Variation in hospital utilization of palliative interventions for patients with advanced gastrointestinal cancer near end of life.

BACKGROUND: Patients with advanced gastrointestinal (GI) cancer often undergo noncurative interventions with palliative intent to relieve high symptom burden near end of life. Hospital-level variation in intervention utilization remains unclear. METHODS: National cohort study of 142 304 patients with stage III or IV GI cancer within the National Cancer Database (2004-2014) who died within 1-year of diagnosis. Hospitals were stratified by palliative intervention utilization (surgery, chemotherapy, radiation, pain management). Multivariable, multinomial regression evaluated the association between patient/hospital factors and palliative intervention utilization. RESULTS: Across 1322 hospitals, median hospital palliative intervention utilization was 12.0% [interquartile range: 0.0%-26.1%]. Utilization increased over time in all but lowest utilizing hospitals. Relative to lowest utilizing hospitals, factors associated with a lower likelihood of care at highest utilizing hospitals included: race (White [ref]; Black-Relative Risk Ratio [RRR] 0.81, 95% confidence interval [0.77-0.85]) and lower income (RRR 0.81 [0.78-0.84]). Factors associated with a higher likelihood included: lower education level (RRR 1.62 [1.55-1.69]) and hospital type (community program [ref]; comprehensive community-RRR 1.33 [1.26-1.41]; academic-RRR 1.88 [1.77-1.99]; integrated network-RRR 1.79 [1.66-1.93]). CONCLUSION: Hospital variation in palliative intervention use is substantial and potentially associated with sociodemographic and hospital characteristics. Future work can examine how differences in hospital care processes translate to quantity/quality of life for cancer patients.

Hospital Variation in Mortality and Ventilator Management among Mechanically Ventilated Patients with ARDS.

RATIONALE: Acute Respiratory Distress Syndrome (ARDS) is associated with significant mortality. Despite the mortality benefits of lung protective ventilation, adherence rates to evidence-based ventilator practice have remained low and ARDS mortality has remained high. OBJECTIVE: Determine variation in ARDS mortality and adherence to low tidal volume ventilation (LTV) across US hospitals. MATERIALS AND METHODS: We identified mechanically ventilated patients with ARDS using data from Philips eICU (2014-2015). We then used multi-variable hierarchical logistic regression models with hospital site as the random effect and patient and hospital level factors as fixed effects to assess the hospital risk adjusted mortality rate and median odds ratio for the association between mortality and hospital site. We then assessed associations between adherence to LTV (defined as 4-8 mL/kg PBW) and hospital risk adjusted mortality rates using Spearman correlation. RESULTS: Among 4441 patients admitted at 110 hospitals with ARDS, the hospital risk-adjusted mortality rate ranged from 19% to 39%, and the MOR for hospital of admission was 1.33 (95% CI 1.25-1.41). Among 3070 patients at 72 hospitals with available ventilator data, 73% of patients had a median set Vt between 4 to 8 mL/kg PBW; hospital adherence rates to LTV ranged from 13% to 95%. There was no association between hospital adherence to LTV and risk-adjusted mortality rate (spearman correlation coefficient -0.01, p = .93). Similarly, among 956 patients who started with a Vt > 8 mL/kg PBW, there was no association between the percent of patients at each hospital whose Vt was decreased to ≤ 8 mL/kg PBW and risk adjusted mortality rate (spearman correlation coefficient .05, p = .73). CONCLUSION: Risk adjusted mortality and use of LTV for patients with ARDS varied widely across hospitals. However, hospital adherence to LTV was not associated with ARDS mortality rates. Further evaluation of hospital practices associated with lower ARDS mortality are warranted.

Variation of ventral and incisional hernia repairs in kidney transplant recipients.

