Examining Potentially Avoidable Emergency Department Encounters and Hospital Admissions

On September 28, 2010, Claudia Steiner made this presentation at the 2010 Annual Conference. Select to access the PowerPoint® presentation (3.2 MB). Free PowerPoint® Viewer (Plugin Software Help).

Slide 1

Examining Potentially Avoidable Emergency Department Encounters and Hospital Admissions

Claudia Steiner MD, MPH
Agency for Healthcare Research and Quality
AHRQ Annual Meeting • September 2010

Slide 2

Assess the impact of including both IP and ED data when evaluating potentially preventable admissions and visits

Two images: a doctor/patients and an ambulance are shown.

Previous research primarily considers IP data

Friedman B., Basu J. The Rate and Cost of Hospital Readmissions for Preventable Conditions. Med Care Res Rev 2004; 61; 225.

Slide 3

Objectives

• Understand the impact of including both IP and ED data when evaluating:
  • The utilization of potentially preventable healthcare encounters.
  • The costs of potentially preventable healthcare encounters.

Slide 4

Study Design

• >Design: retrospective, observational cohort study
• Timeframe: 23-months (January 2005—November 2006)
• Four States: AZ, FL, NE, and TN:
  • Geographic and demographic variability
  • Represent 8.6 million discharges

Slide 5

Three Primary AHRQ Resources

Image: The cover of the Guide to Prevention Quality Indictors.

Prevention Quality Indictors (Asthma {pediatric, adult, elderly}, Diabetes {pediatric, adult}, CHF, Bacterial Pneumonia, and Pediatric Gastroenteritis) } Quality Measures.

Image: A green data bin labeled SID [State Inpatient Databases].

HCUP State Inpatient Databases (AZ, FL, NE, TN) } Hospital discharge databases.

Image: A red data bin labeled SEDD [State Emergency Department Databases].

HCUP State ED Databases (AZ, FL, NE, TN) } Hospital discharge databases.

Slide 6

Study Design

• HCUP SID:
  • Encounter / discharge level
  • All discharges from all community hospitals in participating States
• HCUP SEDD:
  • Encounter / visit level
  • All treat and release encounters from all community hospital emergency departments in participating States
• AHRQ PQIs:
  • Applied the standard definitions (numerators and denominators) provided by the software package available through AHRQ
  • Each condition defined using the principle diagnosis field
  • Readmissions limited to the same condition

Slide 7

Impact of using all-listed vs. first-listed diagnosis (HCUP SEDD)

Source: Agency for Healthcare Quality and Research, Healthcare Cost and Utilization Project, State Emergency Department Databases, Arizona and Nebraska, 2006-2007.

Slide 8

Study Results: Across 4 States

Slide 9

Study Results: Across 4 States

Slide 10

Study Results: Across 4 States

Slide 11

Study Results: Across 4 States

Slide 12

Study Results: Across 4 States

Slide 13

Distribution of IP and ED Events by PQI Condition

Image: A bar chart shows the Distribution of IP and ED Events by PQI Condition:

Slide 14

Which PQIs were most impacted by adding ED data?

Percent of visits that were treat-and-release ED visits versus IP admissions:

• Pediatric Gastroenteritis (83%)
• Asthma, Pediatric (82%)
• Asthma, Adult (81%)
• Asthma, Elderly (41%)
• Bacterial Pneumonia (28%)
• Diabetes, Adult (24%)
• Diabetes, Pediatric (20%)
• Congestive heart failure (11%)

Greatest ED impact on utilization:
Pediatric gastroenteritis and Non-elderly asthma.

Lowest ED impact on utilization:
CHF

Slide 15

Conclusions

• Substantial impact of ED visits on overall hospital utilization for eight potentially preventable admissions.
  • Accounting for ED visits more than doubled the number of visits (by 500K).
• Variable impact of ED visits on overall hospital costs for eight potentially preventable admissions.
  • Increased overall costs by 7% (by $243M).

Slide 16

Project Team

AHRQ [Image: AHRQ logo]:

• Claudia Steiner, MD, MPH
• Barry Friedman, PhD
• Joanna Jiang, PhD

Thomson Reuters [Image: Thomson Reuters logo]:

• Dan Whalen
• Marguerite Barrett, MS
• Minya Sheng
• Chaya Merrill, PhD

Slide 17

Considerations

Editor, Annals of Emergency Medicine (based upon almost identical comments from Reviewer1 and Reviewer2):

"The eight selected conditions are very relevant to emergency physicians... However, the reviewers had several concerns that limit our ability to publish your manuscript.

First, ED care and hospitalization are non-mutually exclusive steps in a pathway in the United States—the decisions about whether to visit the ED are up to patients while the decisions about whether to get hospitalized are up to emergency physicians and physicians in the community.

In addition, PQIs were developed exclusively for use in the inpatient setting, and are not necessarily valid for ED visits."

Slide 18

Considerations

Reviewer 1, Annals of Emergency Medicine:

Misclassification Bias:

"One means of assessing the degree of misclassification would be to calculate the proportion of the inpatient (IP) diagnoses that match the ED "reason for visit" for the IP hospitalizations that were admitted through the ED."

Slide 19

Considerations

Reviewer 1, Annals of Emergency Medicine:

"Should preventable hospital care be viewed the same as "preventable" ED care? Perhaps these entities are different and should be analyzed as distinct entities."

Slide 20

Considerations

Reviewer 2, Annals of Emergency Medicine:

"ED care is ambulatory care and not inpatient care. This paper assumes that ED patients that have diagnoses that when admitted fall within PQI measures are the same as admitted patients with these diagnoses. But EDs treat and release patients who were likely less sick, did have their hospitalization prevented by definition, and the importance of including such treat and release patients in the PQI measure is unclear."

Slide 21

Considerations

Reviewer 2, Annals of Emergency Medicine:

"The PQI was developed for inpatient hospitalization measurement, whereas alternative methods such as the Billings algorithm have been developed to assess preventable ED care."

Slide 22

Considerations

Reviewer 2, Annals of Emergency Medicine:

"The authors even note that many ED codes are symptom-based codes, so the application of PQI to ED codes raises many concerns:

a. PQI is designed for inpatient discharge codes not ambulatory codes.

b. PQI has never been validated for application to ED patients; there is no reason to believe that ED patients with treat and release conditions and/or these codes for CHF, asthma or diabetes are similar. This needs to be studied."

Slide 23

Healthcare Cost and Utilization Project (HCUP)

The largest collection of multi-level, all-payer, encounter-level, health care data.

Image: The HCUP logo is shown superimposed over photographs of medical professionals doing a variety of jobs.