Testing the Re-Engineered Discharge (Text Version)

Slide Presentation from the AHRQ 2008 Annual Conference

On September 9, 2008, Brian Jack, M.D., made this presentation at the 2008 Annual Conference. Select to access the PowerPoint® presentation (53 MB).

Slide 1

Testing the Re-Engineered Discharge (RED)

Principal Investigator: Brian Jack, MD
Associate Professor and Vice Chair

Department of Family Medicine
Boston Medical Center /
Boston University School of Medicine

Slide 2

The slide shows an emblem for the Boston Red Sox baseball team.

Slide 3

The slide shows an emblem for the Boston Celtics basketball team.

Slide 4

Plan for Today

1. The Problem
2. How We Got Started
3. National Quality Forum (NQF) "Safe Practice"
4. Randomized Controlled Trial (RCT): Is "Safe Practice" Safer?
5. Can Health Information Technology (IT) Deliver?

Slide 5

"Perfect Storm" of Patient Safety

• Loose Ends—workups NOT completed.
• Communication—DC [discharge] summary not available.
• Poor Quality Info—DC summary lack results.
• Poor Preparation—few pts know meds/dx.
• Fragmentation—who is in charge?

Slide 6

There are Many Discharges and they are Costly

• In 2003 there were over 38 million discharges:
  • That's over $753 BILLION
• 13% of patients are recurrently hospitalized—and use 60% of resources

Slide 7

Patients Are Not Prepared?

Original Article:

Patients' Understanding of Their Treatment Plans and Diagnosis at Discharge
Amgad N. Makaryus, MD, and Eli A. Friedman, MD Mayo Clinic Proceedings August 2005; 80(8):991-4

At Discharge:

• 37.2% able to state purpose of all their medications.
• 14% knew their medication's common side effects.
• 41.9% able to state their diagnosis.

Slide 8

Little Time Spent on DC

• Audiotaped 97 Discharge Encounters.
• 8 Elements—Roter Interactional Analysis:
  • Nurse, Pharmacist, Physician, Nurse Case Manager.
• Averaged 8 minutes (range of 2 to 28.5 min).
• No teachback 84% of the time.
• Patient is a passive participant:
  • Two initiated questions.
• Not comprehensive:
  • 4 or fewer elements covered 50%.

Slide 9

Pending Tests not Followed

Patient Safety Concerns Arising from Test Results That Return after Hospital Discharge. Christopher L. Roy, MD; Eric G. Poon, MD, MPH; Andrew S. Karson, MD, MPH; Zahra Ladak-Merchant, BDS, MPH; Robin E. Johnson, BA; Saverio M. Maviglia, MD, MSc; and Tejal K. Gandhi, MD, MPH Ann Intern Med 2005; 143(2): 121-8.

1095 of 2644 (41%) inpatients discharged with a test result pending

• 9.4% potentially required action.
• 2/3 of MDs unaware of results.
• 37% actionable and 13% urgent.

Slide 10

Work-ups Not Completed

Tying Up Loose Ends: Discharging Patients With Unresolved Medical Issues Carlton Moore, MD; Thomas McGinn, MD, MPH; Ethan Hahn, MD, MPH. Arch Intern Med 2007; 167:1305-11

• ¼ of discharged patients require additional outpatient work-ups.
• >1/3 not completed.

Slide 11

Communication

Impact of patient communication problems on the risk of preventable adverse events in acute care settings. Gillian Bartlett, PhD; Régis Blais, PhD; Robyn Tamblyn, PhD; Richard J. Clermont, MD; and Brenda MacGibbon, PhD. CMAJ June 3, 2008; 178(12).

Patients with communication problems:

• 3X more likely to have adverse event.
• 46% had multiple adverse events.

