Skip to main content
Loading navigation
Back to challenges

MCED Pre-test Counseling and Risk Stratification with AI SDK

Rising consumer interest in multi-cancer early detection testing demands structured pre-test risk assessment and patient guidance. Develop a web-based clinical decision agent using Vercel AI SDK and Moonshot AI to evaluate patient risk criteria and generate tailored clinical summaries. Achieve 100% compliance with standard NCCN screening guidelines.

Data ScienceHosted by Vera
Sign in free to start

What you are building

Develop an interactive MCED risk stratification agent using Vercel AI SDK and Moonshot AI for patient pre-test counseling.

How work is evaluated

Evaluation

Evaluates guideline compliance and clinical accuracy of generated MCED pre-test counseling summaries.

Datasets

Public datasets, plus the private datasets you have access to.

How submissions are scored

Dimensions
2 scoring checks
Binary
2 pass or fail dimensions
Ordinal
0 scaled dimensions
Dimension 1

guideline_compliance_check

Requires compliance score of 1.0 against NCCN rules.

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Dimension 2

guideline_compliance

Adherence to established oncology risk stratification standards • target: 1 • range: 0-1

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Learning goals

  • Implement interactive AI streaming interfaces using Vercel AI SDK

  • Utilize Moonshot AI models for clear, empathetic clinical text generation

  • Implement rule-based NCCN guideline validation checks during multi-turn interactions

  • Format structured risk assessments for integration into Electronic Health Record portals

Resources

Dataset notes

Synthetic dataset of 100 patient clinical profiles covering diverse age ranges, risk profiles, and cancer screening histories.

Challenge setup

Evaluates guideline compliance and clinical accuracy of generated MCED pre-test counseling summaries.

Challenge input

JSON containing patient age, family history, smoking history, and prior cancer diagnoses.

AI SDK

Required framework for building streaming user-facing AI applications.

Moonshot AI

Specified language model engine for nuanced patient communication.

Evaluated output

JSON object with recommended_for_mced boolean, guideline_compliance_score, and patient_summary text.

Evaluation checks
  • Requires compliance score of 1.0 against NCCN rules.
  • Adherence to established oncology risk stratification standards • target: 1 • range: 0-1
Configured evaluation evidence
  • Benchmark: NCCN Guideline Adherence Benchmark
  • Guideline Compliance target: 1
  • 1 public reference case
Configured runtime
  • JavaScript execution harness
  • javascript sandbox (unavailable on Versalist)

Shows the first task, up to three tools, and up to three evidence items.

View technical recipe

Configured tools

Action Space
  • Moonshot AI · Required
  • Guidance · Optional
  • Vercel AI SDK · Optional
Orchestration
  • Vercel AI SDK · Optional

Evaluation contract

  • guideline_compliance_check · Weight 1
  • guideline_compliance · Weight 1

Metrics

  • guideline_compliance · Target 1

Benchmarks

  • NCCN Guideline Adherence Benchmark
  • biotech

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: JSON containing patient age, family history, smoking history, and prior cancer diagnoses.

    Output: JSON object with recommended_for_mced boolean, guideline_compliance_score, and patient_summary text.

    Execution harness: JavaScript

Evaluation tools

  • AI SDK · TypeScript Streaming Framework
  • Moonshot AI · LLM Provider Engine

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about MCED Pre-test Counseling and Risk Stratification with AI SDK