Faculty Bibliography

BACKGROUND:Reducing costs while increasing or maintaining quality is crucial to delivering high value care. OBJECTIVE:To assess the impact of a hospital value-based management programme on cost and quality. DESIGN/METHODS:Time series analysis of non-psychiatric, non-rehabilitation, non-newborn patients discharged between 1 September 2011 and 31 December 2017 from a US urban, academic medical centre. INTERVENTION/METHODS:NYU Langone Health instituted an institution-wide programme in April 2014 to increase value of healthcare, defined as health outcomes achieved per dollar spent. Key features included joint clinical and operational leadership; granular and transparent cost accounting; dedicated project support staff; information technology support; and a departmental shared savings programme. MEASUREMENTS/METHODS:Change in variable direct costs; secondary outcomes included changes in length of stay, readmission and in-hospital mortality. RESULTS:The programme chartered 74 projects targeting opportunities in supply chain management (eg, surgical trays), operational efficiency (eg, discharge optimisation), care of outlier patients (eg, those at end of life) and resource utilisation (eg, blood management). The study cohort included 160 434 hospitalisations. Adjusted variable costs decreased 7.7% over the study period. Admissions with medical diagnosis related groups (DRG) declined an average 0.20% per month relative to baseline. Admissions with surgical DRGs had an early increase in costs of 2.7% followed by 0.37% decrease in costs per month. Mean expense per hospitalisation improved from 13% above median for teaching hospitals to 2% above median. Length of stay decreased by 0.25% per month relative to prior trends (95% CI -0.34 to 0.17): approximately half a day by the end of the study period. There were no significant changes in 30-day same-hospital readmission or in-hospital mortality. Estimated institutional savings after intervention costs were approximately $53.9 million. LIMITATIONS/CONCLUSIONS:Observational analysis. CONCLUSION/CONCLUSIONS:A systematic programme to increase healthcare value by lowering the cost of care without compromising quality is achievable and sustainable over several years.

Background: Clinical decision support (CDS) tools which incorporate clinical prediction rules (CPRs) have the potential to successfully deliver accurate information and guide decision-making at the point of care. Our previously validated integrated clinical prediction rule (iCPR) was designed to guide evidence-based treatment within an electronic health record for streptococcal pharyngitis and pneumonia based on chief complaints of sore throat, cough or upper respiratory infection. In initial testing at a single site, it resulted in high provider tool adoption (58%) and decreased antibiotic prescribing rates (35%) for acute respiratory infections. Our objective for this study was to assess the impact of this tool when adapted and implemented in diverse primary care settings.
Method(s): This was a randomized controlled trial including 33 primary care practices at two large academic health systems in Wisconsin and Utah. Between October 2015 and June 2018 providers in the intervention group were prompted to complete either Centor Score or Heckerling Rule for Pneumonia based onthe chief complaint of the patient encounter. EHR data on provider and patient demographics, tool use rates, and antibiotic order rates from 541 providers and 100,573 monitored patient encounters were collected for analysis. Risk ratios, CIs, and P values are calculated from a generalized estimating equation log-binomial model adjusting for clustering of orders or visits by provider and using robust standard error estimators.
Result(s): The tool was triggered 42,126 times among 214 intervention providers and was completed in 6.9% of eligible visits. The intervention and control groups prescribed antibiotics in 35% and 36% of visits respectively and were not significantly different. There were no differences in rates for rapid streptococcal test or chest X-ray orders between groups (Strep: relative risk, 1.0; P=.11; Pneumonia: relative risk, 1.8; P=.64).
Conclusion(s): In diverse primary care settings, the tool was not effective at reducing unnecessary antibiotic prescription and diagnostic testing. This outcome was possibly driven by low overall use of CDS tools highlighting the growing impact of " alert fatigue" and the need for new approaches to enhance provider engagement with CDS tools. New strategies for reducing the persistently high rates of inappropriate antibiotic prescribing for acute respiratory infections are needed. Novel approaches in future studies are necessary for reducing barriers to CDS tools in order to increase use and engagement