Slide 12

Communication Deficits at Hospital Discharge are common

• Discharge summary availability:
  • 1st post-discharge appt 12-34%.
  • 51-77% at 4 weeks.
• Discharge summaries often lack:
  • Test results (33-63%).
  • Hospital course (7-22%).
  • Discharge meds (2-40%).
  • Pending test results (65%).
  • Follow-up plans (2-43%).
• Direct communication 3-20%

Note: Kripalani S et al. JAMA 2007;297:831-41.

Slide 13

Discharges are Variable by Day of the Week

Screen shot of a line graph showing how discharges are variable by the day of the week. There is a line representing each day of the week using a different color and/or patterned line. The "x" axis is Days to Rehospitalization from 0-80 days in increments of 20 days. The "y" axis is the discharge rate from 0 to 1.0% in increments of 0.4, 0.6, 0.8 and 1.0.

Slide 14

An Etiologic Classification of Adverse Events at Hospital Discharge

The slide presents a chart of various, adverse discharge events that could occur under the "Health Care System," "Patient," and "Clinician" leading to "Rehospitalization." Red circles are placed around "Health Care System," "Patient," "Clinician," and "Rehospitalization."

The three categories are listed as follows:

• Health Care System:
  • Lapse of communication:
    • Discharge summary to PC.
    • Inpatient team to Primary Care Physician (PCP).
    • Community services with PCP.
  • Inadequate Patient Education.
  • Medication Error.
  • Lack of timely follow-up.
  • Lapse in community services.

• Patient:
  • New Medical Problem.
  • Deteriorization of known medical problem:
    • Distant from discharge.
    • Early Post-discharge.
  • Drug/Alcohol use.
  • Language/Cultural barrier.
  • Medication non-adherence.
  • Doesn't keep follow-up appointment.

• Clinician::
  • Lab/Test error:
    • Not ordered.
    • Not performed.
    • Not seen.
    • Not acted upon.
  • Inappropriate discharge.
  • Inappropriate medication..
  • Inadequate use of community services.

Slide 15

Errors lead to Adverse Events

Arch Intern Med 2003; 138

• 19% of patients had a post discharge adverse events (AE).
  • 1/3 preventable and 1/3 ameliorable.

CMAJ 2004; 170(3)

• 23% of patients had a post discharge AE:
  • 28% preventable and 22% ameliorable.

Slide 16

Two Questions

We asked:

• Can improving the discharge process reduce adverse events and unplanned hospital utilization?

Grant reviewer asked:

• What is the "discharge process?"

Slide 17

Principles of the RED: Creating the Toolkit

Screen shot of a chart that shows:

• Readmission Within 6 Months
• Between Readmission Within 6 Months and Hospital Discharge:
  • Upward arrow with text box underneath stating Probabilistic Risk Assessment
• Hospital Discharge
• Upward arrow from the text box Process Mapping pointing to Hospital Discharge.
• Between Hospital Discharge and Patient Readmitted Within 3 Months:
  • Failure Mode and Effects Analysis
• Patient Readmitted Within 3 Months
• Qualitative Analysis and upward arrow.
• Root Cause Analysis and upward arrow.

Slide 18

Process Mapping-1: Ready for Discharge?

Screen shot of a flowchart presenting:

• 1st row: Hospitalization leading to:
• Ready For Discharge? (Yes or No):
  • Medical:
    • Physician Team*:
    • Other Consultants
  • Social/Behavioral:
    • Case Manager/Social Worker
    • Psychiatry
    • Substance Abuse Counselor
  • Physical:
    • Physical Therapy
    • Occupational Therapy
  • Nutrition:
    • Nutritionist
  • Preparedness:
    • Patient
    • Family
    • Supports
    • Facilities
  • B1, B2, B3 (Homefund)
    7:30-8:30—Morning Report
    8:30-10:15—Rounds
    10:15-11am—(sit down) Rounds,
    15 minutes per Team
  • B4 Team
    7:30-8:30—Morning Report
    8:30-11am—Rounds
    9:15-9:30—Case Manager Joins Rounds
  • Nursing/Case Management Rounds
    11-11:15am—Morning Meeting to Discuss Patients' Discharge Status

Note: *Physician Team includes: Sub-I, Medical Student, Intern, Junior Resident, Senior Resident and Attending Physician.

