ED Throughput Improvements. Dr. David Allard, Chief Medical Information Officer Dr. Bruce Muma, Chief Medical Officer, Henry Ford Physician Network

Transcription

1 ED Throughput Improvements Dr. David Allard, Chief Medical Information Officer Dr. Bruce Muma, Chief Medical Officer, Henry Ford Physician Network

2 Problem Statement The Henry Ford Accountable Care Organization (HF ACO) is participating in the CMS Next Generation Model which offers significant upside/downside financial risk for total cost of care and measurable quality of care The primary challenge (problem) is to reduce non value added medical expense while simultaneously improving preventive care and chronic disease care. The secondary challenge is to organize the beneficiary data into a valid registry within the EMR/EDW to facilitate population analytics and develop real time interventions.

3 Problem Statement Analysis of data sources led the HF ACO to focus on three hot spots of value improvement opportunity: High Risk Populations The top 30% based on future risk predictive models using claims and EMR data. The waste is driven by discontinuity and inadequate clinical/social supports Variation in Medical Decision Making Focus on high cost medical decisions (global). The waste is driven by variation in risk aversion and inconsistent application of standard protocols Post Acute Care Focus on SNF expenses. The waste is driven by inconsistent adherence to communication protocols, variability in care planning, and absence of provider oversight.

4 Design and implementation The ACO is governed by a Board of Trustees and managed by a team of dedicated Population Health division staff who are accountable to the board for achieving performance targets The Population Health staff developed a comprehensive work plan in late 2015 with a focus on reducing total cost of care, improving selected quality measures, engaging providers and patients, communication and building analytic tools with dashboards to drive continued success. The population was analyzed using a projected beneficiary attribution model derived from both EMR and EDW data elements. This model allowed us to estimate financial and utilization performance in real time and identify hot spot opportunities.

5 Design and implementation Specific interventions were developed for each of the hot spots utilizing the ACO registry and creating patient lists for targeted interventions. The following interventions were developed and launched: High Risk Populations Top 5%: Comprehensive Care Centers Top 30%: Universal Case Management Program Variation in Medical Decision Making Hospital Admission Decision in ED: EDS Program Specialty Referral Decision: Referring Wisely Program Targeted treatment/testing: Choosing Wisely BPA Program Post Acute Care SNF length of stay/transition pathways: PAC Surveillance

6 EDS Program Objective: provide alternative pathways for ED physicians to avoid hospitalization of patients not meeting IQ criteria for admission Scope: all HF ACO, HAP MA and HAP commercial patients presenting to the HFH/HFWBH/Fairlane ED Intervention: EDS navigator resides in the ED and is equipped with tools to orchestrate OPD care pathways (~15 pathways) Technology: EMR track board identifies targeted patients for EDS navigator EMR flow sheets to track intervention/impact Patient scheduling software to facilitate appointments/treatments EDW database to track utilization impact and potential harm

7 Design and implementation (how we used technology) ACO Registries were established for HAP Medicare Advantage and NextGen ACO. Patient lists are received from the insurer, matched with HFHS MRN and then automatically moved into Epic using a datalink process on an ongoing basis CMS sends periodic updates on enrollees in the program which sometimes need to get back dated to January or sometimes only affect membership going forward Enrollees are compared to known patients with a matching algorithm using demographics.

8 Registry population CMS update files can be additions or removal updates which must be back dated to the beginning of the year Expired patients are removed from reports as of the date of death Updates are moved up to the EMR to keep registries current Transactional patient data fed to EDW Registries updates based on upload CMS ACO Update files received by EDW Files matched to MRN with weighted algorithm Patient updates migrated to EMR

9 Registries used to alert users of Status

10 ED Trackboard Alerts

11 Creating actions

12 How was HIT utilized 2 Registries used to drive improvements in CMS hcc capture Multi provider schedules shows the number of conditions that need refreshed Best Practice Advisories automatically fire for the providers and identify the specific conditions that need to be addressed

13 Dashboard Reports are available for various populations by Provider, Department and Care Team member In addition to PCPs, Ambulatory Case Managers, Population Health Coaches and Diabetes Educators can follow their own panels of patients

14 Risk Scoring Tools are available to identify patients that may need more intensive management Chronic Disease Registries allow CareTeam members to focus on specific cohorts of patients this example is CHF patients and is stratified by risk score

15 Allignment of clinical data with registries allows comprehensive dashboard reporting

16 Self Serve Access to Near Real Time Data Mart On the fly queries (one day delay in data) Allows identification of patients and immediate bulk actions

17 EDS Results 29% of all ED visits at participant hospitals by target patient population were reviewed by EDS Navigator To date 235 admission avoided (IPD and OBS) which is equivalent to an avoidance rate of 4.9%. To date 141 additional patients had interventions which provided direct benefit (potential avoidance of future admission or other additional costs associated with delays).

