Date of Award


Degree Type


Degree Name

Master of Science (MS)


Pharmaceutical Sciences


Objective: This study assessed the macro level effects of multiple and varied forms of clinical guidance for medication based treatment for heart failure. Drug mention rates for physician visits by patients with heart failure were evaluated with respect to the dates of publication of large randomized trial evidence and guidelines. Design: Retrospective, cross-sectional series study Methods: We used the National Ambulatory Medical Care Survey (NAMCS) for years 1993-2000, which captures a probability sample of visits to United States physicians to provide national estimates. We examined heart failure coded visit drug mentions alongside research published during the same period to examine trends in medication prescribing and the aggregate influence of the dissemination of research findings. Multi year estimation equations from the National Center for Health Statistics (NCHS) were used for calculation of sampling error. Measurements: Medication mention rates were calculated for four sequential two year periods. Relative standard errors (RSEs) were generated for measuring reliability and stability of our findings of changes in medication mention rates for beta blockers, angiotensin converting enzyme inhibitors, spironolactone, and angiotensin receptor blockers. Stratification and logistic regression models were used to provide insight into other possible predictors. Results: The number of visits by a patient with heart failure to physicians was not statistically significantly different across the eight years of interest. The estimated medication mention rate of beta blockers, spironolactone, and angiotensin receptor blockers increased dramatically, but the number and rate of mentions was too low for statistical reliability. There was an adequate number of drug mentions of angiotensin converting enzyme inhibitors for reliable aggregate estimates, but there were not adequate numbers of mentions to demonstrate statistically significant increases over the eight years. Logistic regression models showed strong associations between increased drug mentions and later two year periods. This association was demonstrated by progressively larger odds ratios (ORs) for subsequent periods when the first two year period is used as a referent baseline. Discussion: The increases in medication mention rates for all medications corresponded with the findings of the major trials and evidence which we assessed. The NAM CS sample size and the low percent of drug mentions in the given therapeutic categories resulted in a lack of statistical power for determining statistical significance of the changes in medication mention rate. Conclusion: We conclude from our collected information, and statistical analyses that the NAM CS demonstrated marked trends, but this study was inadequately powered to establish statistical significance.