Statement of Problem Or Question (One Sentence): Underutilization of clinical prediction rules and poor uptake of provider-oriented clinical decision support (CDS) has contributed to overuse of antibiotics for sore throat. Objectives of Program/Intervention (No More Than Three Objectives): 1. Adapt CDS for registered nurses (RNs) to evaluate and treat patients with sore throat 2. Demonstrate the feasibility of RN visits using CDS to evaluate and treat patients with sore throat Description of Program/Intervention, Including Organizational Context (E.G. Inpatient Vs. Outpatient, Practice or Community Characteristics): We performed a 12-week pilot study to evaluate the feasibility of RN visits using an integrated clinical prediction rule (iCPR) tool to determine patient risk for strep throat and provide appropriate treatment at a family medicine clinic in a Midwest academic healthcare system. iCPR, originally developed for use by primary care physicians (PCPS), includes a risk calculator using Centor strep throat criteria and ordersets based on patient's risk for strep throat: education for low-risk, testing for intermediate-risk, and testing or antibiotics for high-risk. To adapt the process for RN visits, we developed triage protocols so appropriate patients received nurse visits, very low risk received education and more complex patients received provider visits. No major changes were made to the risk calculator or ordersets. Four RNs, with 2-24 years of experience, received a 10-minute online training session on sore throat evaluation followed by a 45-minute in-person training on physical examination and iCPR use. RNs triaged patients by phone and conducted RN visits using iCPR and following orderset recommendations. RNs could transition to a PCP visit if they were uncomfortable evaluating the patient. Measures of Success (DISCUSS QUALITATIVE AND/OR QUANTITATIVE METRICS WHICH WILL BE USED TO EVALUATE PROGRAM/INTERVENTION): Electronic health record data was used to determine the number of nurse visits, frequency of tool use and antibiotic and diagnostic test ordering. RNs completed a self-efficacy survey prior to training and 8-weeks after implementation. At 12 weeks, we interviewed RNs to understand barriers and facilitators to using the tool. Findings To Date (It Is Not Sufficient To State Findings Will Be Discussed): 162 triage calls for sore throat resulted in 77(48%) patients with RN-only visits, 45(28%) with provider visits, 38(23%) with no visit. Only 2 RN visits (< 3%) converted to provider visit due to patient complexity. RNs completed the risk calculator for 99% of visits and followed recommendations in all cases except for ordering antibiotics in 1 high-risk patient with a negative rapid strep. RN confidence in their ability to evaluate and treat a patient with sore throat was 85 (SD 5.8) (0 cannot do at all; 100 highly certain I can do) prior to training and 97.5 (SD 5.0) at 8-weeks. RNs felt the tool decreased provider visits and strep testing in patients. RN's also felt that the tool increased patient and RN satisfaction. Key Lessons For Dissemination (What Can Others Take Away For Implementation To Their Practice Or Community?): This pilot study demonstrates that RNs can use CDS to appropriately triage, evaluate and treat acute low-complexity sore throat patients. Implementation of an RN-driven iCPR tool shows promise to reduce inappropriate antibiotic prescribing and represents a potential model for expanding RN practice using CDS

Background: Discharge to skilled nursing facilities (SNF) is common in patients with heart failure (HF). The goal of a SNF stay is to improve functional status to allow patients to return home safely. However, the second transition from SNF to home may also be risky. Here, we examine the association between receipt of home health care (HHC) and readmission risk among patients discharged from SNF to home following HF hospitalization.
Method(s): We examined all Medicare fee-for-service beneficiaries 65 and older admitted 2012-2015 with a HF diagnosis discharged to SNF then subsequently discharged home. The primary outcome was unplanned read-mission within 30 days of SNF to home discharge, using CMS's HF read-mission methodology. We plotted time to readmission with Kaplan-Meier curves and compared these groups with a log-rank test. Then, we compared time to readmission using an adjusted Cox model; this model included a frailty term to account for correlation of patient outcome by SNF.
Result(s): There were 67,585 HF hospitalizations discharged to SNF and subsequently discharged home; 13,257 (19.6%) were discharged with HHC, 54,328 (80.4%) without. Patients discharged home from SNF with HHC had lower 30-day readmission rates than patients discharged without HHC (22.8% vs 24.5%, p< 0.0001). Kaplan-Meier curves demonstrated that patients discharged home from SNF with HHC have a longer unadjusted time to readmission. Of those readmitted within 30 days, median time to readmission for those discharged home from SNF with HHC was 11 days and 9 days for those discharged home without HHC (p< 0.0001). After risk-adjustment, patients discharged home with HHC still had a lower hazard of 30-day readmission.
Conclusion(s): Patients who received HHC were less likely to be readmitted within 30 days compared to those discharged home without HHC. This is unexpected as patients discharged with HHC likely have more functional impairments and therefore at higher readmission risk. Since patients requiring a SNF stay after hospital discharge may have additional needs, they may be especially likely to benefit from restorative therapy through HHC; however only about 20% received such services

