Faculty Bibliography

Objective: The purpose of this study was to determine whether an electronic health record-based sepsis alert system could improve quality of care and clinical outcomes for patients with sepsis. Materials and Methods: We performed a patient-level interrupted time series study of emergency department patients with severe sepsis or septic shock between January 2013 and April 2015. The intervention, introduced in February 2014, was a system of interruptive sepsis alerts triggered by abnormal vital signs or laboratory results. Primary outcomes were length of stay (LOS) and in-hospital mortality; other outcomes included time to first lactate and blood cultures prior to antibiotics. We also assessed sensitivity, positive predictive value (PPV), and clinician response to the alerts. Results: Mean LOS for patients with sepsis decreased from 10.1 to 8.6 days ( P < .001) following alert introduction. In adjusted time series analysis, the intervention was associated with a decreased LOS of 16% (95% CI, 5%-25%; P = .007, with significance of alpha = 0.006) and no change thereafter (0%; 95% CI, -2%, 2%). The sepsis alert system had no effect on mortality or other clinical or process measures. The intervention had a sensitivity of 80.4% and a PPV of 14.6%. Discussion: Alerting based on simple laboratory and vital sign criteria was insufficient to improve sepsis outcomes. Alert fatigue due to the low PPV is likely the primary contributor to these results. Conclusion: A more sophisticated algorithm for sepsis identification is needed to improve outcomes.

Background: Paper-based Early Recovery after Surgery (ERAS) path-ways have been shown to reduce length of stay, but there have been limited evaluations of electronic health record (EHR) based pathways. The objective of this study was to evaluate whether ERAS processes implemented with a novel pathway activity integrated with the EHR (E-Pathway) can reduce costs without resulting in increased 30 day readmissions. Methods: We performed a retrospective cohort study of surgical patients age>= 18 years hospitalized at an academic medical center from March 1, 2013 to August 31, 2016. The primary cohort consisted of patients admitted for elective colon surgery. We also studied a control group of patients undergoing elective procedures with similar length of stay as colon surgery (3-5 days). The E-Pathway was based on a pathway template developed by a common EHR vendor (Epic Systems, Verona, WI) with content developed by a multidisciplinary team based on ERAS principles. The E-Pathway was implemented on March 2, 2015. The primary outcome was variable costs per case. Secondary outcomes were observed to expected length of stay (O:E LOS) and 30 day readmissions to our hospital. For both groups, we performed an interrupted time series with segmented regression analysis with month being the unit of time. We used gamma regression for cost models and logistic models for the secondary outcomes. Results: We included 823 (470 and 353 in the pre-and post-intervention, respectively) colon surgery patients and 3415 (1,819 and 1,596 in the pre-and post-intervention) surgical control patients. Among the colon surgery cohort, we observed no changesin cost eitheratbaseline [-0.1% (95% CI-0.8%, 0.5%) per month] or with immediate introduction of the pathway. However, there was statistically significant (p = 0.040) decrease in costs of 1.3% (0.6% to 2.5%) per surgical encounter per month over the 18 month post intervention period. The surgical comparator group had no change in cost either at baseline or at the time of intervention and had a nonsignificant decrease in monthly costs of 0.6% (p = 0.231) post-intervention. There was statistically significant (p = 0.039) decrease in the O:E slope after the intervention of 1.49% per surgical encounter per month. The surgical comparator group had a nonsignificant (p = 0.761) increase in slope of 1.87%. For the 30 day readmission rates, there were no statistically significant changes in either the colon surgery or control groups. Conclusions: Our study is the first, to our knowledge, to report on the outcomes of a novel sophisticated E pathway integrated into an EHR. Following implementation of the E-pathway for colon surgery patients, we observed decreasing direct variable costs and O:E LOS over time. These improvements were not observed among comparable surgical patients. Consequently, as institutions continue to place increased emphasis on standardization of best practice care, E-pathways can be powerful vehicles to support those changes in the new EHR-centric care model

