Implementing QI Validation Tools for Coding and Clinical Quality Improvement in Academic Medical Centers

Select to access the PowerPoint® presentation (1.6 MB).

Slide 0

Implementing QI Validation Tools for Coding and Clinical Quality Improvement in Academic Medical Centers

Leslie Prellwitz, MBA, CCS, CCS-P
Director, Performance Improvement Analytics
September 11, 2012

Slide 1

Objectives

• Recognize the importance of validation efforts to healthcare providers (Academic Medical Centers in particular):
  • National Landscape.
  • Describe the Annual UHC Performance Ranking and the use of the PSI's.
• Identify the tools and techniques used in QI validation:
  • Chart Review.
  • Case Control Study.
  • Case Scenarios.
• Assess the role validation serves in successfully implementing improvement activities
  • Improving Practice and Outcomes: Success Stories.
  • Why other providers should also be interested in QI validation.

Slide 2

Why is QI Validation Important to Academic Medical Centers?

Proliferation of QI indicators over time

Image: A series of text boxes shows the following timeline:

• HCUP Quality Indicators, 1994.
• AHRQ Quality Indicators, 1999:
  • Prevention Quality Indicators (PQI), Nov. 2000.
  • Inpatient Quality Indicators (IQI), May 2002.
  • Patient Safety Indicators (PSI), Mar. 2003.
  • Pediatric Quality Indicators (PQI), Apr. 2006.

Slide 3

Why is QI Validation Important to Academic Medical Centers?

Visibility in an array of (increasingly public) venues

RAND Analysis of environmental scan results, 2007
"Evaluation of the Use of AHRQ and other Quality Indicators", AHRQ Publication No. 08-M012-EF
December 2007

Slide 4

Why is QI Validation Important to Academic Medical Centers

Reimbursement

• Similar metrics in DRG-based reimbursement now.
• QI indicators to be incorporated into Value Based Purchasing starting in Federal Fiscal Year 2015.

Improvement

• More complex patients more prone to certain conditions.
• To guide improvement, AMCs need to be confident that the QIs are identifying:
  • The appropriate target populations.
  • The appropriate risk factors.

Slide 5

Comparative Ranking of UHC Members has driven a focus on all aspects of improvement

Since 2005, on an annual basis, UHC has ranked performance of all of its principal members on selected dimensions of quality.

Y = performance levels provided but no included as a component in the overall ranking.

Slide 6

UHC Quality & Accountability Study: Safety Domain, 2011

Slide 7

Improving Patient Care is a Team Sport

Image: A chart depicts the following process:

Clinicians:

• Patient Characteristics & Clinical Profile →
• Clinical Care: Diagnosis, Intervention, Prevention →
• Documentation of Care.

plus

Coders

• Principal Diagnosis.
• Secondary Diagnosis.
• Principal Procedures.
• Secondary Procedures →

ICD-9 Codes Auto-Mapped to MS-DRG's →

MS-DRG Assignment of Severity-Level Profiles →

• Risk-adjusted Profiles.
• Public Reporting and Ranking.
• Quality Measurement.

It takes effort from all parties to improve Quality and Safety.

Slide 8

UHC's Approach to Improvement—Connections to QI validation

Benchmark and Share Best Practices for Clinical Care, Documentation and Coding

Clinical Care—Appropriate Population and Risk Factors

• Understand measures of performance (i.e. numerator and denominator).
• Understand the evidence based practice associated with the treatment of a condition or prophylaxis.
• Evaluate actual patient care provided in relation to evidence based practice.
• Determine the factors that influence the outcome of interest.

Documentation—Accurate reflection?

• Timely documentation to define the condition as co-morbid vs. complication.
• Appropriate terminology to represent the severity of illness of the patient.
• Terms that describe the severity of the condition.
• Clarification regarding conditions that are ruled-out.

Coding—Correct, Consistent translation

• Consistent interpretation of the condition.
• Correct selection of codes to represent patient condition and care.

