Complications Methodologies

Methodology 1: Potential Inpatient Complications

Complications are defined as certain clinical conditions that occurred after patients were admitted into the facility. Those clinical conditions often cause higher mortalities, extended length of stay, and spiked treatment costs. There are 114 such clinical conditions: 14 conditions are defined by CMS as Hospital Acquired Conditions (HACs) and 100 conditions have been identified by Premier.

This methodology is applied on the following analyses:

These reports utilize the appropriate fiscal year version for Potential Inpatient Complications (PIC) based on the timeframe selected for the report:

  • Patients discharged from 10/1/19 to 9/30/2020 group to the FY20 PIC list

  • Patients discharged from 10/1/20 to 9/30/2021 group to the FY21 PIC list

  • Patients discharged from 10/1/21 to 9/30/2022 group to the FY22 PIC list

  • Patients discharged from 10/1/22 to 9/30/2023 group to the FY23 PIC list

  • Patients discharged from 10/1/23 to 9/30/2024 group to the FY24 PIC list

  • Patients discharged from 10/1/24 and forward group to the FY25 PIC list

  • Reports run for time periods that cross over multiple fiscal years (i.e., 10/1/2019, 10/1/2020, 10/1/2021, 10/1/2022, 10/1/2023, and/or 10/1/2024) will see a mix of the different PIC FY versions.

The complications in this methodology were defined similarly to the CMS Hospital Acquired Conditions (HAC) methodology where one or more ICD-10 CM code(s) are grouped together and paired with present-on-admission (POA) flags to identify if the event occurred after admission.

Premier Research Services reviewed CareScience Analytics high volume complications to develop the disease groupings. The POA (present-on-admit) flags of N (No: not present on admission) or U (Unknown: documentation insufficient) are used to delineate if the condition occurred after admission.

In order for the condition to be measured, the POA flag must be set to N or U for any one diagnosis in the condition.

  • Numerator: The secondary diagnosis code(s) for each complication creates the numerator populations

  • Denominator: The total population for the analysis.

Only complications with volume are included in the rows. This means that if there are no cases for a complication that complication will not appear on the analyses.

POA Flags in the System

The following POA flags are in the system. The POA flag on secondary diagnoses identifies a comorbidity or a complication for CareScience Analytics.

Flag

Description

Identifies a Comorbidity

Identifies a Complication

Y

Diagnosis was present at the time of the inpatient admission

Yes

No

W

Clinically undetermined.

Provider unable to clinically determine whether the condition was present at the time of the inpatient admission

Yes

No

N

Diagnosis was not present at the time of the inpatient admission

No

Yes

U

Documentation is insufficient to determine if the condition was present at the time of the inpatient admission

No

Yes

I

Unreported / Not used

Exempt from POA reporting. This code is equivalent to a blank on the UB-04®

N/A

N/A

POA Flags and Comorbidity

The Present on Admission (POA) flag determines which conditions were already present when the patient arrived at the facility, regardless of the patient’s principal diagnosis for the visit. Conditions already present when the patient arrives are considered comorbidities and impact the care administered for the principal diagnosis as well as the overall outcome.

For example, if a patient arrives at the facility with a principal diagnosis of Heart Failure (HF) and is also a diabetic, diabetes is considered a comorbidity because it impacts how care is administered to that patient. In addition, diabetics are more likely to develop certain complications due to their condition. As a result, comorbidities are very important in assessing overall patient care.

Defining Comorbidity

In order to qualify as a comorbidity, the POA flag must be set to Y or W on any one of the secondary diagnoses to ensure that the condition is present when the patient was admitted.

  • Y (Yes: Diagnosis was present at time of the inpatient admission)

  • W (Clinically undetermined: Provider unable to clinically determine whether the condition was present at the time of the inpatient admission)

For each federal fiscal year, CMS publishes a list of ICD diagnosis codes that are exempt from POA coding. Many of them are status codes, e.g. history of tobacco use. In CareScience Analytics, those ICD diagnosis codes are regarded as comorbidity.