OBJECTIVE:Discharge to skilled nursing facilities (SNFs) is common in patients with heart failure (HF). It is unknown whether the transition from SNF to home is risky for these patients. Our objective was to study outcomes for the 30 days after discharge from SNF to home among Medicare patients hospitalized with HF who had subsequent SNF stays of 30 days or less. DESIGN/METHODS:Retrospective cohort study. SETTING AND PARTICIPANTS/METHODS:All Medicare fee-for-service beneficiaries 65 and older admitted during 2012-2015 with a HF diagnosis discharged to SNF then subsequently discharged home. MEASURES/METHODS:Patients were followed for 30 days following SNF discharge. We categorized patients by SNF length of stay: 1 to 6 days, 7 to 13 days, and 14 to 30 days. For each group, we modeled time to a composite outcome of unplanned readmission or death after SNF discharge. Our model examined 0-2 days and 3-30 days post-SNF discharge. RESULTS:Our study included 67,585 HF hospitalizations discharged to SNF and subsequently discharged home. Overall, 16,333 (24.2%) SNF discharges to home were readmitted within 30 days of SNF discharge. The hazard rate of the composite outcome for each group was significantly increased on days 0 to 2 after SNF discharge compared to days 3 to 30, as reflected in their hazard rate ratios: for patients with SNF length of stay 1 to 6 days, 4.60 (4.23-5.00); SNF length of stay 7 to 13 days, 2.61 (2.45-2.78); SNF length of stay 14 to 30 days, 1.70 (1.62-1.78). CONCLUSIONS/IMPLICATIONS/CONCLUSIONS:The hazard rate of readmission after SNF discharge following HF hospitalization is highest during the first 2 days home. This risk attenuated with longer SNF length of stay. Interventions to improve postdischarge outcomes have primarily focused on hospital discharge. This evidence suggests that interventions to reduce readmissions may be more effective if they also incorporate the SNF-to-home transition.

OBJECTIVES/OBJECTIVE:Respiratory infections are associated with acute exacerbations of asthma and accompanying morbidity and mortality. In this study we explore inter-practice variations in respiratory infections in children with asthma and study the effect of practice-level factors on these variations. DESIGN/METHODS:Cross-sectional study. SETTING/METHODS:We analysed data from 164 general practices in the Royal College of General PractitionersResearch and Surveillance Centresentinel network in England. PARTICIPANTS/METHODS:Children 5-12 years. INTERVENTIONS/METHODS:None. In this observational study, we used regression analysis to explore the impact of practice-level determinants on the number of respiratory infections in children with asthma. PRIMARY AND SECONDARY OUTCOME MEASURES/UNASSIGNED:We describe the distribution of childhood asthma and the determinants of upper/lower respiratory tract infections in these children. RESULTS:83.5% (137/164) practices were in urban locations; the mean number of general practitioners per practice was 7; and the mean duration since qualification 19.7 years. We found almost 10-fold difference in the rate of asthma (1.5-11.8 per 100 children) and 50-fold variation in respiratory infection rates between practices. Larger practices with larger lists of asthmatic children had greater rates of respiratory infections among these children. CONCLUSION/CONCLUSIONS:We showed that structural/environmental variables are consistent predictors of a range of respiratory infections among children with asthma. However, contradictory results between measures of practice clinical care show that a purely structural explanation for variability in respiratory infections is limited. Further research is needed to understand how the practice factors influence individual risk behaviours relevant to respiratory infections.