Background: Clinical decision support (CDS) has been shown to im-prove compliance with evidence-based care but its impact is often diminished due to issues such as poor usability, insufficient integration into workflow, and alert fatigue. Non-interruptive CDS may be less subject to alert fatigue but there has been little assessment of its usability. The purpose of this study was to perform usability testing on interruptive and non-interruptive versions of a CDS. Methods: We conducted a usability study ofa CDS tool that recommended prescribing an angiotensin converting enzyme (ACE) inhibitor for inpatients with heart failure. We developed two versions of the CDS that varied in its format: an interruptive alert, in which the CDS popped-up at the time of order entry, and a non-interruptive alert, which was displayed in a checklist section of the Electronic Health Record (EHR). We recruited inpatient providers to use both versions in a laboratory setting. We randomly assigned providers to first trigger the interruptive or non-interruptive alert. Providers were given a clinical scenario and asked to " think aloud" as they worked through the CDS; we then conducted a brief semi-structured interview about usability. We used a constant comparative analysis informed by the Five Rights of CDS framework to analyze the interviews. Inpatient providers from different disciplines were recruited until thematic saturation was reached. Results: Of 12 providers who participated in usability testing, seven used the interruptive followed by the non-interruptive CDS and five used the non-interruptive CDS initially. We categorized codes into four themes related to the Five Rights of CDS and determined some codes to be general to the CDS while others were specific to the interruptive or non-interruptive version (Table). Providers noted that the interruptive alert was readily noticed but generally impeded workflow. Providers found the non-interruptive CDS to be less annoying but had lower visibility; although they liked the ability to address the non-interruptive CDS at any time, some providers questioned whether it would ultimately be used. Conclusions: Providers expressed annoyance in working with an inter-ruptive CDS. Although the non-interruptive CDS was more appealing, providers admitted that it may not be used unless integrated with workflow. One potential solution was a combination of the two formats: supplementing a non-interruptive alert with an occasional, well-timed interruptive alert if uptake was insufficient

INTRODUCTION: Orthopaedic care is shifting to alternative payment models. We examined whether New York University Langone Medical Center achieved savings under the Centers for Medicare and Medicaid Services Bundled Payments for Care Improvement initiative. METHODS: This study was a difference-in-differences study of Medicare fee-for-service patients hospitalized from April 2011 to June 2012 and October 2013 to December 2014 for lower extremity joint arthroplasty, cardiac valve procedures, or spine surgery (intervention groups), or for congestive heart failure, major bowel procedures, medical peripheral vascular disorders, medical noninfectious orthopaedic care, or stroke (control group). We examined total episode costs and costs by service category. RESULTS: We included 2,940 intervention episodes and 1,474 control episodes. Relative to the trend in the control group, lower extremity joint arthroplasty episodes achieved the greatest savings: adjusted average episode cost during the intervention period decreased by $3,017 (95% confidence interval [CI], -$6,066 to $31). For cardiac procedures, the adjusted average episode cost decreased by $2,999 (95% CI, -$8,103 to $2,105), and for spinal fusion, it increased by $8,291 (95% CI, $2,879 to $13,703). Savings were driven predominantly by shifting postdischarge care from inpatient rehabilitation facilities to home. Spinal fusion index admission costs increased because of changes in surgical technique. DISCUSSION: Under bundled payment, New York University Langone Medical Center decreased total episode costs in patients undergoing lower extremity joint arthroplasty. For patients undergoing cardiac valve procedures, evidence of savings was not as strong, and for patients undergoing spinal fusion, total episode costs increased. For all three conditions, the proportion of patients referred to inpatient rehabilitation facilities upon discharge decreased. These changes were not associated with an increase in index hospital length of stay or readmission rate. CONCLUSION: Opportunities for savings under bundled payment may be greater for lower extremity joint arthroplasty than for other conditions.

Background: Heart failure is among the most common reasons for admission to hospital and is associated with high rates of readmission. Studies have shown that the frequency of heart failure hospitalisation in temperate climates increases in winter months. While such findings suggest that the risk of hospitalisation is higher in winter as compared to summer months, there has been little evaluation of whether there is seasonal variation in the readmission risk of patients who are hospitalised for heart failure. Purpose: To examine seasonal variations in: 1) readmission risk for patients hospitalised with heart failure; and 2) frequency of heart failure hospitalisation for patients at different risk of readmission. Methods: We performed a retrospective study of United States Medicare beneficiaries age >=65 who were hospitalised for heart failure between January 1, 2009 and June 30, 2015. We used a predictive model for 30-day unplanned hospital readmission to assign each hospitalisation a predicted risk of readmission; this model adds demographic and prior utilization data to the hospital-wide readmission model used by the Centres for Medicare and Medicaid Services (CMS). Each hospitalisation was categorized as lowest 20%, middle 60%, and highest 20% of predicted readmission risk. We calculated rate of hospitalisations per calendar month, predicted readmission risk, and monthly rate of hospitalisations for each risk stratum in the study period; monthly hospitalisations were standardized to 30 days. When comparing risk strata, we divided the totals for the middle 60% stratum by three. Results: Among 2,661,837 heart failure hospitalisations, we observed the highest rates of hospitalisation in January through March (range 37,185-37,949 per month) and the lowest rates in July through September (range 29,901-30,603 per month). Conversely, predicted readmission rates were lowest in January and highest in August, with rates of 23.1% and 24.0%, respectively. The number of hospitalisations increased in winter months for patients in all three risk strata, with greatest variation in seasonal differences observed for patients in the lowest 20% of predicted risk (Figure). For example, hospitalisation rates for highest risk patients were 7058 per 30 days in January versus 6473 per 30 days in August; for lowest risk patients, these values were 7852 versus 5595, respectively. (Figure Presented) Conclusion: Readmission risk decreased in winter versus summer months for patients hospitalised for heart failure. Much of the seasonal variation in heart failure hospitalisations appears to be due to a large excess of hospitalisations of these low risk patients in winter months. Our results suggest that preventative measures, such as vaccinations or dietary education, that target lower risk patients in colder months may reduce overall utilisation

