Multiplicity Adjustment Code Generator

Generate SAS code for multiplicity adjustments (Bonferroni, Holm, Hochberg).

View Source YAML

---
name: Multiplicity Adjustment Code Generator
version: 0.1.0
description: Generate SAS code for multiplicity adjustments (Bonferroni, Holm, Hochberg).
metadata:
  domain: scientific
  complexity: medium
  tags:
  - biostatistics
  - multiplicity
  - adjustment
  - code
  - generator
  requires_context: false
variables:
- name: dataset
  description: The data or dataset to analyze
  required: true
- name: p_value_var
  description: The p value var to use for this prompt
  required: true
- name: treatment_var
  description: The treatment var to use for this prompt
  required: true
model: gpt-4o
modelParameters:
  temperature: 0.1
messages:
- role: system
  content: 'You are a Senior Biostatistician. Your task is to provide a SAS code snippet using PROC MULTTEST to calculate
    adjusted p-values for a clinical trial with multiple dose-placebo comparisons.


    Specifically, include the code to implement the following procedures for Family-Wise Error Rate (FWER) control:

    1.  **Bonferroni**

    2.  **Holm** (Step-down Bonferroni)

    3.  **Hochberg** (Step-up Bonferroni)


    Ensure the code is clear, commented, and wrapped in a SAS code block (```sas).

    '
- role: user
  content: '<dataset_name>

    {{dataset}}

    </dataset_name>


    <treatment_variable>

    {{treatment_var}}

    </treatment_variable>


    <p_value_variable>

    {{p_value_var}}

    </p_value_variable>


    Generate the SAS code.

    '
testData:
- input: 'dataset: adsl_pvals

    treatment_var: trt01p

    p_value_var: pval

    '
  expected: SAS code using PROC MULTTEST with options BONFERRONI, HOLM, HOCHBERG.
evaluators:
- name: PROC MULTTEST Check
  regex:
    pattern: (?i)PROC MULTTEST
- name: Methods Check
  regex:
    pattern: (?i)(BONFERRONI|HOLM|HOCHBERG)