3122 — Successful Generalization of a Program to Improve HIV Testing

Goetz MB,  Hoang T,  Knapp H, and Burgess J, VA Greater Los Angeles HCS; Goffin A, VA New York Harbor Healthcare System; Kertz B, Houston VA Medical Center; Fletcher M, VA Greater Los Angeles HCS; Gifford AL, VA New England Healthcare System; Asch S, VA Palo Alto Healthcare System;

Objectives:
To assess the regional role out of an effective multimodal organizational intervention to improve HIV testing rates in unexposed facilities and the need for sustained involvement by project staff.

Methods:
Fifteen facilities in three VISNs were randomly assigned to the NATIONAL vs LOCAL strategy, while 7 facilities in two other VISNs served as controls matched by baseline testing rates, size and time. At NATIONAL and LOCAL facilities, the project staff met with facility at project launch to describe the interventional components including a clinical reminder, provider activation, quarterly clinic-level feedback, and changed testing policies. For NATIONAL facilities, the project team conducted additional monthly conference calls with local leadership and made an in-person follow-up to further promote HIV testing. This initial phase of the project promoted HIV testing to persons with known risk factors, while a subsequent expanded phase promoted routine offer of testing in accordance with changed VA policy. We compared testing rates among previously untested patients for the 6 months before and after implementation, adjusting for clustering, patient, provider and facility characteristics.

Results:
Adjusted incident rate of risk-based HIV testing increased from 7 to 15% for NATIONAL versus from 7% to 11% for LOCAL facilities in ten randomized facilities. Subsequent routine testing in fifteen randomized facilities increased from 3 to 16% for NATIONAL from 2% to 9% for LOCAL facilities. No change in risk based testing occurred in the control sites, and only modest changes occurred in routine testing rates (1.6% to 2.3%). At the facility-level (p <0.05) and the patient-level (p <0.001), increases in testing rates were greater at NATIONAL than at LOCAL sites. Prevalent risk-based testing rates reached 24% at NATIONAL and 19% at LOCAL sites.

Implications:
Although both implementation methods increased HIV testing, these results show that ongoing support from a national project team was more effective than lower intensity local implementation in bringing this successful intervention to scale.

Impacts:
Our results confirm the effectiveness of this multi-modal program in promoting HIV testing. Economies of scale in large integrated systems like the VA may ease development of centralized teams, and future research must evaluate their cost effectiveness.