Slide 19

Process Mapping-2: Discharge Summaries

Screen shot of a flow chart showing:

• MD/Nursing Complete Discharge Paperwork leading to:
  • Page 1: MD-Clinical Resume/Discharge Summary leading to:
  • Physician Team which conducts:
    • Clinical Course
    • Labs
    • Tests
    • Medication List
    • Follow up Appointments
    • Outstanding Issues
    • Other Services
  • Page 2: Nursing-Discharge Paperwork leading to:
    • Clinical Course
    • Medication List
    • Follow up Appointments
    • Patient's Condition/Discharge
  • Page 3: Occupational, Physical and Speech Therapy, Nutrition
  • All 3 Pages leading to: Discharge Summary Completed leading to:
  • Intern Writes Order

Slide 20

Process Mapping-4: Patient Education

Screen shot of a flowchart that presents:

• Patient Preparation leading to:
  • Physician Team:
    • Intern/Unit Clerk Prints Discharge Summary and Puts in Chart.
    • Writes Prescriptions and Puts in Chart.
    • Discusses Follow up Appointments and Medications.
  • Nursing Team:
    • Gives Patient Discharge Summary.
    • Gives Prescriptions to Patient.
    • Gives Follow up Appointments.
  • Case Manager/Social Worker:
    • Facilitates any Further Details Regarding Discharge Support Services for Patient Destination.
  • All three teams lead to: Patient Leaves Hospital Bed

Slide 21

Re-engineering the Discharge

• Iterative Group Process.
• Identification of Potential Failures.
• Prioritization.

Note: The slide shows a photograph of medical staff looking through paperwork.

Slide 22

Re-engineering the Discharge-2

• Brainstorming of Alternatives.
• Re-design of Process Map.

Note: The slide shows a photograph of two men at a table looking at paperwork and a chart.

Slide 23

Principles of the Newly Re-Engineered Hospital Discharge

• Explicit delineation of roles and responsibilities.
• Patient education throughout hospitalization.
• Easy Information flow:
  • From PCP.
  • Among hospital team.
  • Back to PCP.
• Written Discharge Plan.
• All information organized and delivered to PCP.
• Waiting until discharge order is written before beginning discharge process is error-prone.
• Efficient and safe hospital discharge is significantly more challenging if discharge personnel work only 7AM to 3 PM shift.
• All patients have access to their discharge information in their language and at their literacy level.
• Those at-risk have discharge plan re-enforced after discharge.
• Discharge processes benchmarked, measured and subject to continuous quality improvement programs.

Slide 24

RED Checklist

Eleven mutually reinforcing components:

1. Medication Reconciliation.
2. Reconcile Plan with National Guidelines.
3. Follow-up Appointments.
4. Outstanding Tests and Studies.
5. Post-discharge Services.
6. Written discharge plan.
7. What to do if a problem arises.
8. Patient Education.
9. Assess patient understanding.
10. Dc summary to PCP:
    >Telephone Reinforcement.

Note: Adopted by National Quality Forum as one of 30 "Safe Practices" (SP-11)

Slide 25

Should the NQF/RED be Done at Discharge at Every Hospital?

Hypotheses

The RED will:

• Improve readiness for discharge.
• Lower adverse events.
• Lower hospital utilization.
• The intervention will be especially effective for those with limited health literacy.

Slide 26

Testing the RED Schematic

Screen show of a flowchart showing:

• Enrollment N=750
• Informed Consent
• Randomization
• RED Intervention and Usual Care
• 30 Day Outcome Data; Telephone Call; Chart Review

Slide 27

Intervention to Administer RED

• In Hospital—Discharge Advocate (DA):
  • Nurse.
  • Interact with care team—med rec and guidelines.
  • Prepare the After Hospital Discharge Plan (AHCP).
  • Teach the AHCP.
• After Discharge—Clinical Pharmacist:
  • Follow-up call @ 2-3 days.
• The DA and Pharm manual:
  • Scripts for each task.