18 All ED Navigator Impact on Disposition (08/13/ /02/2017) IPD avoided IPD downgraded to observation Observation avoided Number of Cases Aug 2016 Sep 2016 Oct 2016 Nov 2016 Dec 2016 Jan 2017 Feb 2017 Mar 2017 Apr 2017 May 2017 Jun 2017 Jul 2017 Aug 2017

19 All ED Population (08/01/ /01/2017) HF ACO Patients HFACO Patients HAP Commercial Patients Number of Patients Aug 2016 Sep 2016 Oct 2016 Nov 2016 Dec 2016 Jan 2017 Feb 2017 Mar 2017 Apr 2017 May 2017 Jun 2017 Jul 2017 All ED Arrival Trends (08/13/ /31/2017) Average number of Patients per hour

20 Admits/1000 (YTD)

21 All ED Navigator Facilitated Interventions: Impacted Disposition Scheduled Diagnostic Test (blank) Facilitated direct admission to a Skilled Nursing Facility Referred Patient to Community Paramedicine Referred to Diabetes Care Center DME PACE Count of Interventions Percentage of Patients with Revisits to Any ED within 72 hours (where Navigator Impacted Disposition OBS & IPD Avoided) No Revisits to ED 86% Revisits to ED 14%

22 Comprehensive Care Centers Objective: Provide a more effective care model for high risk patients to reduce total cost of care, hospitalizations, ED visits and improve health and quality of live. Scope: All HF ACO, HAP MA and HAP commercial patients living within the 10 mile radius of the 2 sites (Taylor and DNW) Intervention: Enroll targets patients in the program which offers comprehensive support with 24/7 availability of providers, case managers, virtual behavioral health services, pharmacy support, home care support, longer visits, management of disease exacerbations in clinic (vs ED/OBS).

23 Comprehensive Care Centers (CCC) Technology: Ongoing surveillance of EDW to identify high risk patients and initiate review and recruitment processes (Optum One) EMR embedded flow sheets and intake processes for CCC staff Patient portal services to promote communication with patient/caregiver Virtual video services EDW database and dashboard to track utilization impact and potential harm

24 CCC: an at risk population

25 CCC Results 20% reductions in ED usage and Admission rates

26 CCC Results Overall 4% reduction in PMPM charges

27 PAC Surveillance (PACS) Objective: Conduct ongoing review of care plans for target patients who are admitted to a SNF and provide support for transitional care back to patient centered medical home Scope: all HF ACO, HAP MA and HAP commercial patients admitted to facilities who belong to the HF preferred SNF network Intervention: PACS case managers conduct outreach (phone, onsite visits, chart review) in real time for patients admitted to the SNF

28 PAC Surveillance Technology: EDW registry to identify target patients HIE tracking system to identify target patients admitted to selected SNF s EDS track board to identify potential candidates for direct to SNF transfer Flow sheet tools to document intervention and report to PCP EDW analytic tools to track length of stay and admission rates.

29 PAC Graceful transfer to SNF

30 PAC: Facilitating PCP followup

31 SNF Admits/1000 (YTD) PACS Results With real time monitoring of clinical events, SNF usage and length of stay can be closely monitored and optimized Average Length of Stay (Preferred SNF vs. Non Preferred SNF) Jan 2016 Feb 2016 Mar 2016 Apr 2016 May 2016 Jun 2016 Jul 2016 Aug 2016 Sep 2016 Non Preferred Preferred Oct 2016 Nov 2016 Dec 2016 Jan 2017 Feb 2017 Mar 2017 Apr 2017 May 2017 Jun 2017 Jul 2017