HSR&D Home » Research » SDP 09-158 – HSR&D Study
Intervention for Stroke Improvement using Redesign Engineering
Linda S. Williams, MD
Richard L. Roudebush VA Medical Center, Indianapolis, IN
Funding Period: February 2010 - February 2013
There are approximately 6,000 annual admissions for acute ischemic stroke (AIS) in VA hospitals. The VA OQP Stroke Study reported national VA acute stroke care quality indicators (QIs) to all VA hospitals in 2009, demonstrating opportunities for improvement in early/acute processes of stroke care.
The objectives of this project were: 1) to evaluate the effect of the OQP report on organizational change to improve stroke care quality in VA; 2) to conduct a formative developmental evaluation in 12 high-volume VAMCs to understand organizational barriers and facilitators in the delivery of high quality acute stroke care; and 3) to compare a System Redesign tailored intervention plus performance data feedback to data feedback alone to improve two inpatient stroke QIs: 1) deep vein thrombosis prophylaxis (DVT), and 2) dysphagia screening.
In Aims 1 and 2, we conducted 104 semi-structured interviews with stroke care providers and managers at 12 VAMCs. Interviews were transcribed and analyzed using the constant comparison technique and matrix-based analyses. A preliminary list of codes were independently developed by coders using an initial set of interviews and reduced to a standard list of codes by consensus review. All transcripts were reviewed by at least two coders using this standard list of codes; subsequently, all qualitative data were merged into a single analytic file and the NVivo9 matrix query function was used to explore connections between qualitative themes, quotations, and quantitative data.
In Aim 3, we randomized 12 sites to intervention vs. control, stratified by the presence at baseline of an existing stroke data collection program. One site was unable to complete timely regulatory documentation necessary for participation in the intervention. Intervention stroke teams were trained in-person on System Redesign improvement techniques and developed a locally-tailored plan for improving the two QIs, then were coached for 6 months via bimonthly telephone calls and one on-site visit. After the intervention phase, we provided QI reports for the two stroke QIs plus 9 other Joint Commission (JC)/VA inpatient stroke QIs (monthly to intervention sites and quarterly to control sites). Performance data were collected via chart review for the 12 months prior to study initiation (pre-implementation) and the 12 months post-implementation. We developed QI numerator, denominator, and scoring algorithms based on existing JC/VA stroke QIs. We used VistAWeb to complete central chart review for the QI results with 10% random reliability resampling. We conducted analyses of baseline patient and facility data to examine differences between intervention and control sites, and calculated QI passing rates for all indicators. We evaluated the intervention effect by constructing logistic models for the DVT and Dysphagia indicators, with adjustment for time (pre- vs. post-implementation), group, time-group interaction, patient data, and presence of a data collection program at baseline. We are also modeling overall quality by constructing similar models of a composite quality indicator (passes/opportunities summarized at the facility level) and defect-free care (a patient level categorical measure of whether an individual patient received all processes for which they were eligible).
In Aims 1 and 2, 104 interviews were completed and analyzed. Emergent themes for Aim 1 included: 1) feedback was not uniformly provided to the relevant staff, 2) staff responses to the data were focused on a single strategy to improve care (e.g. revised electronic order entry), 3) the impression that performance was being measured without their knowledge influenced responses to the data, and 4) in general, feedback of stroke performance data was useful for public reporting. In Aim 2, the following barriers to high quality stroke care were identified by multiple sites: 1) budget constraints for hiring, 2) difficulty in training staff in a specialized area, 3) time constraints in focusing on another new area of care improvement, and 4) lack of physician comfort with providing acute stroke care. Facilitators identified by multiple sites included: 1) tapping into existing continuous staff readiness/training programs, 2) rewarding staff participation/ skill attainment, and 3) presence of an active stroke clinical champion. Sites with some ongoing stroke data collection at baseline were more likely to have multiple unanimous identifications of the stroke clinical champion than sites without an ongoing data collection program, and were also more likely to report strong Neurology-ED interservice communication practices.
In Aim 3, 3078 possible stroke admissions were identified; 2164 were confirmed by chart review and 358 were re-sampled for inter-rater reliability. Of 138 variables, only three had agreement (kappa) < 0.7 and none of these were part of the QI definitions. Agreement on the individual QI results (ineligible, passed, or failed) was 0.84-0.96. Preliminary analyses show that patients at intervention and control sites were similar, with the exception of slightly older patients at control sites (68.3 vs. 66.0 years, p < 0.001). In unadjusted analyses, the odds ratio of improving DVT prophylaxis was 60% higher, but not significantly so (p-value=0.16), at intervention sites (84.8% passing pre-intervention to 91.3% post-intervention, OR 1.88, p = 0.004) than at control sites (89.5% pre to 90.9% post, OR 1.17 p = 0.54). The odds ratio of improving dysphagia screening was 26% lower, but not significantly so (p = 0.14), at intervention (50.8% pre to 58.3% post, OR 1.35, p = .01) compared to control sites (36.5% pre to 50.5% post, OR 1.77, p < 0.001).
This project identified important strategies to improve feedback of quality data to VA clinical staff, and also identified key barriers and facilitators to providing high quality stroke care in the VA. If intervention effects are seen in adjusted analyses for individual QIs or for overall quality, this project will provide a blueprint for a feasible intervention to improve complex processes of inpatient care for both stroke and non-stroke admissions across the VA.
External Links for this Project
Dimensions for VADimensions for VA is a web-based tool available to VA staff that enables detailed searches of published research and research projects.
If you have VA-Intranet access, click here for more information vaww.hsrd.research.va.gov/dimensions/
VA staff not currently on the VA network can access Dimensions by registering for an account using their VA email address. Search Dimensions for this project
DRA: Aging, Older Veterans' Health and Care, Health Systems, Other Conditions
DRE: Technology Development and Assessment
Keywords: Organizational issues, Quality assurance, improvement, Stroke
MeSH Terms: none