VA is mandated to routinely measure, track, and report its performance on a multitude of quality measures. Among these are hospital readmission rates, which have increasingly gained national prominence with endorsement by the National Quality Forum and public posting of 30-day readmission rates for congestive heart failure (CHF), acute myocardial infarction (AMI) and pneumonia (PNA) on the VA and CMS Hospital Compare websites. Hospital readmissions are common, with 20% of hospitalized patients readmitted within 30 days of discharge and 56% within one year. Moreover, recent research has demonstrated considerable variation in readmission rates among hospitals. Recent VA national averages for 30-day readmissions, as reported by the VA Hospital Compare website, are 22%, 21% and 16% for CHF, AMI, and PNA, respectively.
Because prior studies of organizational factors associated with measurement-based quality improvement (MBQI) have examined implementation of QI programs broadly, without consideration of the content of what hospitals were trying to improve, this study sought to examine MBQI in the context of improving condition-specific readmission rates among VA hospitals. In our study, based on a conceptual model derived from organizational theory (QI Fit Model), we hypothesized that improvement in measure performance would be strongly associated with the "fit" between the measure and measure-specific organizational characteristics (e.g., leadership, culture, structure and resources).
The objectives of this study were to: 1) Assess the extent of improvement of VA hospitals on hospital readmission rates for three conditions-CHF, AMI and PNA-over a 3-year timeframe; 2) Assess the contribution of fit between each readmission rate measure and measure-specific organizational characteristics of a hospital's medical service on improvement in measure performance; and 3) Determine specific areas of good and poor fit within VA hospitals that are significantly associated with readmission rate performance, in order to identify actionable facilitators and barriers to improvement.
To achieve a balance high- and low-performing facilities in our study sample, VA hospitals with active cardiac catheterization labs were recruited based on two considerations: 1) their average 2011 readmission rates across the three conditions (data accessed through VA IPEC disease-specific readmission cubes), and 2) the average change in readmissions from 2006 to 2011 across the three conditions. We conducted telephone interviews with hospital service leaders (e.g., Chiefs of Medicine [COMs]) and collected quantitative and qualitative data regarding perceptions of their hospital's CHF, AMI and PNA readmission rates. We identified hospital organizational factors influencing improvement on readmission measures and conducted analyses to assess the influence of these factors on improving measure performance.
We recruited 40 VA hospitals, across 17 VISNs, and interviewed hospital service leaders at each facility (38 COMs; 2 Chiefs of Staff). Twenty-one, 16, and 9 hospitals were classified as high performers (top 50% across all VA facilities) for AMI, CHF and PNA readmission rates, respectively; 19, 24 and 31 hospitals were classified as low performers with respect to the three conditions. On average, the 40 hospitals in our sample demonstrated a 7% worsening (percent change) of readmission rates across the three conditions during our study timeframe; individually, CHF readmissions improved by 3%, while PNA and AMI worsened by 6% and 18%, respectively.
Contrary to our original hypothesis, analyses of interview data revealed that lower-performing hospitals consistently exhibited organizational characteristics indicating active initiatives aimed at reducing readmission rates for CHF and PNA. Conversely, relative to the low performers, higher-performing hospitals demonstrated less of a focus on improving their readmission rates for the two conditions; results for AMI readmissions were mixed across high and low-performing hospitals. On average, in comparison to high-performing hospitals, leadership indicated that reducing readmissions for CHF and PNA was of higher importance (1-5 Likert scale: 1=high, 5=low) relative to other service line priorities (PNA effect size: 1.2, p-value: 0.01). Furthermore, active Section Chief and clinical champion involvement was more apparent in lower-performing hospitals. Feasibility of improving readmissions for CHF and PNA was consistently rated higher (more feasible) in lower-performing hospitals (PNA effect size: 1.2, p-value: 0.05), potentially indicating a "ceiling effect" in high performers in their ability to improve on current rates. Not surprisingly, lower- performing hospitals were consistently less satisfied with their current readmission rates. Communication, availability and active tracking of rates, as well as efforts to reduce readmissions for CHF and PNA, were also more apparent in lower-performing hospitals. Finally, lower-performing hospitals were encouraged more often to improve their rates, and had more resources (e.g., training, education) and requirements (e.g., financial incentives, numerical goals) to reduce readmissions (CHF Education effect size: .93, p-value: 0.02).
While our findings are contrary to our original hypothesis, they suggest that lower-performing hospitals felt a greater urgency to improve readmission rates for CHF and PNA, while high performers were satisfied with their rates and had thus made lowering readmissions for specific conditions a lower priority. Additionally, our study findings represent a cross-section of results amid a continuum of MBQI efforts. To more accurately identify organizational facilitators and barriers to improve readmission rates, additional research is needed to assess whether or not, over time, lower-performing hospitals are able to meet their readmission rate goals for specific conditions. Similarly, identification of determinants of sustained improvement over time, in both low- and high-performing hospitals, is needed.
MBQI has been adopted within and outside the VA as one means of improving quality. However, the effectiveness of MBQI varies and our knowledge of key determinants of its effectiveness continues to be limited. Further examination of organizational factors associated with sustained readmission rate reduction, as well as variation in improved performance on quality measures, would help to better inform health systems, managers and clinicians seeking to improve performance more consistently. This study contributes to improving veterans' healthcare by beginning to identify potential organizational facilitators and barriers to successful MBQI to help address quality problems more effectively. Additionally, the study illustrates that MBQI appears to have salient impact in motivating lower-performing hospitals to become more engaged and supportive of efforts to improve quality. In turn, this suggests that VA-wide efforts to encourage such QI can be effective.
- Shwartz M, Peköz EA, Labonte A, Heineke J, Restuccia JD. Bringing responsibility for small area variations in hospitalization rates back to the hospital: the propensity to hospitalize index and a test of the Roemer's Law. Medical care. 2011 Dec 1; 49(12):1062-7.
- Restuccia JD, Shwartz M, Kreger BE. Underuse of percutaneous coronary intervention. JAMA. 2011 Nov 16; 306(19):2094; author reply 2094-5.
- Wilkinson JE, Lauer E, Freund KM, Rosen AK. Determinants of mammography in women with intellectual disabilities. Journal of the American Board of Family Medicine : JABFM. 2011 Nov 1; 24(6):693-703.
- Restuccia JD. The Association of Hospital Characteristics and Quality Improvement Activities in Veterans Administration Inpatient Medical Services. Paper presented at: Centre for Research on Health and Social Care Management Meeting; 2012 Mar 19; Bocconi University, Milan, Italy.