According to the U.S. census Bureau, close to 20% of the U.S. population speaks a language other than English at home. Home health care (HHC) patients who speak English less than very well or have limited English proficiency (LEP) are at an increased risk for medication mismanagement and serious health consequences. The purpose of this study was to examine if there were differences in medication management between English-speaking patients and patients with LEP receiving HHC services. Data for this cross-sectional observation study were collected from 2010 to 2014. Medication management was measured by two items in the Centers for Medicare and Medicaid Services"“mandated Outcomes Assessment Information Set (OASIS). All patients in the database who were taking medications and had a valid admission and discharge assessment from HHC were included in the analysis. Inverse probability of treatment weighting (IPTW) with a marginal structural model was used to address potential imbalances in observed patient characteristics when estimating the effect of having LEP or being an English-speaking HHC patient on changes in medication management over the course of a HHC episode. Estimates from marginal structural model with inverse probability weighting indicate that being LEP was associated with less improvement in medication management and increased likelihood of getting worse over the course of a HHC episode. This study is one of the first to demonstrate that patients with LEP experience disparities in medication management when compared to English-speaking patients in HHC.

The mechanisms underlying the adverse cardiovascular effects of increased salt intake are incompletely understood, but parallel increases in serum sodium concentration may be of importance. The aim of this retrospective cohort study was to investigate the relationship between serum sodium, hypertension and incident cardiovascular disease (CVD). Routinely collected primary care data from the Royal College of General Practitioners Research and Surveillance Centre were analysed. A total of 231,545 individuals with a measurement of serum sodium concentration at baseline were included. Exclusion criteria were: age < 40 years; abnormal serum sodium; diabetes mellitus; prior CVD event; stage 5 chronic kidney disease; and liver cirrhosis. The primary outcome was incident CVD (myocardial infarction, acute coronary syndrome, coronary revascularisation, stroke, transient ischaemic attack or new heart failure diagnosis) over 5 years. There was a 'J-shaped' relationship between serum sodium concentration and primary cardiovascular events that was independent of established risk factors, medications and other serum electrolytes. The lowest cardiovascular risk was found with a serum sodium between 141 and 143 mmol/l. Higher serum sodium was associated with increased risk in hypertensive individuals, whereas lower concentrations were associated with increased risk in all individuals. Therefore, alterations in serum sodium concentration may be a useful indicator of CVD risk. Higher serum sodium could have a direct effect on the vasculature, particularly in hypertensive individuals. Lower serum sodium may be a reflection of complex volume and neuroendocrine changes.

BACKGROUND:Hypertension is a major cause of preventable disability and death globally and affects more than one in four adults in England. Unwarranted variation is variation in access, quality, outcome or value which is unexplained by differences in the condition or patient characteristics and which reduces quality and efficiency. Distinguishing unwarranted from variation due to clinical, organisational or patient factors can be challenging. We carried out this study to explore inter-practice variation in the diagnosis and management of hypertension in the Royal College of General Practitioners (RCGP) Research and Surveillance Centre (RSC) network database, a large, representative surveillance database. METHODS AND FINDING/UNASSIGNED:We carried out a cross-sectional study using primary care data extracted from the electronic health records of 1,271,419 adults registered at RCGP RSC general practices on 31st December 2016. Logistic regression was used to indirectly standardise practice-level hypertension prevalence and control against the RCGP RSC population, adjusted for age, gender, ethnicity, deprivation, co-morbidity, NHS region and practice size. Inter-practice variation was demonstrated using funnel plots with 95% and 99.8% control limits. The prevalence of detected hypertension was 18.4% (95% CI 18.4-18.5), n = 234,165. Uncontrolled hypertension was present in 146,553 of 196,052 individuals, 25.2% (25.1-25.4), in whom blood pressure had been recorded in the previous year. Hypertension management varied markedly between practices with a three-fold difference in prevalence, 13.5-38.4%, and a four-fold difference in the proportion of uncontrolled hypertension, 11.8-47.9%. Despite adjustment for sociodemographic and practice characteristics funnel plots demonstrated marked over-dispersion. CONCLUSIONS:Substantial variation in the prevalence of diagnosed hypertension and the management of hypertension was only partially explained by characteristics captured within a routine dataset. The over-dispersion suggests variation is not fully explained by these factors and that context, behaviour and processes of care delivery may contribute to variation. Routine data sources in isolation to not provide sufficient contextual data to diagnose the causes of variation.