Comorbidity Composite Scores

CareScience Analytics calculates a comorbidity composite score for the conditions that were present on admission for the following outcomes:

  • Complications
  • Cost and Charge
  • Length of Stay
  • Mortality
  • Readmission

The comorbidity composite score is one of the independent variables like age, gender, or chronic disease that contribute to the risk score calculation for each outcome at the patient level. Risk scores determine the Expected values for risk- adjusted outcomes. In fact, the Expected value column is the average of all the patient-level risk scores for an outcome. Therefore, comorbidity composite scores are an important contributing factor to the Expected values of these outcomes.

In order to determine the impact of comorbidities on risk scores calculated for outcomes, CareScience Analytics has developed a series of regression models. A patient’s secondary diagnosis codes are mapped to a Comorbidity Weight Index table, which is populated with coefficients from these regression models. The sum of the weight of all secondary diagnoses, present on admission or on the list of POA exempt, is the comorbidity composite score of the patient. The more severe a comorbid condition is, the higher the weight applied when calculating the risk score for the patient.

Potential Inpatient Complications

Note: To be a complication, the POA flag must be set to N or U on at least one of the secondary diagnoses.

The Excel spreadsheet posted here includes the full list of CMS HACs and PICs, including the associated methodology for each:

Note: When trying to identify CMS HACs, filter on the specific HAC you are looking for in column C

Complications with “(CMS)” after their name are CMS HACs.

Methodology 2: CMS Hospital Acquired Conditions (HAC)

This methodology provides a snapshot of your facility’s performance on the CMS- defined HACs that CMS deems ineligible for reimbursement if they occur after admission.

This methodology enables you to:

  • Compare costs and charges for cases where the secondary diagnosis is not present on admission versus cases where the secondary diagnosis is present on admission.
  • Review detail information for cases of facility-acquired pressure ulcers, urinary tract infections, and injuries, including specific diagnosis codes and the rate of occurrences after admission.
  • Review patient-level detail for each case after admission.

Analysis

This methodology is applied on the Facility CMS Hospital Acquired Conditions Analysis. Each row on this analysis is a CMS HAC. For information about the columns on this analysis, please review CMS Hospital Acquired Conditions.

CMS HAC Fact Sheet

The Excel spreadsheet posted here includes the full list of CMS HACs and PICs:

Note: When trying to identify CMS HACs, filter on the specific HAC you are looking for in column C

Use the Effective Date and Expiration Date filters to see specific fiscal year lists.

Patients discharged:

  • Patients discharged from 10/1/19 to 9/30/2020 group to the FY20 PIC list

  • Patients discharged from 10/1/20 to 9/30/2021 group to the FY21 PIC list

  • Patients discharged from 10/1/21 to 9/30/2022 group to the FY22 PIC list

  • Patients discharged from 10/1/22 to 9/30/2023 group to the FY23 PIC list

  • Patients discharged from 10/1/23 to 9/30/2024 group to the FY24 PIC list

  • Patients discharged from 10/1/24 and forward group to the FY25 PIC list

  • Reports run for time periods that cross over multiple fiscal years (i.e., 10/1/2019, 10/1/2020, 10/1/2021, 10/1/2022, 10/1/2023, and/or 10/1/2024) will see a mix of the different PIC FY versions.

Note: Reports run for time periods that cross over multiple fiscal years will see a mix of the different CMS HAC versions.

The CMS HACs included in the Facility CMS Hospital Acquired Conditions Analysis are based on the CMS HAC Fact Sheet, which is located in the Downloads section of the CMS.gov Web site.

CMS HAC Categories on the Rows

The main analysis has a main row for each CMS HAC category and a sub- category row for each CMS HAC in that category:

  • Hospital Acquired Infections

  • Serious Preventable Events

  • Manifestations of Poor Glycemic Control

  • Falls and Trauma

When this analysis returns, only those CMS HACs that occurred at the selected facilities for the timeframe appear on the analysis. If a CMS HAC does not appear, that means that the selected facilities did not have an occurrence of that CMS HAC during the timeframe.

A patient can qualify for more than one HAC, however, a patient is counted only once for a particular HAC even though a patient may have multiple diagnosis codes to qualify for that HAC.