mscore 26.1.2

MScore: Risk Scores Made Easy

---

MSCORE^® is a python package designed to streamline the CMS-HCC model risk score calculation process for health plans and providers participating in Medicare Advantage and value-based care. It provides a SAS®-free alternative to produce HCCs and risk scores. It streamlines your data processing for enhanced accessibility and allows you to keep your data secure in-house, so you can focus on execution of your risk adjustment strategy.

Benefits

• Implement Anywhere: MSCORE^® unlocks multiple integration opportunities including real-time scoring, embedding in EHR or enterprise data warehouse, incorporating into cloud applications, and traditional batch scoring.
• SAS^®-Free: Create HCC and risk scores without SAS^® software, reduce software license costs and the headache of recoding and maintaining the software in another programming language.
• Enhanced Output: MSCORE^® provides the numeric relative factor value for each HCC assigned to a person allowing for a deeper understanding of how each factor contributes to an individual's overall risk score.
• Reliable: Validated on over 1 million plus enrollees. MSCORE^® is rigorously tested and updated with each HCC model release.
• CMS-HCC Model Support: Supports HCC models V22 for payment years 2021-2026, V24 for payment years 2021 - 2025, and V28 for payment years 2024 - 2026, with upgrades for each release from CMS (Initial, Mid-Year, Final).
• Easy to Set Up & Maintain: Easily installed and upgraded using pip, Python's standard package manager.
• Platform Compatibility: Compatible with Windows, Mac, and Linux operating systems.
• Optimized Performance: Efficient processing with fast runtimes, ensuring quick and accurate risk score calculations.
• Improved Data Security: Securely process personal health information within your organization's IT infrastructure.

Installation

pip install mscore

Registration

Registration is required to use the mscore package. Please visit https://riskadjustmentmodel.com/registration to create user account and obtain a license key.

Trial License

After you've installed mscore, run the following command to activate your free 30-day trial. An internet connection is required to activate & run the package.

mscore -a <license key>

Subscribe

To use MSCORE^® for internal business operations beyond the 30-day trial period, a pay-as-you-go subscription and payment method is needed. Please set up your payment method in our customer portal on the billing tab. To license MSCORE^® for commercial purposes other than internal business operations, please contact us directly for alternative licensing.

For more information on pricing, please see our product page. If you do not see a price you like, then please email us so we can discuss your requirements.

Academic users are eligible for a free license of our software. Please contact us to obtain your complimentary academic license.

Basic Use

Before you can run your scores, you will need to generate an authorization token object.

import sourcedefender  #This package is required at the top
from mscore import AuthorizeLicense, MScore

auth = AuthorizeLicense(staging_key).validate()

NOTE: You must import sourcedefender at the top of every package that imports MScore.

The mscore class requires 7 arguments in order to run with 1 optional argument.

Required

• authorizer - The authorization token obtained when AuthorizeLicense is ran
• year
• version
• model
• person_data (Pandas DataFrame, csv, or parquet file)
• diag_data (Pandas DataFrame, csv, or parquet file)
• columns

NOTE: The example below assumes you already have a Pandas DataFrame generated for your person and diagnosis datasets. If you don't have this data yet and want to test mscore, download our synthetic data files here:

• Person File
• Diagnosis File

import sourcedefender
import pandas as pd
from mscore import AuthorizeLicense, MScore

# Generate DataFrames if using our test files
person_df = pd.read_csv('person.csv')
diag_df = pd.read_csv('diag.csv')

auth = AuthorizeLicense(staging_key).validate()

model = MScore(
    authorizer = auth,
    year = '2026',
    version = 'V28',
    model = 'CMS-HCC',
    person_data = person_df,
    diag_data = diag_df,
    columns = 'all-fields',
)

scores = model.score_mscore()

risk_scores = scores.risk_scores

The resulting 'scores' output is a data model object of pandas.DataFrames. You can access your risk scores or relative factors data sets by calling the respective model name from the grouped data model.

• scores.risk_scores
• scores.relative_factors

From here you can proceed to use the risk score or relative factor DataFrames for further downstream processing or utilize the Pandas built-in methods to save to a file type of your choosing.

Features

relative_factors (attribute)

MScore offers the ability to view the relative factors for each person run through MSCORE^®. To access this data, you simply access the relative_factors attribute of the mscore output. The relative factors are the individual contributions (from the variables in the model) to the risk score. MSCORE's^® relative_factors output contains these values with 1 row per person scored, and each relative factor in the output as the columns.

To leverage this feature, you would do the following:

relative_factors = scores.relative_factors

get_disease_scores (method)

For users wanting to compute the disease contribution to a person's risk score, you can do this in MScore by using the get_disease_score method of mscore. After instantiating your mscore object, you can use get_disease_score and pass the mscore output as a parameter. This will create a new dataframe containing only the portion of the risk score that comes from disease factors. Alternatively, if you would like this data appended to your MScore risk score output, you can use join_scores parameter to do this. NOTE: Disease interactions with DISABLED variables are not included in the disease score.

You can use the get_disease_score method as here:

    disease_scores = model.get_disease_scores(scores)

Or using the join_scores parameter:

    disease_scores = model.get_disease_scores(scores, join_scores=True)

Supported Data

The mscore class attributes, person_data and diag_data, will accept a Pandas.DataFrame object, CSV or Parquet filepath directly without the need to load your input data to DataFrame first. The resulting outputs will be a Pandas DataFrame object.

Supported Environments

MSCORE^® is designed to be universally compatible, ensuring seamless integration across various platforms.

• Operating Systems: Windows, macOS, Linux
• Python Versions: 3.10 or greater (both 32-bit and 64-bit architectures)

If your required environment is not listed, please contact us as [e-mail] for an alternate solution.

User Resources

For a more in-depth dive into all available model options, arguments, and reference documents, visit our User Guides.

Support

We are experts in Risk Adjustment and we're here to help. Whether it's general questions, billing, or integration support, please contact us📧. We are here to make risk scoring easy.

Stay Connected

Please subscribe to our newsletter, to receive updates on new MSCORE^® features, model releases, tutorials, and to stay up-to-date on the latest news in risk adjustment.

Legal

To review the MSCORE^® license agreement, please visit EULA webpage.