CDI specialists play vital role in capturing pay-for-performance measures

by Shannon Newell, RHIA, CCS, Steve Weichhand, and Sean Johnson

Since the implementation of the Hospital Value-Based Purchasing (HVBP) Program in 2013, CMS has adjusted the MS-DRG payment for each traditional Medicare discharge. The type and amount of the adjustment, which could be a financial penalty and/or an incentive payment, is determined by the hospital’s performance for defined quality measures, such as risk-adjusted mortality.

Since that time, the number of pay for performance (P4P) programs and quality measures has expanded. By 2017, P4P payment adjustments will impact up to 6% of traditional Medicare revenue. Why is this relevant to the coding and CDI team? Because many of the P4P measures are claims based, meaning the performance for claims-based measures is derived from diagnosis codes submitted on claims.

The CDI team are the subject matter experts on accurate and complete assignment of diagnoses, and provider documentation requirements to support code assignment. The CDI team must understand CMS P4P measures in order to improve data quality.

The types of diagnoses that can affect these measures are typically chronic conditions, many of which have been off the radar in the inpatient acute care setting, such as restless leg syndrome or loss of weight but many others directly correlate to current topics currently under CDI review. For starters, let's take a closer look at patient safety indicator 90--a composite measure of eight different PSIs rolled up  including:

• PSI 03, pressure ulcer
• PSI 06, iatrogenic pneumothorax
• PSI 07, central venous catheter-related bloodstream infections (CLABSI)
• PSI 08, postoperative hip fracture
• PSI 12, postoperative pulmonary embolism or deep venous thrombosis
• PSI 13, postoperative sepsis
• PSI 14, postoperative wound dehiscence
• PSI 15, accidental puncture or laceration

CMS provides each hospital with an annual report, referred to as a Hospital Specific Report, that provides feedback on PSI 90 performance for the HVBP Program and HACRP. Feedback provided by the HACRP Hospital Specific Report is more meaningful for the CDI team than that provided for the HVBP.

Unlike the HVBP Program, the HACRP uses the most recent version of the measures. Different measure versions affect how the PSIs are weighted in PSI 90. For example, PSI 15 is weighted at 49% in the HACRP vs. 42% in the HVBP Program.

In addition, the HACRP uses 25 diagnosis and procedure codes, while the HVBP Program uses only nine diagnoses and six procedure codes.

The time period of data included in the performance evaluation differs as well. For FY 2015, the HACRP time period is two years (07/01/2011–06/30/2013) vs. nine months (10/15/2012–06/30/2013) for the HVBP Program.

The HACRP scoring methodology requires continual improvement across all measures to avoid a financial penalty. Each year hospitals are ranked based on PSI 90 risk-adjusted performance. Hospitals with performance that falls into the worst quartile are penalized with a 1% reduction in the MS-DRG payment for each discharge the next fiscal year.

The hospital’s quality department is typically responsible for obtaining and analyzing this report to identify organizational improvement priorities within PSI 90. Given that CMS has adopted this measure for two different P4P programs, PSI 90 performance is of significant importance to most hospitals.

Get engaged in PSI 90 data quality improvement

The following key steps can position the CDI team for successful engagement:

• Meet with the quality department to learn about PSI 90 improvement priorities
• Heighten awareness of the quality department on CDI team contributions to measure performance with improved data quality
• Get a seat at the table for existing organizational improvement initiatives
• Identify coding and documentation vulnerabilities for each PSI
• Develop an action plan to strengthen documentation capture and code assignment for conditions pertinent to the PSI measure(s), including:
• Coding and documentation query processes
• Provider educational initiatives
• Documentation infrastructure refinements
• Additional performance metrics

Identify PSI coding and documentation vulnerabilities
The CDI team (and inpatient coders) must understand PSI measure structure in order to identify data quality vulnerabilities. Three key concepts are associated with the structure of PSIs.

The first is inclusions. These variables trigger a discharge to be counted in one of the measures. Inclusions consist of ICD-9-CM codes for diagnoses and/or procedures.

The second is exclusions. These variables cause a discharge triggered for inclusion in the measure to be excluded from the measure; they will not count.

Exclusions exist to improve capture of the intended population and enhance face validity of the measures with clinicians. As an example, a patient with a stage III, IV, or unstageable pressure ulcer would be included in the PSI 3 measure unless:

• The diagnosis was not present on admission
• The patient had a length of stay less than five days; it is unlikely that a patient would develop a stage III, IV, or unstageable pressure ulcer during the course of a five-day stay

Exclusions consist of ICD-9-CM diagnosis and procedure codes, admission source codes, and discharge disposition codes.

The third concept is risk adjustment. Twenty-five different comorbid categories impact PSI risk adjustment.

Some comorbid categories have a positive impact on risk adjustment, while others can negatively impact (or weaken) risk adjustment.

ICD-9-CM codes are mapped to each comorbid category. The number of ICD-9-CM codes mapped to each comorbid category ranges from a low of 1 to a high of approximately 800.

The capture of one ICD-9-CM code for each comorbid category with a positive impact on risk adjustment would optimize risk adjustment for the PSI.

The impact that a comorbid category has on the risk adjustment is PSI specific. As an example, the capture of restless leg syndrome has a positive risk adjustment impact of 10% on PSI 3 (pressure ulcers).

A review of the measure specifications (www.qualitynet.org) can then be conducted to identify inclusions, exclusions, and risk adjustment variables for each PSI. To support identification of data quality vulnerabilities, this analysis is best performed by someone on the CDI team who understands the coding classification system, coding guidelines, documentation requirements, and associated documentation improvement strategies. Consider the following examples:

For PSI 8, which measures in-house hip fractures, patients are excluded from this measure if the fracture is pathologic.

Patients are excluded from PSI 13 if secondary diagnoses reported are considered an immunocompromised state. Chronic kidney disease (CKD) stage V and malnutrition are examples.

Patients have a strengthened risk adjustment for PSI 15 if secondary diagnosis are reported for defined risk adjustment variables. The capture of obesity, CKD, and peripheral vascular disease would optimize risk adjustment for this PSI.

Editor's Note: This article originally published on JustCoding.com. Newell is a director with CCDI-DQ which provides consulting services to hospitals interested in strengthening their coding and CDI programs and a past speaker at ACDIS National Conferences. Contact her at sknewell1010@gmail.com. Weichhand and Johnson lead the provider documentation improvement service line at Falcon Consulting Group, an EHR consultancy specializing in EHR planning, implementation, optimization, and support. Contact them at Steve.weichhand@falconconsulting.com, Sean.johnson@falconconsulting.com.