Slide 9

QI Validation—Dimensions and Tools

Image: Three boxes containing the following text are connected by arrows in circle, indicating an ongoing process:

• Clinical Care: Case Control, Chart Review.
• Documentation: Chart Review.
• Coding: Case Scenarios.

Slide 10

UHC VTE Benchmarking Project

Project Goals:

• Identify opportunities to improve prophylaxis methods to consistently meet evidence-based practice guidelines.
• Demonstrate that some patients receiving evidence-based prophylaxis still developed VTEs.
• Learn which patient characteristics or other criteria are most commonly present in VTE cases.

Patient Population of Focus: Total Knee Replacement (TKR)

CDB Analysis

In addition to review of impact of prophylaxis methods and guideline compliance, also reviewed accuracy of case identification in PSI 12, Post-operative DVT/PE.

Slide 11

UHC VTE Benchmarking Project - Coding and documentation of VTEs for TKR: Case Control and Chart Review

Work with team from UC Davis Health System

Applied PSI 12 (post-operative DVT/PE) Version 4.1 to eligible cases with POA flags.

• Additional DVT/PE cases were captured by applying the same ICD-9-CM definition to POA diagnoses on records within 90 days of the TKR discharge.

Flagged cases (n=126) and non-flagged controls (n=463) were audited at each participating hospital.

• When there was a discrepancy between PSI-flagged status and the abstractor's determination, a detailed review was conducted to identify reasons for the discrepancy.

Data Collection Tool—element categories captured:

• Administrative*.
• Demographics*.
• Surgery & Screening.
• Prophylaxis.
• Ambulation.
• Outcomes.

* = data linked with UHC's Clinical DataBase/Resource Manager (CDB/RM) for validation.

Slide 12

VTE Benchmarking Project—Data Collection Tool: Administrative

Validation begins with administrative data.

QI measure patient selection: denominator.

CDB provides an external reflection: Is it accurate?

Image: The first page of the data collection tool form, titled "A. Adminstrative," is shown.

Slide 13

VTE Benchmarking Project—Data Collection Tool: Demographics

Present on Admission: Key for Risk Factor identification.

CDB provides enhanced dx depth; have all factors been captured?

Image: The second page of the data collection tool form, titled "B. Demographics," is shown.

Slide 14

VTE Benchmarking Project—Data Collection Tool: Surgery & Screening

Image: The third page of the data collection tool form, titled "C. Surgery and Screening" is shown.

Slide 15

VTE Benchmarking Project—Data Collection Tool: Prophylaxis

Prophylaxis: Connect pharmacologic option with outcomes.

Continuation at Discharge—Implications for readmissions?

Image: The fourth page of the data collection tool form, titled "D. Prophylaxis" is shown.

Slide 16

VTE Benchmarking Project—Data Collection Tool: Prophylaxis (non-pharmacologic)

Image: The fifth page of the data collection tool form, titled "E. Prophylaxis Continued" is shown.

Slide 17

VTE Benchmarking Project—Data Collection Tool: Ambulation

Condition-specific elements of care also captured.

Image: The sixth page of the data collection tool form, titled "F. Ambulation" is shown.

Slide 18

VTE Benchmarking Project—Data Collection Tool: Outcomes

Validation concludes with administrative data.

QI measure patient selection: numerator.

Image: The seventh page of the data collection tool form, titled "G. Outcomes" is shown.

Slide 19

UHC VTE Benchmarking Project: Results

AHRQ PSI 12 can be used with high accuracy to flag post-operative DVT/PE cases and to monitor trends over time.

Slide 20

UHC VTE Benchmarking Project: Practice Improvement Opportunities

Routinely monitor and analyze your hospital's DVT/PE rates against internal and external benchmarks.

Provide patients with guideline-directed prophylaxis and focus on the timing of the first post-operative dose.

