The Association between Hospital Consolidation and Patient Safety - PowerPoint PPT Presentation

The Association between Hospital Consolidation and Patient Safety in the State of New York Kathryn Segal AcademyHealth Annual Research Meeting June 26, 2018

Acknowledgements  Disclosure  No conflict of interest; internally-funded research grant  Team at NORC  Sai Loganathan  Adil Moiduddin  Tim Riddle 2

Outline  Background  Objective  Methodology  Findings  Implications 3

Background  Hospital consolidation efforts have been on the rise throughout the United States, particularly in the state of New York Between 2012 and 2015, New York state ranked 6 th for hospital M&A (Howard  and Feyman, 2016)  Potential causal pathways: Hospital consolidation → Patient safety  ↓ Market competition → Patient safety risk ↑ (Gaynor & Town, 2012)  ↑ Economy of scale → Patient safety risk ↓ (Noether & May, 2017; Mutter et al, 2011) ↑ Clinical standardization → Patient safety risk ↓ (Noether & May, 2017)   Change in business operations and safety culture → Patient safety risk 4

Objective  Research questions:  Did hospital consolidation in NY adversely impact patient safety outcomes within the consolidating hospitals? – Do impacts vary by type of consolidation?  Following consolidation, is there a change in patient safety outcomes among the lower performing entities? – Do impacts vary by type of consolidation? 5

Methods: Study Population and Data Source  Study Design: Retrospective cohort with a concurrent propensity score weighted comparison group  Data Source: 2010-2014 New York State Inpatient Databases (SID) from the Healthcare Cost and Utilization Project (HCUP); 2010-2014 AHA Annual Survey Dataset  Unit of Analysis: Inpatient Episode  Study Population: Treatment Comparison NY Hospitals that Propensity score weighted experienced consolidation comparison group of during 2010-2014 consolidating hospitals from the same HRRs Number of Consolidation Events 15 Number of Hospital Facilities 51 139 Number of Inpatient Episodes 3,949,747 7,654,283 6

Methods: Outcome Variables  Inpatient admissions with a potentially preventable complication (PPC)  The 3M™ Potentially Preventable Complications Grouping Software (version 31.0) identifies: – “Conditions not present on admission and determines whether the conditions were potentially preventable given patient characteristics, reason for admission, clinical procedures, and interrelationships between underlying medical conditions”  Inpatient mortality 7

Methods: Defining Consolidation  We categorized consolidation events into three types:  Consolidation of multiple hospital systems (N = 5 consolidation events)  A hospital system purchases an individual hospital (N = 3 consolidation events)  Two or more individual hospitals consolidate to form a single entity (N = 7 consolidation events)  Identifying consolidation using AHA Annual Survey Data – SYSID: system identification number – HOSPID: hospital identification number – DSHOSPID: data-source (aka unique hospital facility) identification number  Supplemented the above method with extensive internet research  Hospital “history” pages on the website, news articles, etc. 8

Methods: Study Design  Difference-in-Difference model comparing consolidating hospitals to all other propensity score weighted hospitals from within the same Hospital Referral Region (HRR)  Covariates in propensity score model : number of hospital beds; ownership; teaching status; Critical Access Hospital (CAH) status; urban/rural; number of inpatient surgical operations; adjusted average daily census; age; sex; and baseline outcomes  Logit model with HRR level and time fixed effects, std. error clustered at hospital level, and episode (age; race; sex; select Elixhauser Comorbidities) and hospital-level (bed size; ownership; teaching status; CAH status; rurality) covariates 9

