Date of Award

December 2017

Degree Type


Degree Name

Doctor of Philosophy


Health Sciences

First Advisor

Timothy Patrick

Second Advisor

Kin Wah Fung

Committee Members

Paul Fontelo, Virginia Stoffel, Mike Conway, Hadi Kharrazi, Jake Luo, Anthony Faiola, Rohit Kate


Non-adherence to antidepressants is a major obstacle to antidepressants therapeutic benefits, resulting in increased risk of relapse, emergency visits, and significant burden on individuals and the healthcare system. Several studies showed that non-adherence is weakly associated with personal and clinical variables, but strongly associated with patients’ beliefs and attitudes towards medications. The traditional methods for identifying the key dimensions of patients’ attitudes towards antidepressants are associated with some methodological limitations, such as concern about confidentiality of personal information. In this study, attempts have been made to address the limitations by utilizing patients’ self report experiences in online healthcare forums to identify underlying factors affecting patients attitudes towards antidepressants. The data source of the study was a healthcare forum called “”. 892 patients’ reviews were randomly collected from the forum for the four most commonly prescribed antidepressants including Sertraline (Zoloft) and Escitalopram (Lexapro) from SSRI class, and Venlafaxine (Effexor) and duloxetine (Cymbalta) from SNRI class. Methodology of this study is composed of two main phases: I) generating structured data from unstructured patients’ drug reviews and testing hypotheses concerning attitude, II) identification and normalization of Adverse Drug Reactions (ADRs), Withdrawal Symptoms (WDs) and Drug Indications (DIs) from the posts, and mapping them to both The UMLS and SNOMED CT concepts. Phase II also includes testing the association between ADRs and attitude. The result of the first phase of this study showed that “experience of adverse drug reactions”, “perceived distress received from ADRs”, “lack of knowledge about medication’s mechanism”, “withdrawal experience”, “duration of usage”, and “drug effectiveness” are strongly associated with patients attitudes. However, demographic variables including “age” and “gender” are not associated with attitude. Analysis of the data in second phase of the study showed that from 6,534 identified entities, 73% are ADRs, 12% are WDs, and 15 % are drug indications. In addition, psychological and cognitive expressions have higher variability than physiological expressions. All three types of entities were mapped to 811 UMLS and SNOMED CT concepts. Testing the association between ADRs and attitude showed that from twenty-one physiological ADRs specified in the ASEC questionnaire, “dry mouth”, “increased appetite”, “disorientation”, “yawning”, “weight gain”, and “problem with sexual dysfunction” are associated with attitude. A set of psychological and cognitive ADRs, such as “emotional indifference” and “memory problem" were also tested that showed significance association between these types of ADRs and attitude. The findings of this study have important implications for designing clinical interventions aiming to improve patients' adherence towards antidepressants. In addition, the dataset generated in this study has significant implications for improving performance of text-mining algorithms aiming to identify health related information from consumer health posts. Moreover, the dataset can be used for generating and testing hypotheses related to ADRs associated with psychiatric mediations, and identifying factors associated with discontinuation of antidepressants. The dataset and guidelines of this study are available at