Promote early ambulation (within 24 hours after surgery) to guard against DVT/PE.

Reduce practice variation and standardize guidelines within the organization and across providers.

• Integrate standardization into the order sets.

Identify and empower a physician champion who can promote best practices and provide education and feedback to all stakeholders.

Establish and support an effective review forum for VTE events.

Slide 21

Postoperative Respiratory Complications Benchmarking Project

Postoperative Respiratory Complications Documentation and Coding Survey

• Follow-up to the Postoperative Respiratory Failure 2007 Benchmarking Project.
• Survey purpose: to understand the variation in coding postoperative respiratory failure (PSI 11).
• Case scenario and multiple-choice questions.
• Requested that 3 coders from each organization respond.
• Sent to UHC full members.

CDB Analysis

• Purpose was to examine preferences for the use of PSI 11 codes.

Slide 22

Definition PSI 11: Postoperative Respiratory Failure (version 3.1 current at time of study)

Numerator Codes:
Respiratory failure ICD-9-CM secondary diagnosis code

• 518.81: Diagnosis of acute respiratory failure.
• 518.84: Diagnosis of acute and chronic respiratory failure.

OR

Intubation or ventilation ICD-9-CM procedure code with appropriate timing after a qualifying surgical procedure

• 96.04: Endotracheal tube insertion procedure takes place 1 or more days after a major operating room procedure—i.e., reintubation.
• 96.70: Continuous ventilation (unspecified duration) or 96.71: Continuous ventilation (less than 96 hours) identified 2 or more days after a major operating room procedure.
• 96.72: Continuous ventilation (for 96 hours or more) identified on or any time after the day of a major operating room procedure.

Denominator:

• Adults undergoing elective operations.
• Excludes:
  • Diagnoses of respiratory failure on admission.
  • Tracheostomy before or during the main procedure.
  • Patients with primary respiratory, circulatory, or pregnancy-related process or a neuromuscular disorder.

Slide 23

Predictive Value of PSI 11 Data Collection Tool for Chart Review

• 609 flagged cases from 18 UHC-affiliated centers.
• Medical records reviewed.

Data Collection Form—Categories covered

• Administrative Data*.
• Demographics/Patient Factors*.
• Surgical Procedures (first, additional)*.
• Invasive intubation.
• Additional invasive intubations or ventilator support episodes for chronic trach patients.
• Outcome.

* = data linked with UHC/s Clinical DataBase/Resource Manager (CDB/RM) for validation.

Slide 24

Predictive Value of PSI 11 Benchmarking Project Experience—Chart Review

• 90% of cases had accurate coding.
• Hospitalization not elective in 5%.
• Inaccurate diagnosis, procedure codes in 3%.
• 83% of cases represented true PRF.

* p<0.05

Romano et al., Health Serv Res, 2009

Slide 25

Predictive Value of PSI 11 Coding Experience—Case Scenarios

17 organizations participated

• 3 coders per organization were requested to respond.

Total of 56 coders responded.

Coding experience

• Average = 13.5 years.
• Median = 11.5 years.
• Range = 1.5 to 30 years.
• Two Case Scenarios concerning postoperative respiratory failure presented for interpretation.

Image: A pie chart shows the following data:

• 0-9 Years - 36% (20).
• 10-19 Years - 36% (20).
• 20-29 Years - 25% (14).
• 30 + Years - 3% (2).

Slide 26

Coding Experience Summary and Review: What Do the Survey Results Tell Us?

Inter- and intra-organization variation in coding postoperative respiratory failure.

• Inter-organizational variation was apparent based on the number of different ICD-9-CM diagnosis codes identified by survey respondents.
• Intra-organizational variation was identified by differences in the responses to the case scenarios by coders from the same organizations.
• Variation in coding was also identified through the responses to 2 statements regarding documentation and coding of postoperative respiratory failure.
  • For each statement, about half of the respondents agreed or strongly agreed that they would use the identified code and about one-third disagreed or strongly disagreed with using the identified code.