INTRODUCTION: As survivorship following kidney transplant continues to improve, so does the probability of intervening on common surgical conditions, such as ventral or incisional hernia, in this population. Ventral hernia management is known to vary across institutions and this variation has an impact on patient outcomes. We sought to evaluate hospital level variation of ventral or incisional hernia repair (VIHR) in the kidney transplant population. METHODS: We performed a retrospective review of 100% inpatient Medicare claims to identify patients who underwent kidney transplant between 2007 and 2018. The primary outcome was 1- and 3-year ventral or incisional risk- and reliability-adjusted VIHR rates. Patient and hospital characteristics were evaluated across risk- and reliability-adjusted VIHR rate tertiles. Models were adjusted for age, sex, race, and Elixhauser comorbidities. RESULTS: Overall, 139,741 patients underwent kidney transplant during the study period with a mean age (SD) of 51.6 (13.7) years. 84,717 (60.6%) were male, and 72,657 (52.0%) were white. Median follow up time was 5.4 years. 2098 (1.50%) patients underwent VIHR. the 1 year risk- and reliability-adjusted hernia repair rates were 0.49% (95% Conf idence Interval (CI) 0.48-0.51, range 0.31-0.59) in tertile 1, 0.63% (95% CI 0.62-0.63, range 0.59-0.68) in tertile 2, and 0.98 (95% CI 0.91-1.05, range 0.68-2.94) in tertile 3. Accordingly, compared to hospitals in tertile 1, the odds of post-transplant hernia repair tertile 2 hospitals were 1.78 (95% CI 1.37-2.31) and at tertile 3 hospitals 3.53 (95% CI 2.87-4.33). CONCLUSIONS: In a large cohort of Medicare patients undergoing kidney transplant, the overall cumulative incidence of hernia repair varied substantially across hospital tertiles. Patient and hospital characteristics varied across tertile, most notably in diabetes and obesity. Future research is needed to understand if program and surgeon level factors are contributing to the observed variation in treatment of this common disease.

Practice variation and outcomes of minimally invasive minor liver resections in patients with colorectal liver metastases: a population-based study.

INTRODUCTION: In 2017, the Southampton guideline stated that minimally invasive liver resections (MILR) should considered standard practice for minor liver resections. This study aimed to assess recent implementation rates of minor MILR, factors associated with performing MILR, hospital variation, and outcomes in patients with colorectal liver metastases (CRLM). METHODS: This population-based study included all patients who underwent minor liver resection for CRLM in the Netherlands between 2014 and 2021. Factors associated with MILR and nationwide hospital variation were assessed using multilevel multivariable logistic regression. Propensity-score matching (PSM) was applied to compare outcomes between minor MILR and minor open liver resections. Overall survival (OS) was assessed with Kaplan-Meier analysis on patients operated until 2018. RESULTS: Of 4,488 patients included, 1,695 (37.8%) underwent MILR. PSM resulted in 1,338 patients in each group. Implementation of MILR increased to 51.2% in 2021. Factors associated with not performing MILR included treatment with preoperative chemotherapy (aOR 0.61 CI:0.50-0.75, p < 0.001), treatment in a tertiary referral hospital (aOR 0.57 CI:0.50-0.67, p < 0.001), and larger diameter and number of CRLM. Significant hospital variation was observed in use of MILR (7.5% to 93.0%). After case-mix correction, six hospitals performed fewer, and six hospitals performed more MILRs than expected. In the PSM cohort, MILR was associated with a decrease in blood loss (aOR 0.99 CI:0.99-0.99, p < 0.01), cardiac complications (aOR 0.29, CI:0.10-0.70, p = 0.009), IC admissions (aOR 0.66, CI:0.50-0.89, p = 0.005), and shorter hospital stay (aOR CI:0.94-0.99, p < 0.01). Five-year OS rates for MILR and OLR were 53.7% versus 48.6%, p = 0.21. CONCLUSION: Although uptake of MILR is increasing in the Netherlands, significant hospital variation remains. MILR benefits short-term outcomes, while overall survival is comparable to open liver surgery.

Hospital variation in revision rates after primary knee arthroplasty was not explained by patient selection: baseline data from 1452 patients in the Danish prospective multicenter cohort study, SPARK.

PURPOSE: Revision rates following primary knee arthroplasty vary by country, region and hospital. The SPARK study was initiated to compare primary surgery across three Danish regions with consistently different revision rates. The present study investigated whether the variations were associated with differences in the primary patient selection. METHODS: A prospective observational cohort study included patients scheduled Sep 2016 Dec 2017 for primary knee arthroplasty (total, medial/lateral unicompartmental or patellofemoral) at three high-volume hospitals, representing regions with 2-year cumulative revision rates of 1, 2 and 5%, respectively. Hospitals were compared with respects to patient demographics, preoperative patient-reported outcome measures, motivations for surgery, implant selection, radiological osteoarthritis and the regional incidence of primary surgery. Statistical tests (parametric and non-parametric) comprised all three hospitals. RESULTS: Baseline data was provided by 1452 patients (89% of included patients, 56% of available patients). Patients in Copenhagen (Herlev-Gentofte Hospital, high-revision) were older (68.6 ± 9 years) than those in low-revision hospitals (Aarhus 66.6 ± 10 y. and Aalborg (Farsø) 67.3 ± 9 y., p = 0.002). In Aalborg, patients who had higher Body Mass Index (mean 30.2 kg/m2 versus 28.2 (Aarhus) and 28.7 kg/m2 (Copenhagen), p < 0.001), were more likely to be male (56% versus 45 and 43%, respectively, p = 0.002), and exhibited fewer anxiety and depression symptoms (EQ-5D-5L) (24% versus 34 and 38%, p = 0.01). The preoperative Oxford Knee Score (23.3 ± 7), UCLA Activity Scale (4.7 ± 2), range of motion (Copenhagen Knee ROM Scale) and patient motivations were comparable across hospitals but varied with implant type. Radiological classification ≥ 2 was observed in 94% (Kellgren-Lawrence) and 67% (Ahlbäck) and was more frequent in Aarhus (low-revision) (p ≤ 0.02), where unicompartmental implants were utilized most (49% versus 14 (Aalborg) and 23% (Copenhagen), p < 0.001). In the Capital Region (Copenhagen), the incidence of surgery was 15-28% higher (p < 0.001). CONCLUSION: Patient-reported outcome measures prior to primary knee arthroplasty were comparable across hospitals with differing revision rates. While radiographic classifications and surgical incidence indicated higher thresholds for primary surgery in one low-revision hospital, most variations in patient and implant selection were contrary to well-known revision risk factors, suggesting that patient selection differences alone were unlikely to be responsible for the observed variation in revision rates across Danish hospitals. LEVEL OF EVIDENCE: II, Prospective cohort study.