BACKGROUND: Interventions to reduce readmission following hospitalization for acute decompensated heart failure (ADHF) require early identification of patients. Electronic health record (EHR)-based approaches to identify patients can range in complexity from simple algorithms based on few data elements to machine learning algorithms using big data. The purpose of this study was to compare performance of a range of algorithms to identify patients with ADHF. Given the emphasis that hospitals currently place on patients with ADHF, we developed models with high sensitivity to avoid missed opportunities for care improvement; this approach assumed secondary chart review would be necessary to confirm a diagnosis. We therefore estimated the time needed for secondary review by providers following initial screening with each algorithm. METHODS: We performed a retrospective study of hospitalizations for patients age > 18 at an academicmedical center in 2013-2015. Using a random 75% development set, we developed four algorithms to identify hospitalizations with a principal diagnosis of ADHF using EHR data through the second midnight of hospitalization: 1) one of three characteristics: heart failure on the problem list, loop diuretic use, or brain natriuretic peptide > 500 pg/ml; 2) logistic regression of 31 clinically-relevant data elements; 3) machine learning approach using L1-regularization logistic regression and unstructured data, including provider notes and radiology reports; 4) machine learning with both structured and unstructured data. We assessed performance of each algorithm in the 25% validation set. We also conducted a brief survey of providers who perform chart review for ADHF to estimate time needed for secondary screening following primary screening with each algorithm. RESULTS: We included 37,229 hospitalizations in the study, of which 1,294 (3.5%) carried a principal diagnosis of ADHF. Algorithm 1 had a sensitivity of 0.98 and positive predictive value (PPV) of 0.14 for ADHF. Algorithm 2 had an area under the receiver operating characteristic curve (AUC) of 0.96 and a PPV of 0.15 when setting the sensitivity at 0.98. Both machine learning algorithms had AUCs of 0.99; with a sensitivity of 0.98, algorithms 3 and 4 had PPVs of 0.30 and 0.34, respectively. Based on survey of three providers, we estimated providers spent 8.6 min per chart review; using this this parameter, providers would spend 61.4, 57.3, 28.7, and 25.3 min on secondary chart review for each case of ADHF if initial screening was done with algorithms 1, 2, 3, and 4, respectively. CONCLUSIONS: Traditional approaches with structured data can accurately identify nearly all patients with ADHF but are limited by a large number of false positives. Machine learning algorithms with unstructured notes and radiology reports can retain this sensitivity while significantly reducing false positives, thereby improving provider efficiency for delivery of quality improvement interventions

BACKGROUND: Electronic health record (EHR)-based clinical decision support tools (CDS) incorporate individualized data to produce patient-specific recommendations at the point-of-care. However, these tools are often limited in their effectiveness, which may be due to poor consideration of usability. The purpose of this study was to evaluate the utilization of a CDS intervention to increase prescription of Angiotensin Converting Enzyme inhibitor (ACEi) or Angiotensin Receptor Blocker (ARB) for patients with heart failure. METHODS: We performed a retrospective study of hospitalized patients with heart failure from the time of CDS implementation, 7/10/13, through 11/30/15. The CDS that we investigated offers providers an opportunity to prescribe an ACEi or ARB or report a contraindication to therapy for patients with documented heart failure. All patients with an EF < 40% who were not on an ACEi or ARB at time of discharge were included in the study. We identified the number of patients for whom the CDS triggered; of those, we categorized provider response as: dismissed, ordered an ACEi/ARB, or contraindication reported. We then performed manual chart review to identify the CDS reported contraindication and structured chart abstraction with standard guidelines to identify gold standard contraindications. We compared each CDS contraindication to gold standard contraindications to determine their accuracy. RESULTS: Out of the 618 subjects who had an EF < 40% but no ACEi or ARB at the time or discharge, 435/618 (70%) had a triggered CDS. Of these 435 subjects for who a CDS was triggered, 180 (41%) were dismissed, 225 (52%) had a contraindication response and 30 (7%) had a prescription for an ACEi/ARB therapy. Overall the accuracy of the documented CDS was 42% (Table 1). CONCLUSIONS: The CDS that we reviewed was poorly utilized and contraindications documented in the tool poorly correlated with patient clinical status reflected elsewhere in the EHR. These findings identify this CDS as a possible impedance to user workflow. One way to improve CDS tools at the point of care is through thorough usability testing and consideration of physician workflow prior to implementation. (Table Presented)