Date of Award
January 2014
Document Type
Open Access Thesis
Degree Name
Master of Public Health (MPH)
Department
School of Public Health
First Advisor
Hongyu Zhao
Abstract
The quality of clinical studies rests on the reliability of the disease diagnosis, and it is important to assess various factors associated with the ability of a physician to provide an accurate diagnosis. Endometriosis is a gynecological disorder in women, which has typically been difficult to diagnose and assess accurately. We focus on the analysis of data collected in the Physician Reliability Study of the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD), National Institutes of Health (NIH), on the agreement between physicians in obstetrics and gynecology in the diagnosis of endometriosis. The objective of our analysis is to investigate the performance of the physicians in diagnosing endometriosis and examine whether there are statistically significant differences in average diagnostic performance among the three groups of gynecologists in the study: international academic experts, regional expert surgeons, and residents. Given the diagnostic rating of each physician expert for every patient, we propose an expectation-maximization (EM) algorithm to infer the true patient disease status, and measure the performance of each physician in diagnosing endometriosis. This is achieved by estimating the true disease status, and then calculating the sensitivity and specificity of each physician rater in diagnosing the disease. The results show that, although there is a marked difference in performance among the physicians, there is no significant difference among the three different groups of experts. This approach can be broadly used to estimate the sensitivity and specificity of a diagnostic test, when the true disease status is not known.
Recommended Citation
Jin, Susan, "Measuring Performance Of Physicians In The Diagnosis Of Endometriosis Using An Expectation-Maximization Algorithm" (2014). Public Health Theses. 1141.
https://elischolar.library.yale.edu/ysphtdl/1141
This Article is Open Access
Comments
This is an Open Access Thesis.