Note: The slide shows a photograph of a woman.

Slide 28

The slide shows a sample cover of an "After Hospital Care Plan" for a discharged patient from Boston Medical Center.

Slide 29

The slide shows a sample page from the "After Hospital Care Plan" entitled "Medicines." It shows the name of the medicine, dosage, what it is for, and what time of day to take it.

Slide 30

The slide shows a continuation of the previous slide"s sample page, "Medicines."

Slide 31

The slide shows a sample page from the plan presenting "Follow-up Appointments." It gives the doctor's name, location, reason for appointment, and phone numbers.

Slide 32

The slide shows a sample calendar from the plan which highlights when the patient left the hospital, when the pharmacist will call, and when future appointments are scheduled.

Slide 33

The slide shows a sample page from the plan which gives information on "Noncardiac Chest Pain."

Slide 34

Enrollment Criteria

• Admitted to Boston Medical Center.
• ≥18 years old.
• English speaking.
• Not on precautions.
• Does not live in an institutionalized setting.
• Has telephone.
• Able to consent.
• Not previously enrolled.

Slide 35

Enrollment

Screen shot of a flowchart showing:

• Top box: Admitted to hospital service during study dates (n=5,489)
  Assessed for eligibility (n=3,873)
  Not assessed for eligibility (n=1,616)- lack of staffing
• Excluded (n=3,124)
  Did not meet inclusion criteria (n=1,049)
  Refused to participate (n=527)
  Reached maximum subjects enrolled/day (n=954)
  Subject unavailable (n=474)
  Subject previously enrolled (n=120)
• Enrollment
• Randomized (n=749)

Notes: non-English speaking (n=371), on hospital precautions (n=274), unable to consent (n=181), admitted from or planned discharge to an institutional setting (n=74), no telephone (n=71), sickle cell disease (n=38), on suicide watch with sitter (n=10), patient had privacy status (n=8), planned discharged to non-US community (n=5), transferred to different hospital service prior to enrollment (n=8), planned hospitalization (n=3), deaf or blind (n=2), other (n=4)

Slide 36

Allocation

Screen shot of a flowchart showing:

• Randomized (n=749) pointing to:
• Choice 1: Allocated to usual care (n=376)
• 30-day Outcome Assessment; Not reached (n=68)
  Could not be contacted (n=66)
  Died after index discharge (n=2)
• Choice 2: Allocated to intervention (n=373)
• Received in-hospital intervention (n=335)
  Received Pharmacist call (n=228)
• 30-day Outcome Assessment; Not reached (n=66)
  Could not be contacted (n=65)
  Died after index discharge (n=1)

Slide 37

Randomization Worked

No significant differences by group (n=749)

Income, Education Level, Literacy, Employment

Slide 38

Randomization Worked (continued)

No significant differences by group (n=749)

Prior ED visits, length of stay (LOS) (2.7), PHQ-9

Slide 39

How Successfully was the Intervention Applied?

Slide 40

How Successfully were Outcomes Collected?

Outcome Assessment:
Telephone Contact at 30 days: 82%
Chart Review at 30 days: 100%

Average Clinical Time Required:
DA: 121 minutes
PharmD: 30 minutes

Slide 41

Medication Errors (MEs)
PharmD Telephone Call
2-4 days after discharge (n=169)

MEs due to failure to take medication

Slide 42

Medication Errors (MEs)
PharmD Telephone Call
2-4 days after discharge (n=169)

MEs due to incorrect self-administration:

Slide 43

Medication Errors (MEs)
PharmD Telephone Call
2-4 days after discharge (n=169)

MEs due to system error:

Slide 44

Interventions
PharmD Telephone Call
2-4 days after discharge (n=169)

Slide 45

Results

Slide 46

AHCP Evaluation: 30 days post-discharge

The pie chart presents the results to the question, "How useful was the booklet to you?"