Instrumental variable analysis to estimate treatment effects: a simulation study showing potential benefits of conditioning on hospital.

BACKGROUND: Instrumental variable (IV) analysis holds the potential to estimate treatment effects from observational data. IV analysis potentially circumvents unmeasured confounding but makes a number of assumptions, such as that the IV shares no common cause with the outcome. When using treatment preference as an instrument, a common cause, such as a preference regarding related treatments, may exist. We aimed to explore the validity and precision of a variant of IV analysis where we additionally adjust for the provider: adjusted IV analysis. METHODS: A treatment effect on an ordinal outcome was simulated (beta - 0.5 in logistic regression) for 15.000 patients, based on a large data set (the IMPACT data, n = 8799) using different scenarios including measured and unmeasured confounders, and a common cause of IV and outcome. We compared estimated treatment effects with patient-level adjustment for confounders, IV with treatment preference as the instrument, and adjusted IV, with hospital added as a fixed effect in the regression models. RESULTS: The use of patient-level adjustment resulted in biased estimates for all the analyses that included unmeasured confounders, IV analysis was less confounded, but also less reliable. With correlation between treatment preference and hospital characteristics (a common cause) estimates were skewed for regular IV analysis, but not for adjusted IV analysis. CONCLUSION: When using IV analysis for comparing hospitals, some limitations of regular IV analysis can be overcome by adjusting for a common cause. TRIAL REGISTRATION: We do not report the results of a health care intervention.

Evidence-practice gaps in P2Y12 inhibitor use after hospitalisation for acute myocardial infarction: findings from a new population-level data linkage in Australia.

BACKGROUND: P2Y12 inhibitor therapy is recommended for 12 months in patients hospitalised for acute myocardial infarction (AMI) unless the bleeding risk is high. AIMS: To describe real-world use of P2Y12 inhibitor therapy following AMI hospitalisation. METHODS: We used population-level linked hospital data to identify all patients discharged from a public hospital with a primary diagnosis of AMI between July 2011 and June 2013 in New South Wales and Victoria, Australia. We used dispensing claims to examine dispensing of a P2Y12 inhibitor (clopidogrel, prasugrel or ticagrelor) within 30 days of discharge and multilevel models to identify predictors of post-discharge dispensing and persistence of therapy to 1 year. RESULTS: We identified 31 848 patients hospitalised for AMI, of whom 56.8% were dispensed a P2Y12 inhibitor within 30 days of discharge. The proportion of patients with post-discharge dispensing varied between hospitals (interquartile range: 25.0-56.5%), and significant between-hospital variation remained after adjusting for patient characteristics. Patient factors associated with the lowest likelihood of post-discharge dispensing were: having undergone coronary artery bypass grafting (odds ratio (OR): 0.17; 95% confidence intervals (CI): 0.15-0.20); having oral anticoagulants dispensed 180 days before or 30 days after discharge (OR: 0.39, 95% CI: 0.35-0.44); major bleeding (OR: 0.68, 95% CI: 0.61-0.76); or being aged ≥85 years (OR: 0.68, 95% CI: 0.62-0.75). A total of 26.8% of patients who were dispensed a P2Y12 inhibitor post-discharge discontinued therapy within 1 year. CONCLUSION: Post-hospitalisation use of P2Y12 inhibitor therapy in AMI patients is low and varies substantially by hospital of discharge. Our findings suggest strategies addressing both health system (hospital and physician) and patient factors are needed to close this evidence-practice gap.

Can We Cluster ICU Treatment Strategies for Traumatic Brain Injury by Hospital Treatment Preferences?

