MOSI

Metabolic & Obesity Staging Index

Decision support and prospective validation for bariatric surgery.

Visit the live tool View on GitHub Contact about this project

The Problem

What is broken clinically?

Bariatric surgery decisions are made with substantial inter-surgeon variation. There is no validated, quantitative staging system that integrates BMI, comorbidity severity, and predicted weight-loss outcomes into a single actionable score — leaving patients under-staged or over-treated based on institutional custom rather than evidence.

Clinical Need

Why this matters in surgery

Over 250,000 bariatric procedures are performed annually in North America. Procedure selection (sleeve gastrectomy vs. Roux-en-Y gastric bypass vs. revisional surgery) significantly impacts total weight loss and comorbidity resolution. A validated decision-support tool that stratifies patients objectively can reduce variation, improve outcomes, and standardize audit-ready documentation.

Data Sources

Where the data comes from

Derivation cohort: 3,097 patients who underwent primary bariatric surgery at Mayo Clinic, Rochester, MN (2010-2022). Prospective validation cohort: multi-institutional, external, currently enrolling. Data include: preoperative BMI, HbA1c, blood pressure, comorbidity index, procedure type, and 12-month postoperative total weight loss (TWL%).

Methods

Technical approach

MOSI assigns patients a composite score across four domains: (1) BMI stage (WHO classification adjusted for metabolic risk), (2) comorbidity burden (Charlson score weighted for metabolic-specific conditions), (3) predicted TWL tier (derived from a validated regression model), and (4) surgical complexity adjustment. The algorithm runs on a web-based calculator with exportable clinical reports. The staging system achieved 100% algorithm accuracy on the derivation cohort and is currently undergoing prospective external validation.

Validation Plan

How the claims will be tested

A prospective, multi-site external validation study is ongoing across three Mayo Clinic campuses (Rochester, Phoenix, Jacksonville). Primary endpoint: concordance between MOSI-recommended procedure and surgeon decision. Secondary endpoints: 12-month TWL%, comorbidity resolution rates, and readmission within 30 days. Target n = 800 patients. IRB approval obtained.

Current Status

What stage we're at

MOSI is in prospective external validation (multi-site). The web tool is live and accessible to clinicians. The derivation paper was published in Surgery for Obesity and Related Diseases (SOARD) in 2026. A validation-cohort manuscript is in preparation.

Model Card

Intended use, readiness, and limitations

A-STAR model card

MOSI

Metabolic & Obesity Staging Index

Intended use: Pre-operative decision support for bariatric surgery candidate evaluation. Advisory only; clinical judgment remains with the treating surgeon.
Clinical phase: Risk stratification, planning, and outcomes validation
Input data: BMI and metabolic risk profile
Comorbidity burden
Prior bariatric procedure history
Procedure and outcomes variables used in validation workflows
Output: MOSI stage
Procedure category recommendation for review
Predicted total weight-loss tier
Audit-ready report
Model/pipeline: Rule-guided staging and decision-support workflow with prospective external validation in progress.
Validation status: Derivation and publication complete; prospective external multi-site validation is in progress.
Deployment readiness: Validation phase. Available as a clinical decision-support aid, not for autonomous use.
Limitations: Derived from Mayo Clinic data and not yet characterized across all populations or revisional surgery contexts.