Findings RQ: Did hospital consolidation in NY adversely impact patient safety outcomes within the consolidating hospitals? Consolidating Hospitals Comparison Hospitals Difference-in- [90% CI] Pre-Period Post-Period Difference Pre-Period Post-Period Difference Difference Inpatient Admissions with a Potentially Preventable Complication (%) All 7.59 [7.35, 7.83] 7.55 [7.25, 7.85] -0.04 [-0.24, 0.16] 7.80 [7.59, 8.01] 7.85 [7.61, 8.09] 0.05 [-0.17, 0.27] -0.09 [-0.42, 0.24] System-System 7.87 [7.54, 8.20] 7.54 [7.25, 7.83] -0.33 [-0.56, -0.10] 8.17 [7.91, 8.43] 8.18 [7.90, 8.46] 0.01 [-0.23, 0.25] -0.34 [-0.68, -0.00] System-Hospital 7.17 [6.74, 7.60] 7.54 [7.01, 8.07] 0.37 [0.11, 0.63] 7.50 [7.28, 7.72] 7.65 [7.39, 7.91] 0.14 [-0.13, 0.41] 0.23 [-0.18, 0.64] Hospital-Hospital 7.63 [7.31, 7.95] 7.65 [7.32, 7.98] 0.01 [-0.32, 0.34] 7.62 [7.37, 7.87] 7.55 [7.38, 7.72] -0.07 [-0.30, 0.16] 0.08 [-0.31, 0.47] Inpatient Mortality (%) All 2.14 [1.95, 2.33] 1.99 [1.84, 2.14] -0.15 [-0.28, -0.02] 1.94 [1.80, 2.08] 1.88 [1.70, 2.06] -0.06 [-0.18, 0.06] -0.09 [-0.33, 0.15] System-System 1.92 [1.78, 2.06] 1.87 [1.70, 2.04] -0.05 [-0.14, 0.04] 2.08 [1.93, 2.23] 1.97 [1.78, 2.16] -0.10 [-0.19, -0.01] 0.05 [-0.09, 0.19] System-Hospital 2.25 [2.09, 2.41] 2.03 [1.86, 2.20] -0.23 [-0.34, -0.12] 1.85 [1.69, 2.01] 1.77 [1.58, 1.96] -0.09 [-0.19, 0.01] -0.14 [-0.31, 0.03] Hospital-Hospital 2.34 [2.09, 2.59] 2.31 [1.99, 2.63] -0.03 [-0.17, 0.11] 1.90 [1.80, 2.00] 1.83 [1.68, 1.98] -0.07 [-0.17, 0.03] 0.04 [-0.14, 0.22] • No evidence of an adverse impact on patient safety outcomes after consolidation • Statistically insignificant 4% relative decrease in the percentage of inpatient admissions with a PPC among system-system consolidations • Statistically insignificant 4% relative decrease in inpatient mortality among all hospital consolidations • Statistically insignificant 6% relative decrease in inpatient mortality among system-hospital consolidations 10

Findings RQ: Following consolidation, is there a change in patient safety outcomes among the lower performing entities? Consolidating Hospitals Comparison Hospitals Difference-in- [90% CI] Pre-Period Post-Period Difference Pre-Period Post-Period Difference Difference Inpatient Admissions with a Potentially Preventable Complication (%) All 8.65 [8.12, 9.18] 8.12 [7.60, 8.64] -0.53 [-0.88, -0.18] 8.54 [7.81, 9.27] 8.07 [7.75, 8.39] -0.46 [-1.38, 0.46] -0.07 [-1.03, 0.89] System-System 8.80 [8.31, 9.29] 8.17 [7.68, 8.66] -0.63 [-1.01, -0.25] 8.58 [7.99, 9.17] 8.03 [7.59, 8.47] -0.55 [-1.47, 0.37] -0.08 [-1.04, 0.88] System-Hospital 8.25 [7.82, 8.68] 7.92 [7.82, 8.02] -0.33 [-0.66, -0.00] 8.62 [8.02, 9.22] 8.00 [7.57, 8.43] -0.62 [-1.55, 0.31] 0.29 [-0.70, 1.28] Hospital-Hospital 8.66 [7.61, 9.71] 7.96 [7.15, 8.77] -0.69 [-1.71, 0.33] 8.55 [7.91, 9.19] 8.06 [7.67, 8.45] -0.48 [-1.40, 0.44] -0.21 [-1.53, 1.11] Inpatient Mortality (%) All 2.28 [2.15, 2.41] 2.06 [1.91, 2.21] -0.23 [-0.32, -0.14] 1.85 [1.72, 1.98] 1.97 [1.74, 2.20] 0.12 [-0.02, 0.26] -0.35 [-0.53, -0.17] System-System 2.09 [2.01, 2.17] 1.82 [1.72, 1.92] -0.27 [-0.38, -0.16] 1.93 [1.84, 2.02] 1.88 [1.76, 2.00] -0.05 [-0.15, 0.05] -0.22 [-0.39, -0.05] System-Hospital 2.23 [2.08, 2.38] 2.03 [1.85, 2.21] -0.20 [-0.32, -0.08] 1.90 [1.81, 1.99] 1.91 [1.78, 2.04] 0.01 [-0.11, 0.13] -0.21 [-0.39, -0.03] Hospital-Hospital 2.65 [2.47, 2.83] 2.53 [2.21, 2.85] -0.12 [-0.29, 0.05] 1.93 [1.83, 2.03] 1.88 [1.73, 2.03] -0.05 [-0.14, 0.04] -0.07 [-0.26, 0.12] • Inpatient mortality decreased by 15% among the lower performing entities after consolidation • For system-system consolidations, inpatient mortality decreased by 11% among lower performing entities • For system-hospital consolidations, inpatient mortality decreased by 9% among lower performing entities 11

Implications  We did not find evidence that hospital consolidation in the state of New York during this time had an adverse impact on patient safety outcomes  Rather, patient safety outcomes of the originally lower performing hospitals may have improved after consolidation, especially among system consolidations and systems that purchased individual hospitals  Study limitations:  Generalizability  Short post period  Potential spillover effects of consolidation on the comparison hospitals since they operate in the same HRR  Change in coding and billing practices following consolidation could impact the outcomes measured  We did not explicitly verify whether clinical standardization occurred 12

Kathryn Segal, BA Research Assistant, Health Care Evaluation segal-kathryn@norc.org Thank You!