BACKGROUND: In traumatic brain injury (TBI), large between-center differences in treatment and outcome for patients managed in the intensive care unit (ICU) have been shown. The aim of this study is to explore if European neurotrauma centers can be clustered, based on their treatment preference in different domains of TBI care in the ICU. METHODS: Provider profiles of centers participating in the Collaborative European Neurotrauma Effectiveness Research in TBI study were used to assess correlations within and between the predefined domains: intracranial pressure monitoring, coagulation and transfusion, surgery, prophylactic antibiotics, and more general ICU treatment policies. Hierarchical clustering using Ward's minimum variance method was applied to group data with the highest similarity. Heat maps were used to visualize whether hospitals could be grouped to uncover types of hospitals adhering to certain treatment strategies. RESULTS: Provider profiles were available from 66 centers in 20 different countries in Europe and Israel. Correlations within most of the predefined domains varied from low to high correlations (mean correlation coefficients 0.2-0.7). Correlations between domains were lower, with mean correlation coefficients of 0.2. Cluster analysis showed that policies could be grouped, but hospitals could not be grouped based on their preference. CONCLUSIONS: Although correlations between treatment policies within domains were found, the failure to cluster hospitals indicates that a specific treatment choice within a domain is not a proxy for other treatment choices within or outside the domain. These results imply that studying the effects of specific TBI interventions on outcome can be based on between-center variation without being substantially confounded by other treatments. TRIAL REGISTRATION: We do not report the results of a health care intervention.

Hospital variation and outcomes of simultaneous resection of primary colorectal tumour and liver metastases: a population-based study.

BACKGROUND: The optimal treatment sequence for patients with synchronous colorectal liver metastases (CRLM) remains uncertain. This study aimed to assess factors associated with the use of simultaneous resections and impact on hospital variation. METHOD: This population-based study included all patients who underwent liver surgery for synchronous colorectal liver metastases between 2014 and 2019 in the Netherlands. Factors associated with simultaneous resection were identified. Short-term surgical outcomes of simultaneous resections and factors associated with 30-day major morbidity were evaluated. RESULTS: Of 2146 patients included, 589 (27%) underwent simultaneous resection in 28 hospitals. Simultaneous resection was associated with age, sex, BMI, number, size and bilobar distribution of CRLM, and administration of preoperative chemotherapy. More minimally invasive and minor resections were performed in the simultaneous group. Hospital variation was present (range 2.4%-83.3%) with several hospitals performing simultaneous procedures more and less frequently than expected. Simultaneous resection resulted in 13% 30-day major morbidity, and 1% mortality. ASA classification ≥3 was independently associated with higher 30-day major morbidity after simultaneous resection (aOR 1.97, CI 1.10-3.42, p = 0.018). CONCLUSION: Distinctive patient and tumour characteristics influence the choice for simultaneous resection. Remarkable hospital variation is present in the Netherlands.

Neonatal Intensive Care Utilization and Postdischarge Newborn Outcomes: A Population-based Study of Texas Medicaid Insured Infants.

OBJECTIVE: To test the hypothesis that newborn infants cared for in hospitals with greater utilization of neonatal intensive care experienced fewer postdischarge adverse events. STUDY DESIGN: We developed 3 retrospective population-based cohorts of Texas Medicaid insured singletons born in 2010-2014 (very low birth weight [VLBW n = 11 139], late preterm [n = 57 509], and non-preterm [n = 664 447]) who received care in higher volume hospitals with level III/IV neonatal intensive care units (NICUs). Measures of NICU care were hospital-level risk adjusted NICU admission rates, special care days (days of nonroutine care) per infant, and the percent of intensive (highest billable care code) special care days. The units of analysis were hospitals (n = 80) and the primary outcome was an adverse event, (defined as admission, emergency department visit, or death) within 30 days postdischarge. RESULTS: Higher use of NICU care at a hospital level was not associated with lower postdischarge 30-day adverse event. Infants cared for in hospitals with above vs below median special care day rates experienced slightly higher postdischarge adverse event per 100 infants (VLBW: 14.01 [95% CI 12.74-15.27] vs 11.84 [10.52-13.16], P < .05; late preterm: 7.33 [6.68-7.97] vs 6.28 [5.87-6.69], P < .01; non-preterm: 4.47 [4.17-4.76] vs 3.97 [3.75-4.18], P < .01). Weak positive associations (Pearson correlations of 0.31-0.37, P < .01) were observed for adverse event with special care days; in no instance was a negative association observed between NICU utilization and adverse event. CONCLUSION: Higher utilization of NICU care was not associated with lower rates of short-term events suggesting that there may be opportunities to safely decrease admission rates and length of NICU stays.