Comparing Results Between Survey Analysis Software and the National Cancer Institute Joinpoint Regression Software...

Area of Science:

• Biostatistics
• Epidemiology
• Health Data Analysis

Background:

• National Center for Health Statistics (NCHS) guidelines recommend specific software for trend analysis of survey data.
• Joinpoint Regression Software and survey analysis software are commonly used for trend analysis.
• Apparent inconsistencies can arise between these software packages when analyzing survey data.

Purpose of the Study:

• To provide guidance on handling and explaining inconsistencies between Joinpoint software and survey analysis software.
• To supplement existing NCHS guidelines for analyzing trends in survey data.
• To assist analysts in making informed decisions for their data analysis.

Main Methods:

• Identified scenarios where inconsistencies between Joinpoint and survey analysis software may occur.
• Discussed plausible explanations for observed differences using text, examples, and FAQs.
• Focused on providing recommendations, cautions, and additional information rather than direct solutions.

Main Results:

• Inconsistencies frequently arise when piecewise regression models from Joinpoint software are estimated using survey analysis software.
• Successive slopes may not be statistically significant, leading to the removal of joinpoints in the final model.
• Potential explanations for these discrepancies are categorized into five main areas.

Conclusions:

• Understanding the reasons for software discrepancies is crucial for accurate trend analysis.
• Analysts should be aware of potential pitfalls when using Joinpoint and survey analysis software together.
• This report serves as a valuable resource for navigating complex trend analysis with NCHS survey data.