API-Powered Scoring

Response Pattern Scoring

The Assessment Center API uses “response pattern scoring” to automatically calculate T-scores. This item response theory (IRT)-based method individually scores each item’s response as part of the overall computation of the T-score. Response pattern scoring accounts for differences in symptom severity and functional status that are measured by different items. For example, selecting “Almost always”, a raw response score of 4, to a question about running 10 miles would result in a higher T-score than selecting ”Almost always” to a question about walking 1 block since running 10 miles represents greater physical function than walking a block. This is the gold standard scoring approach because it:

• Produces more precise scores (less error).
• Accommodates missing data (only items that have been answered contribute to the T-score for a given individual).
• Increases power to detect change in scores over time or between groups.
• Allows for computer adaptive tests (CATs) to further reduce respondent burden with fewer questions.
• Avoids error-prone manual score calculation (software automatically produces T-scores).

Learn more about Item Response Theory and Computer Adaptive Tests.

When to Use API-Powered Scoring

API-powered scoring is the preferred scoring approach when feasible. This scoring method is especially valuable when you:

• Need immediate scoring (scores automatically calculated).
• Want participants who skip items to still receive a score.
• Want to administer CATs.
• Want to compare scores for small subgroups of respondents (fewer than 75 people).
• Want to interpret scores at the individual level.

API-Powered Platforms

If you are administering HealthMeasures with the Epic PROMIS application or another API-powered platform, such as REDCap, the platform will automatically calculate scores for you. Confirm with the platform that you are administering API-powered HealthMeasures; for example, REDCap indicates API-powered measures with the labels “auto-scoring” and “adaptive instrument”.

Learn more about options and licensing with the Epic PROMIS application and API-powered Platforms.

API Integration

If you are using your own digital administration platform, you can integrate the Assessment Center API to administer a comprehensive library of HealthMeasures, including computer adaptive tests (CATs) and custom forms. Scores are automatically generated. Additionally, score interpretation cut-points (e.g., mild, moderate, within normal limits) and population norms are available for many measures.

Learn more about licensing and integrating the Assessment Center API.

Direct API Scoring

If you have already administered a short form or custom form, you can score the data from within a statistical software package by calling the Assessment Center API as a web service. The API is hosted on your own server or on the statistician’s computer which allows you to keep data behind your own firewall. To implement the API for this simpler scoring-only approach:

1. License the Assessment Center API.
2. Install the API on your server or computer.
3. Identify the documentation for calling a web service for your desired statistical package:
   • SAS uses Proc SOAP to call a web service hosted within the SAS environment.
   • R can use the httr library to call a web service.
   • SPSS Modeler can call a web service via CLEF.
   • Stata and other software packages have similar capabilities.
4. Write code to call the relevant API endpoints for the items that have been administered. Translated HealthMeasures can be scored using the English versions of the measures implemented in the API.

Learn more about licensing the Assessment Center API. Contact api@assessmentcenter.net for scoring support.