Pitney Bowes                     Evaluation Summary

• Uses a relational database with five years of individual-level historical health claim, enrollment, and disability experience.
• Three years of data are available to Pitney Bowes on-line for immediate access through Hewit's Integrated Health Information System^TM (IHIS).
• Uses a total episode of care approach as a global measure of provider performance, health care cost and access, absence and disability days, and indicators of quality of care and health status.

Methodology
A multifaceted evaluation of the Power of 2 program was undertaken that had the following features:

• Multiple years of data were available (1993 through 1997) allowing for time series analysis for many programs;
• When possible, a pre- and post- comparison group design was utilized to measure program impact;
• Multivariate statistical methods controlled for differences in the groups compared.  These differences included age, sex, illness, severity, geographic cost differences, and plan design factors; and
• A comprehensive array of cost and utilization measures were analyzed, including inpatient and outpatient medical costs, prescription drug costs, emergency room utilization, disability days, incidental absence days, and  income replacement payments.  Clinical outcomes as well as self-reported data were also analyzed for some programs.

Data were obtained from several sources: health care claims data from 1993 through 1997, Pitney Bowes enrollment and payroll data, and disability management program experience came directly from IHIS.  Participation data from HCU and on-site medical services encounters were integrated into this person-level relational database for certain analyses.  Disease management results were obtained from surveys to participants and from Eli Lilly, the program sponsor.