Talk to the Veterans Crisis Line now
U.S. flag
An official website of the United States government

VA Health Systems Research

Go to the VA ORD website
Go to the QUERI website
HSRD Conference Logo

2023 HSR&D/QUERI National Conference Abstract

Printable View

4099 — Comparing Trends in Safety Events in VA-Delivered Care versus VA-Purchased Care in the Community

Lead/Presenter: Amy Rosen,  COIN - Bedford/Boston
All Authors: Rosen AK (Bedford/Boston COIN), Beilstein-Wedel E (Bedford/Boston COIN) Chan J (Bedford/Boston COIN) Miech E (Indianapolis VA, VA EXTEND QUERI Borzecki A (Bedford/Boston COIN) Yackel E (VHA National Center for Patient Safety) Flynn J (VHA Office of Integrated Veteran Care) Shwartz M (Bedford/Boston COIN)

Implementation of the 2014 Veterans Choice Program expanded Veterans’ access to community care (CC) paid for by the VA. While some studies have examined quality differences between VA and CC, little is known about the safety of VA-purchased care. National implementation of the 2018 “VHA Office of Community Care (OCC) Patient Safety Guidebook” aimed to ensure that longstanding VA safety practices (safety event reporting, investigation, and improvement) become standardized across VA and CC. We partnered with OCC and the National Center for Patient Safety (NCPS) to examine the Guidebook’s impact on national and facility-level trends in VA and CC safety event reporting.

Retrospective study of de-identified safety events voluntarily reported by VA staff into the VA’s Joint Patient Safety Reporting System (JPSR). We obtained 8 quarters of JPSR data (FY2020, quarter 2 to FY2022, quarter 1 [01/01/2020-12/31/21]). We examined changes in counts of safety events by setting from Year 1 to Year 2; low-harm vs. serious safety events; and types of safety events. We ran linear regression models with time (i.e., quarter) as a single independent variable to determine the national average change in safety events per 10,000 unique patients per quarter, separately for VA and CC. We also ran similar linear regression models at each individual facility to identify statistically significant facility-level trends over the 8 quarters.

There were a total of 330,579 safety events reported nationwide in Years 1 and 2; 94% and 5.8% of total events occurred in VA and CC, respectively. Of these, the largest proportion (82.7%) in both settings each year were “Care Management Events;” these included adverse drug reactions; delays in care; and failure to follow-up on laboratory tests. There was an overall increase of 9% in events reported; CC events increased 25% vs. 6.8% in VA. Almost all events (99.6%) were considered low-harm events. While both types of events were more prevalent in VA than CC, low-harm and serious safety events in CC increased 25.7% and 144.7% vs. 6.7% and 22.3% in VA, respectively. Despite these relatively large increases from Year 1 to Year 2, there was no statistically significant increase in events per 10,000 unique patients over the 8 quarters in either setting. There was wide variation in trends in VA and CC event reporting. Across VA facilities, the average change in events per 10,000 patients per quarter ranged from -39.7 to 22.2; in CC facilities (i.e., those facilities that purchased VA care through their VA “parent” facility), the average change in events ranged from -51.3 to 15.6. Of 140 facilities, trends were statistically significant at 21 VA and 30 CC facilities over the 8 quarters.

Safety event reporting increased after Guidebook implementation, particularly in CC. Improving Guidebook implementation among those sites where event reporting was unchanged (or worsened) would help to decrease variation in facility-level reporting.

As VA strives to become a High-Reliability Organization, it is imperative that safety practices become consistent across settings of care within the organization and also between facilities within the same setting.