Slide 27

Hospital Successes using QIs

Slide 28

Improving Outcomes: Success Stories

Pressure Ulcer Performance

Image: A line graph shows the numbers of cases of pressure ulcers declining from the first quarter of 2010 to the final quarter of 2011.

Slide 29

Success Stories: Pressure Ulcer Reduction

Goal: Commitment to top decile performance

Background: In 2010, UAB Hospital sought to streamline the commitment to quality through the appointment of a new Chief Quality and Safety Officer (CQSO) as well as a reorganization of the Nursing Quality Council (NQC). Both changes align with the Health Systems clearly articulated goal: to provide exceptionally safe and high quality health care as measured by national quality indicators. NEW STRUCTURE = NEW APPROACH TO QUALITY MEASUREMENT

Interventions:

1. education and increased awareness by all disciplines of causes and preventative measures.
2. creation of unit based quality dashboards.
3. implementation of monthly quality variance meetings, where all HACs are discussed and action plans determined.
4. hospital wide monthly trending to identify targeted opportunities.
5. identification of unit based staff nurse pressure ulcer experts.

Results: The number of hospital acquired pressure ulcers decreased from 33 in first quarter 2010 to 8 in the fourth quarter 2011.

Slide 30

UHC's Work on PSI and HAC Coding

Slide 31

UHC's Work on PSI and HAC Coding

Image: The following timeline is set within an arrow pointing from left to right:

• 2011: Coding Post-Operative Respiratory Failure.
• 2011: Accidental Punctures and Lacerations Networking Collaborative.
• 2012: AHRQ Quality Indicator Documentation and Coding Toolkit.
• 2011-16: Battelle/UHC Quality Metrics Project for AHRQ.
• 2012: Develop PSI and HAC Recommendations Influence National Agenda.

Slide 32

Consensus Recommendations Development Project: Accurate Documentation and Coding

• Develop consensus recommendations for documentation/reporting PSIs and HACs:
  • Compliant with national definitions and existing guidelines.
• Provide consistent interpretation in areas of uncertainty.
• Promote standardized reporting across members.
• Enhance the accuracy and comparability of data.

Image: A child's alphabet blocks spell out G-O-A-L-S.

Slide 33

Patient Safety Expert Panel

Accidental puncture or laceration
Postoperative respiratory failure
Iatrogenic pneumothorax
Foreign body left during procedure

• The Cleveland Clinic Foundation.
• NewYork-Presbyterian Hospital.
• NYU Langone Medical Center.
• UC Davis Medical Center.
• University of Kentucky Hospital.
• University of Michigan Hospitals & Health Centers.
• Vanderbilt University Medical Center.
• Wexner Medical Center at The Ohio State University.
• University of Washington.
• Emory University Hospital.

Slide 34

Obstetric Expert Panel

OB trauma—with instrument
OB trauma—without instrument
Birth trauma—injury to neonate

• Beaumont Hospital, Royal Oak.
• Froedtert & The Medical College of Wisconsin.
• Massachusetts General Hospital.
• Medical University of South Carolina.
• The Nebraska Medical Center.
• The University of Kansas Hospital Authority.
• UC Davis Medical Center.
• University of North Carolina Hospitals.
• University Hospitals Case Medical Center.
• University of Washington Medical Center.
• UT Southwestern Medical Center University Hospitals—Zale Lipshy and St. Paul.

Slide 35

Why other providers should be interested in QI Validation

Metrics will eventually affect all provider types.

• Long term care, ambulatory surgery, others.
• Value Based Purchasing extension into episodes of care; improvement will move into an extended collaborative effort across these care settings.
• Pace of usage will only increase over time as budget constraints increase.
• ICD-10 provides an opportunity to reset the slate.

Where do you want to be? Ahead of the curve and informing the decision, or behind the curve and accepting the result?

Slide 36

Questions?