Date of Award


Degree Type


Degree Name

Master of Public Health (MPH)


Public Health

First Advisor

Christine E. Stauber, Ph.D

Second Advisor

Lisa Casanova, Ph.D


Background: Access to improved water and sanitation are key measures of the World Health Organizations. However, while a community can be classified as having access to improved water and sanitation, the possibility of microbiological contaminations exists. Globally, there is a need to assess the quality of drinking water to better classify levels of microbiological quality in attempts to reduce diarrheal disease burden. Utilizing the test for hydrogen sulfide (H2S) producing bacteria test is a cost effective and easy to use method that may be comparable to the traditional yet more costly method (IDEXX Colilert Quantitray). Due to a paucity of data on the test for H2S producing bacteria, this study was performed to examine how well the test for hydrogen sulfide (H2S) producing bacteria compared to traditional measure of fecal indicator bacteria total coliforms and E. coli in drinking water. Furthermore, an analysis of the ability of the test for H2S producing bacteria to predict diarrheal disease was also examined.

Methods: The following conditions for the H2S were examined in the study: 2 volumes (10mL or 90mL), 2 incubation times (24 and 48 hours) and the use of a semi-quantitative scoring system that measured the intensity of the black precipitate formed (H2S). To examine how well these conditions compared to E. coli and total coliform results, the following analyses were performed: 1) analysis of sensitivity and specificity to examine presence/absence of bacteria in both samples, 2) linear regression to examine how well a semi-quantitative H2S scoring system predicted bacterial concentrations and 3) logistic regression to examine how well the H2S test predicted risk of diarrheal disease.

Results: Within the dataset, there were 816 observations among the 7 communities involved in the study. The H2S test condition that had the highest sensitivity and specificity (94.23% and 36.07% respectively) for total coliforms was 90mL volume at 48 hours. This test condition also produced the highest sensitivity and specificity for E. coli (97.82% and 78.67%, respectively). An analysis using linear regression demonstrated that a semi-quantitative H2S scoring system was able to predict both total coliform and E. coli concentrations in the same samples. In a logistic regression analysis of diarrheal disease, the test of H2S producing bacteria suggested an increase in diarrheal disease risk for higher levels of H2S (OR of 1.18 (p=0.03; 1.02 – 1.35)).

Discussion: The initial results here suggest that the use of the test for H2S producing bacteria has potential with high sensitivity (>90%) for E. coli and total coliforms. The application of the semi-quantitative scoring system may also have applications in predicting concentration of E. coli and total coliforms and well as possibly predicting diarrheal disease. However, more work needs to be completed to standardize the semi-quantitative approach to reduce subjectivity of scoring as well as examine the role of the test in additional epidemiologic studies.

INDEX WORDS: waterborne disease, E. coli, Dominican Republic, microbial testing