Not Your Mother's Administrative Data Anymore: Present on Admission (POA) Collection and Use In States (Text Version)

On September 28, 2010, Denise Love made this presentation at the 2010 Annual Conference. Select to access the PowerPoint® presentation (284 KB).

---

Slide 1

Not Your Mother's Administrative Data Anymore:
Present on Admission (POA) Collection and Use In States

Denise Love
National Association of Health Data Organizations (NANDO)
September 2010
 

Image: The NANDO logo appears on the left-hand side of this slide and all subsequent slides.

Slide 2

Overview

• POA survey results (NAHDO)
• Highlights of states' experience collecting and using POA
• Best Practices and Technical Assistance Needs
• Other data set enhancements

Slide 3

In 1999...

• NAHDO conducted an inventory of state data collection practices for the HCUP.
• Two out of 44 states collecting inpatient discharge data reported they collect POA.
• Identified by states as a priority for national standards.

Slide 4

1999 NAHDO recommendations:

• POA inclusion in the next X12N Implementation Guide.
• Education of discharge data agencies about the value of the field to outcomes studies.
• Provision of technical support to enable its collection where needed.

Slide 5

NAHDO Member Survey, 2010—POA Collection/Use

• Survey and Webinar conducted in response to member inquiries about POA practices
• Surveymonkey
• 27 / 48 respondents

Slide 6

Does your agency collect POA with inpatient data?

A table shows the following information:

Survey Responses	27 State and Private Health Data Organizations
Yes, collecting POA	21
No, not collecting POA	6

Slide 7

Does your agency release POA in its public use files?

A table shows the following information:

Survey Responses	27 State and Private Health Data Organizations
Yes	13
No	12
Non response	2

Of those releasing in public use files:

• 8 validate POA
• 5 do not validate POA
  

Slide 8

Reasons for not releasing POA in public use files:

• Data quality concerns:
  • "2009 data was not reliable."
  • "We will consider release of the 2010 data after reliability analysis."
• Data are incomplete:
  • "Some hospitals are only recording POA on Medicare patients, and critical access hospitals are not required to code POA."
• In the process of deciding

Slide 9

Have there been any issues related to the use of POA in the Public Use File?
 

• Most respondents reported no problems.
• Issues identified:
  • POA is not coded in Critical Access, LTAC & Rehab Hospitals.
  • Reporting for non Medicare patients is not part of normal billing requirements.
  • Worry about liability from the hospital perspective.
  • Some facilities do not have the capabilities to report it at this time.

Slide 10

Does your agency use POA in public reports?

A table shows the following information:

Survey Responses	27 State and Private Health Data Organizations
Yes	6
No	17
Non response	4

If yes, how is POA used?

• "We only use POA to the extent is is used in AHRQ quality indicators."
• Risk adjustment modeling only
• PSI statistics in hospital performance reports
• Feedback to hospitals
  

Slide 11

Why is POA not used in your agency reports?

• POA reporting is not complete and/or reliable
• Cannot verify accuracy of coding across providers
• When data quality is better understood, we will use POA in reports
• Will disclose POA only upon request

Slide 12

Does your agency validate the POA code?

A table shows the following information:

Survey Responses	27 State and Private Health Data Organizations
Yes	14
No	9
Non response	4

Slide 13

Briefly describe your validation/editing process:

• Provider feedback:
  • Aggregate reports of CMS HAC's given to hospitals quarterly
  • Provider statistics against benchmarks
• Limited system edits:
  • Vendor does the screens
  • Flag invalid, missing code
  • Check for exempt POA assigned to non-exempt dx codes and vice versa
  • Check for percent of "N" POAs and blanks in monthly submissions

Slide 14

State Presentation: Quality of POA Coding

• Accurate distinction between hospital-acquired complications and comorbidities that were present on admission is essential to ensure validity of estimates of predicted mortality rates.
• Quality of POA coding was measured using 15 screens applied to high risk medical, elective surgical, and obstetrical cases.
• Predictive models were derived using only data from hospitals with acceptable POA coding.

Slide 15

POA Reporting—Select AHRQ Patient Safety Indicators

Image: A bar chart shows the following rates per 1,000 discharges:

PSI	Virginia Pre-POA	National Inpatient Sample (2006)	Virginia Post-POA
Decubitus Ulcer	33.1	24.57	8.7
Post-operative pulmonary embolism	15.4	11.18	8.7
Post-operative sepsis	18.1	15.06	15.9

Slide 16

Quick "HAC" Query—California

Image: A sad "smiley" face.

A table shows the following information:

First half of 2009	POA Values	ICD-9-CM Code	Excluding Principal Diagnosis
Number of discharges			ICD-9-CM Description for Pressure Ulcer
13,429	Y	707.23 707.24	Pressure Ulcer, Stage III  Pressure Ulcer, Stage IV
398	N	707.23 707.24	Pressure Ulcer, Stage III  Pressure Ulcer, Stage IV
9	U	707.23 707.24	Pressure Ulcer, Stage III  Pressure Ulcer, Stage IV
2	W	707.23 707.24	Pressure Ulcer, Stage III  Pressure Ulcer, Stage IV
13,837	Total	707.23 707.24	Pressure Ulcer, Stage III  Pressure Ulcer, Stage IV

Slide 17

Examples of POA Data Requests

• Outcomes studies before and after surgery
• Trends of conditions associated with alcoholic liver disease
• Factors associated with admissions and readmissions
• Design a rural health system to meet the broadest level of health conditions

Image: A cartoon figure with a lightbulb going on over its head.

Slide 18

Best Practices will Evolve

• Current Best Practices:
  • Provider feedback reports
  • Release of POA in public use files
  • Use of POA in agency reports
• State resources and funding constraints will limit:
  • Provider POA training activities in states
  • Analytic workforce capacity to fully use this field

Slide 19

State Technical Assistance Requests Related to POA

• Analysis of potential or real legal issues related to public reports revealing provider comparisons
• Sharing of open-source POA edits/screens
• How other states are using POA
• Implications for previously-released quality reports

Slide 20

Conclusions

• POA collection in states for all payer data bases is growing.
• Divergence in POA Philosophy:
  • Some states are not releasing due to validity concerns.
  • Other states don't validate and are releasing.
• States are seeking guidance on minimum edit protocols:
  • This will support enforcement of data quality thresholds.

Slide 21

 

Thank You.

Thank You

• http://www.nahdo.org
• Dlove@nahdo.org