙂

    Post a Reply

Submit a Comment

Your email address will not be published.