Fluoroquinolone antibiotics have been implicated in cases of central nervous system (CNS), peripheral nervous system (PNS), and metabolic adverse events. In this dissertation, we investigate causal associations between fluoroquinolones and these adverse events. We used administrative health care claims data from a commercially insured adult population (Optum) prescribed either an oral fluoroquinolone or a comparator antibiotic between January 2000 and September 2015 for specific indications of interest. In Chapters 1&2, we conducted new-user, propensity score-matched cohort studies. In Chapter 1, the outcomes of interest were CNS and PNS dysfunction. Cox proportional hazards models were estimated after matching. The hazard ratio associated with fluoroquinolone exposure was 1.08 (95% confidence interval [CI]: 1.05-1.11) for CNS dysfunction, and 1.09 (95% CI: 1.07-1.11) for PNS dysfunction. In Chapter 2, our outcome was serious hypoglycemia. Conditional logistic regression was conducted after matching. The odds ratios associated with fluoroquinolone exposure were 1.30 (95% CI: 1.05-1.62) among individuals with diabetes and 1.06 (95% CI: 0.53-2.13) in those without diabetes. In Chapter 3, we used our cohort of fluoroquinolone users to develop and validate two types of risk prediction models, LASSO and random forest, in predicting CNS and PNS dysfunction, which were outcomes found to be associated with fluoroquinolones in the first chapter. We assessed the accuracy and calibration of these models in a validation subset using AUC and calibration curves. For CNS dysfunction, LASSO had an AUC of 0.81 (95% CI: 0.80-0.82), while random forest had an AUC of 0.80 (95% CI: 0.80-0.81). For PNS dysfunction, LASSO had an AUC of 0.75 (95% CI: 0.74-0.76) vs. an AUC of 0.73 (95% CI: 0.73-0.74) for random forest. Both LASSO models had better calibration than the corresponding random forest models. Fluoroquinolone antibiotic use was associated with the development of neurological and metabolic dysfunctions versus comparator antibiotic use in the adult population. LASSO outperformed random forest in predicting neurological dysfunction among fluoroquinolone users and should be considered for generating clinical risk prediction models when the cohort is modest in size, when the number of model predictors is modest, and when predictors are primarily binary.