|
Abstract:
|
Increasingly, educational, administrative and policy interventions are being carried out to improve public health, quality of medication use and/or contain costs. Several statistical methods have been suggested to implement the intervention procedures and to assess its impact. These methods may vary from randomized controlled trials to Auto-Regressive Integrated Moving Average (ARIMA) to segmented regression analysis or parameter-driven models in regression analysis. Clearly, the methodology is partially guided by nature and availability of data. Thus, ARIMA is more suitable for data with serial dependence. But, it also requires a large number of data points for better results. Segmented regression assumes linearity which often may hold only over short intervals. This method may also not be appropriate for highly auto-correlated data. In this paper, these methods will be compared.
|
