Venue: The Fuqua School of Business, Duke University, 1 Towerview Drive, Durham, NC 27708-0120
Presentation
Implementation of Inpatient Information Technology Systems: Assessing the Impact on Patient Safety and Staff
As individual hospitals and hospital systems contemplate implementation of inpatient information technology (IT) to improve patient care and outcomes, the impact and benefits accrue to particular hospital constituencies differently. These differences in the expected and actual utility of IT affect the ability to implement these complex information systems effectively when incentives for those constituencies conflict with each other. For inpatient information technology, one key constituency is nurses, whose work is essential to achieving good patient outcomes and protecting patient safety.
The U.S. Department of Veterans Affairs (VA) has been a leader in implementation of Computerized Patient Record Systems (CPRS) and Bar Code Medication Administration (BCMA). We take advantage of detailed patient records in the VA system, staffing data, and qualitative interview methodologies to undertake a mixed method longitudinal analysis of the VA CPRS/BCMA implementation. We conducted a quarterly analysis of inpatient data on 120 VA Medical Centers over 44 quarters of patient safety outcomes using the AHRQ Patient Safety Indicators (PSIs) and Inpatient Quality Indicators (IQIs), taking advantage of variation in CPRS/BCMA implementation across the VA system. We also conducted site visits at 8 VA facilities, with 10 to 26 interviews at each facility. The qualitative data were analyzed using Atlas TI.
Our key preliminary findings from the quantitative analysis are that: (1) CPRS implementation was associated with significant increases in accidental puncture/laceration rates (b ranges from 0.0008 to 0.0012; mean rate in 2005 was 0.0043); (2) CPRS was associated with a decline in risk-adjusted pneumonia mortality (b ranges from -0.019 to -0.022; mean rate in 2005 was 0.065); (3) BCMA implementation was associated with significant declines in decubitus ulcer rates in both acute and ICU inpatients (b ranges from -0.0007 to -0.002; mean in 2005 was 0.023); (4) BCMA implementation was associated in declines in gastrointestinal hemorrhage mortality (b ranges from -0.007 to -0.013; mean in 2005 was 0.024).
The qualitative research indicates that CPRS implementation was associated with dramatic increases in computerized templates that absorbed significant amounts of registered nursing attention and increased responsibilities and speed of workflow that could be associated with the increase in puncture/laceration rates. The qualitative data indicated that bar coding took more time and reduced medical errors, but the overall effectiveness of BCMA depended on the ability of the implementation team to resolve systems problems between pharmacy, nursing, and the technology.
Additional analyses are underway to learn the effects of BCMA and CPRS on paid nursing hours, as well as to include levels of nurse staffing and nurse staff characteristics in the quantitative models.
This research illustrates the complexities of information technology implementation in hospitals and suggests why efforts in large healthcare systems, such as VA, makes faster progress and works through initial barriers to change to successful implementation.