Bariatric Surgery ECR® Playbook Contents

1. Methodology

   1. Clinical Guidelines. This file contains clinical guidelines developed by HealthPartnersa. These guidelines were used for the development of the ECR.
   2. ECR Definition Summary. This document presents a brief summary of key ECR definitions (e.g. types of models developed, trigger codes, episode period, exclusions, etc.). It is intended to be a quick reference for key definitions, which are described fully in the Bariatric_Playbook_1c_ ECR Dataset Instructions.doc
   3. ECR Dataset Instructions. This document contains instructions for constructing the ECR datasets, which are used for the typical and PAC analyses. This document was formerly referred to as the “Analysis Plan”.
   4. Statistical Analysis Plan. The Statistical Analysis Plan describes the statistical analysis methods used for the typical regression modeling.
   5. All Codes. This workbook contains code definitions (using ICD-9-CM codes, HCPCS, etc.) for the episode triggers, the risk factor variables (“RF” variables), comorbidities (“M”), procedures (“P”), and PAC variables (“PAC”). It does not contain pharmacy variables, which are in the Pharmacy All Codes file (listed below). The Bariatric ECR Dataset Instructions refer to code definitions in the All Codes workbook.
   6. Pharmacy All Codes. This workbook lists the pharmacy variables and, for each, whether claims with the pharmacy variable should be deleted or should be considered typical or PAC. The Bariatric ECR Dataset Instructions refer to the Pharmacy All Codes workbook.
2. Summary Results.

   1. Data flow. This workbook presents the following summary data:
      • Decision Tree. This tree diagram shows an outline of the decisions made in selecting patients during the process of constructing episodes.
      • Complete Data Flow. This table shows patient counts, claim counts, and cost statistics for selected steps from the Bariatric ECR Dataset Instructions.
      • Data Flow Overview. This table shows patient counts and cost statistics for the relevant, typical, and PAC services. The results are broken down further by stay, professional, and pharmacy. This table is also presented in the PAC analysis below.
      • Dollar Allocation Graphs. These graphs show how total costs are distributed by the typical and PAC groups.

   2. Demographics Professional. This file shows demographic data (age, gender) for relevant Bariatric Surgery patients with professional and pharmacy claims.
   3. Demographics Stay. This file shows demographic data (age, gender) for relevant Bariatric Surgery patients with inpatient stay claims.
   4. Predictor Frequencies Professional. This document shows the frequency of each predictor (e.g. comorbidities, procedures, etc.) in the full typical sample for professional episodes which consist of professional and pharmacy claims.
3. Analysis

   1. PAC Allowance. This workbook provides information about potentially avoidable complications (PACs) and how the PAC allowance is calculated :
      • Data Flow Overview. This table shows patient counts and cost statistics for relevant, typical, and PAC services. The results are broken down further by stay, professional, and pharmacy. This table also appears in the Data Flow workbook.
      • Severity-Adjusted Price. This table shows the computation of the severity-adjusted base price of ECR for typical care. Prices for three hypothetical patients are shown. These prices are used in the Construction of the ECR (see tab entitled “PAC Allowance” in this workbook). The final regression model (from the Typical Analysis Results workbook) is used to do the calculations.
      • PAC Allow – Stay & Prof. This table shows how to calculate the PAC allowance separately for stays and for professional services, calculate the margin and, together with the severity-adjusted base price of the ECR (for professional only), helps determine the total ECR price in three hypothetical patient scenarios.
      • PAC Allow – Combined. This table shows how to calculate the PAC allowance for stays and professional services combined together into one complete episode, calculate the margin and, together with the severity-adjusted base price of the ECR, help determine the total ECR price in three hypothetical patient scenarios. The ECR price determined in a combined fashion may be different than the ECR price when determined separately and summed at the end.
      • Payer Savings Graph. This graph shows how typical and PAC dollars can be redistributed to provide incentive for providers to reduce PAC costs within Prometheus.

   2. HACs and PACs. This workbook gives information on the frequency and types of CMS-defined HACs (Hospital Acquired Conditions) and on additional Prometheus defined PACs (Potentially Avoidable Complications).
      • HACs and PACs Summary. This table shows the costs for HACs and PACs and their relation to each other.
      • Graphs – HACs and PACs. These graphs show the added burden of HACs and PACs on stay costs as well as the prevalence and costs for specific HACs and PACs. They demonstrate which PACs have the highest “n” and the biggest dollar impact within the episode.
   3. Typical Stay Analysis. This spreadsheet contains descriptive statistics of allowed amount for the typical stays. When a risk-adjustment model was constructed, the predictive capability of the model was low, and so an average stay price is displayed along with the standard deviation instead of the model results.
   4. Typical Professional Results. This spreadsheet contains the results of analyzing the typical professional and pharmacy claims for the Bariatric Surgery episodes. The data are presented on the trimmed typical sample.
      • Allowed Amount. This page shows descriptive statistics (mean, standard deviation, etc.) for the “allowed amount” which is the equivalent of reimbursable amount, being equal to the paid amount plus the patient portion of the costs in terms of copays, deductibles, coinsurance etc.
      • Age Distribution. This page shows descriptive statistics (mean, standard deviation, etc.) for age.
      • Ln(Allow) vs. Age. This page contains a graph of allowed amount on the natural log scale vs. age. The graph shows whether allowed amount increases or decreases with age, and also whether age should be modeled as a linear, quadratic or some other type of variable.
      • t tests. To determine whether there was a significant difference in allowed amount by each predictor variable, t tests were done. Descriptive statistics (count, percent, mean, standard deviation) for allowed amount by each risk factor variable are shown in addition to the t test results.
      • Final Model fit. This spreadsheet shows the adjusted R-square, number of episodes used and other model fit information for the final model.
      • Final Model. This page shows the final severity-adjusted model used to predict costs of the professional and pharmacy portion of the typical Bariatric surgery episodes. All predictors retained in the final model and their regression coefficients, standard errors, and p-values are shown.
      • ECR Price Calculator. This worksheet allows the user to compute severity-adjusted base prices for typical care for hypothetical patients. The user can select desired comorbidity, procedure, and pharmacy variables (if included in the model). The worksheet uses the Final Model to do the calculations. Instructions are provided at the bottom of the worksheet.

References

a - Evidence-Based Case Rate Development: Gastric Bypass for Treatment of Morbid Obesity. Copyrighted by HealthPartners, 2008. www.healthpartners.com.