• Extremely: 19%
• Very: 39%
• Moderately: 21%
• A little bit: 17%
• Not at all: 4%

Slide 47

AHCP Evaluation: 30 days post-discharge (continued)

The pie chart presents the results to the question, "What was the most helpful part of the booklet?"

• Medical Provider Information: 13%
• RED Medication Schedule: 25%
• Appointment Page: 20%
• Appointment Calendar: 12%
• Diagnosis Information: 15%
• Other: 15%

Slide 48

AHCP Evaluation: 30 days post-discharge (continued)

The pie chart presents the results to the question, "How helpful was the RED medication calendar?"

• Extremely: 26%
• Quite a bit: 45%
• Moderately: 15%
• A little bit: 9%
• Not at all: 4%

Slide 49

Knowledge of Diagnosis and Making PCP visit
30 days post-discharge

Slide 50

Self-Perceived Readiness for Discharge
30 days post-discharge

The bar graph shows RED had higher numbers than Usual Care in the following areas:

• Prepared
• Understand Appointments
• Understand Meds
• Understand Dx
• Questions answered

Slide 51

Primary Outcome

Slide 52

Cumulative Hazard of Patients Experiencing a Hospital Utilization in 30d After Index Discharge

Screen shot of a line graph presenting the "Probability of Survival" for 0 to 30 days after discharge for RED and Usual Care. The results show the probability of survival declining for both RED and Usual Care as time elapses. However, at 10 days after discharge, RED's decline is not as rapid as Usual Care.

Slide 53

Cost

Slide 54

Adjusted Rate Ratio of Hospital Utilization within Subgroups

Slide 55

Conclusions from the RCT

RED:

• Successfully delivered using:
  • RED protocols.
  • AHCP.
• Improves "Readiness for Discharge".
• Decreases hospital use:
  • 32% reduction.
  • Number needed to treat (NNT) = 7.9.
• Helps high hospital utilizers:
  • 35% reduction.
• Is Cost-Effective:
  • $329/patient.
  • 38 million discharges @ $753 billion x 32% eligible = 4 billion

Slide 56

Policy Implications

The components of the RED should be provided to all patients as recommended by the National Quality Forum Safe Practice #11.

Slide 57

Major Problem: RN Time Can Health IT Help?

• Embodied Conversational Agent to Teach the AHCP:
  • Emulate face to face communication.
  • Develop therapeutic alliance:
    • Empathy.
    • Gaze, posture, gesture.
• Workstation database to automatically print AHCP and "feed" Louise.
• Connect hospital IT to workstation.
• Kiosk for patient access.

Note: The slide shows an image of a woman named "Louise".

Slide 58

RED-lit Proposed Methods November 29, 2007

Screen shot of a flowchart which presents:

• Side 1: Hospital
  • Identification of Subjects
    • Software configuration management (SCM) Printout
  • Discuss Study and Obtain Consent
    • In/exclusion
    • Consent Form
  • Baseline Data:
    • Demographics
    • SF-12
    • PHQ-9
    • REALM
  • Randomization
    • Block Randomization by Health Literacy
  • Control Group
  • Intervention Group:
    • DA + RA Complete Workstation
    • Print AHCP
    • Present ECA to Subject
  • ECA Alerts:
    • Responded to by DA, RA
• Side 2: Post-Hospital:
  • Process Outcomes: RA 7 Day Phone Call
    • AEs (Forester method)
    • Satisfaction
      • Therapeutic Alliance Inventory
  • Final Outcomes: 30 Days Electronic Record Review:
    • PCP Visits
    • ED Visits
    • ReHospitalizations
    • Costs
  • TLC leading to PCP Visit
  • Alerts
    • Responded to by DA, RA, and PharmD

Slide 59

The diagram shows how information being entered at the Discharge Management Workstation syncs with the Electronic Health Record System, PCP (via E-mail), Pre-Discharge Patient Education Workstation (Conversational Agent), and Post-Discharge Patient Education System (TLC).

