Publication Date



Open access

Embargo Period


Degree Type


Degree Name

Doctor of Philosophy (PHD)


Epidemiology (Medicine)

Date of Defense


First Committee Member

Guillermo Prado

Second Committee Member

Hermes Florez

Third Committee Member

Kathryn E. McCollister

Fourth Committee Member

Dev Pathak


The main objectives of this dissertation were to examine the effects of bariatric surgery on medication and health services utilization among a cohort of Commercial and Medicare insured members from a large health benefits organization in the U.S.. A total of 1,492 members with morbid obesity underwent gastric bypass (n=785) or gastric banding (n=707) procedure between January 2005 and June 2008. Administrative claims databases were accessed and three data files including a member file, a medical file and a pharmacy file were merged at the member level. Non-parametric Wilcoxon signed rank tests revealed that the average number of all prescription claims were significantly lower during the 12 months post-surgery, compared to the 12 months pre surgery (p<0.0001). Moreover, McNemar’s Chi Square analyses showed that after the surgery, there was a statistically significant (p=<0001) decline in the proportion of members utilizing antihypertensives, antidiabetics and antihyperlipidemics. Our results also showed that the average number of prescription claims for each of these medication groups significantly declined during the 12-month post-surgery period, among members who had at least one prescription for these medications before the surgery (p<0.0001). Logistic regression modeling revealed that members who underwent bypass procedures were more likely to discontinue antihypertensives (OR=2.04; 95% CI= 1.30-3.23) , antihyperlipidemics (OR=3.25; 95% CI 1.96-5.40) and antidiabetics (OR=1.89; 95% CI 1.13–3.08) post-surgery than members who underwent banding procedures. In terms of medical services utilization, our results showed a significant decline in the average number of medical claims for all outpatient services overall from the 12 months pre to the 12 months post-surgery (p<0.0001). In contrast, the average number of medical claims for emergency room and inpatient hospitalizations were significantly increased from the pre to the post-surgery period (p<0.01). Logistic regression modeling revealed that the type of bariatric surgery was a significant positive predictor for inpatient hospitalizations post-surgery (OR =2.33; 95% CI= 1.76-3.08; p<0.0001) but not for emergency room visits (OR=1.23; 95% CI 0.97–1.56). The implications of the findings from a managed care perspective are discussed, along with limitation and future directions.


bariatric surgery; morbid obesity; medication utilization; health services utilization; claims data; co-morbidities