Increasing Prompt Response Rates in the CHICA System

We have developed a novel decision support system for implementing clinical guidelines in pediatric practice. CHICA (Child Health Improvement through Computer Automation) combines three elements: (1) pediatric guidelines encoded in Arden Syntax; (2) a dynamic, scannable paper user interface; and (3) an HL7-compliant interface to existing electronic medical record systems. The result is a system that both delivers "just-in-time" patient-relevant guidelines to physicians during the clinical encounter, and accurately captures structured data from all who interact with it. Preliminary work with CHICA has demonstrated the feasibility of using the system to implement and evaluate clinical guidelines.

However, analyses have shown that physicians ignore a fair number of prompts. We would like to experiment with changes to the system to see if we can increase physicians' response rates. This could include highlighting prompts, reordering them, or flagging them in other ways. The main outcome of interest is the rate at which physicians answer prompts.

This will be a randomized, controlled trial of the CHICA system to see if we can improve prompt response rates. We will randomize physicians or clinics to receive some form of prompt change, including highlighting them, reordering them, or flagging them. No other changes will be made to care.

We will extract data from the CHICA system for all patients seen in our study clinics. This data will include a the clinic location, whether or not a physician responded to a prompt, the prompt's position on the form, the patient's gender, the rule priority of the prompt, rule title, patient insurance category, the patient's age in days, and the name of the provider. The main outcome of interest is whether or not a prompt was answered (discussed/not discussed).