Slide 60

The slide shows a photograph of a woman in a hospital bed adjusting a telemedicine screen.

Slide 61

Skills of the Agent

• Teach the AHCP
• Competency Questions.
  • We know what they know.
• Can drill down in med education.
• Maps of test sites and Community Health Centers (CHCs).
• Instructions:
  • Lovenox.
  • Glucometer.
  • Incentive spirometer.
• Concordancy Studies:
  • Race/ethnicity.
  • Gender.
  • Empathic styles.

Note: The slide shows an image of a woman.

Slide 62

Social Chat

Slide 63

Cover

Slide 64

Medications

Slide 65

Appointments

Slide 66

Diagnosis

Slide 67

Closing

Slide 68

• PI: Brian Jack, MD
• Caroline Hesko, MPH
• Irina Kushnir
• Fiana Gershengorina
• Kim Visconti, RN
• Jared Kutzin, RN, MPH
• Alison Simas, RN
• Mary Goodwin, RN
• Lynn Schipelliti, RN
• Lindsey Hollister
• Maggie Jack
• Kacie Fyrberg, RN
• Vimal Jhaveri
• Laura Pfeifer
• Juan Fernandez
• David Anthony, MD, MSc
• Tim Bickmore PhD
• Gail Burniske, PharmD
• Kevin Casey, MPH
• VK Chetty, PhD
• Allyson Correia, RN
• Larry Culpepper, MD, MPH
• Shaula Forsythe, MPH, MS
• Rob Friedman, MD
• Jeffrey Greenwald, MD
• Anna Johnson
• Anand Kartha, MD
• Christopher Manasseh, MD
• Julie O'Donnell

Slide 69

Pennsylvania Patient Safety Reporting System (PA-PSRS): Problems Reported after at discharge

Since June 2004 to December 2007, more than 800 reports have been submitted through PA-PSRS identifying problems at discharge.
30% of all reports indicated patients left the facility prior to receiving verbal and/or written discharge instructions.

The narratives below illustrate some of the issues reported through PA-PSRS:

• Patient discharged to Nursing Home. Discharge orders for 50 mg fentanyl but were written as 500 mg. The Nursing home did not catch error until patient became very drowsy. Narcan was administered.
• Patient was discharged with the wrong discharge medication list. The discharge medication list was for another patient.
• Patient admitted with diagnosis of DK A. An x-ray of left elbow was ordered. Patient was discharged to an extended care facility via ambulance before left elbow x-ray done. Orthopedic doctor notified of x-ray not being done.
• Patient was discharged to another facility with the right femoral triple lumen catheter still in place. Staff from the other facility called asking how long and how much pressure to hold on the femoral site when removing the catheter.
• Patient's daughter called this nursing unit stating the discharge instructions were unclear. The nurse discovered the medication discharge instructions were not completed. The patient had received a coronary artery stent and the booklet was still with the chart. The daughter was also unclear of the pacemaker. Patient had a 5 second pause on the cardiac monitor. The monitor strip was placed on the medical record but the physician was not notified. The patient was discharged the following morning. The patient's spouse called to report the patient passed out on the way home. As instructed, they returned to the ED and the patient was admitted. The patient had a dual chamber pacemaker inserted the next day.
• Pt resumed Coumadin post-op tonsillectomy and developed bleeding requiring admission to the hospital and return to the OR for cauterizing of bleeding site. Dr. signed standard discharge instruction sheet of surgery center stating pt. to resume medication unless otherwise instructed and did write for pt to not resume Coumadin.
• Discharge instructions for decadron tapering not clearly written. Patient stopped taking medication abruptly and required readmission.

Note: Blanco M. Discharge Planning—A Critical Juncture for Transition to Posthospital Care. Pa Patient Saf Advis 2008 Jun;5[2])