Frailty, Sleep, and Migraine Risk in Older Adults
- Okay,hear's a breakdown of the statistical analyses described in the text,focusing on what was done and how results where presented.
- * Continuous Variables: Reported as meen (standard deviation, SD) or median (interquartile range, IQR).
- * Purpose: To determine the association between frailty status and sleep quality with migraine risk.
Okay,hear’s a breakdown of the statistical analyses described in the text,focusing on what was done and how results where presented.
1. Descriptive Statistics:
* Continuous Variables: Reported as meen (standard deviation, SD) or median (interquartile range, IQR).
* Categorical Variables: Presented as n* (%). (This means the number of participants in each category, and the percentage of the total sample that falls into that category.Percentages for covariates are found in Tables S4 and S5.)
2. Primary Analysis: Cox Proportional Hazard Models
* Purpose: To determine the association between frailty status and sleep quality with migraine risk.
* Output: Hazard Ratios (HR) and 95% Confidence Intervals (CI). (HR > 1 suggests increased risk, HR < 1 suggests decreased risk).
* Assumption Check: Proportional hazard assumption was tested using the Schoenfeld residual method and found to be valid (no violations).
3. Frailty Analysis - Multiple Approaches
* Continuous Score: HR calculated for each 1-point increase in the frailty score.
* Tri-Categorical Variable: frailty status was grouped into three categories (non-frail as the reference group).
* Linear Trend: The categories were treated as continuous,using median values to test for a linear relationship between frailty and migraine risk (P-value reported).
* Individual Indicators: Associations between *each of the five frailty indicators and migraine risk were assessed,mutually adjusted for each other.
* Kaplan-Meier (KM) Curves & Log-Rank Test: Used to visualize the cumulative hazard of migraine by frailty status and to compare the curves.
4. Sleep Quality Analysis – Same approach as Frailty
The same analytical strategies used for frailty were also applied to sleep quality.
5.Cox Models with Stepwise Covariate Adjustments
Three Cox models were built, adding layers of covariates:
* Model 1: Sociodemographic factors + either sleep quality or frailty status.
* Model 2: Model 1 + Lifestyle factors.
* Model 3: Model 2 + Mental health issues and medical histories.
* Restricted Cubic Spline (RCS) Model: Applied after model 3 adjustments to assess dose-response relationships (how changes in frailty/sleep scores relate to migraine risk).
6. Joint Associations
* Reference Group: Non-frail + Healthy Sleep Quality.
* Combinations: Nine combinations of frailty status and sleep quality categories were created and their associations with migraine examined.
* stratified Analysis: Frailty-migraine association was estimated within different sleep quality categories.
* Interaction Analysis:
* Multiplicative Interaction: Tested using an interaction term in a Cox model and a likelihood test.
* Additive Interaction: Tested using Relative Excess Risk due to Interaction (RERI) and Attributable Proportion due to Interaction (AP).Additive interaction was considered important if the 95% CI for RERI and AP did not include zero. (Reference 33)
7. Robustness Checks (Secondary Analyses)
* Additional Covariates: Model 3 was further adjusted for inflammatory bowel diseases (IBD) and specific mental disorders (anxiety, depression, schizophrenia, bipolar disorder).
In essence, the researchers used a variety of statistical techniques to thoroughly investigate the relationships between frailty, sleep quality, and migraine risk, while carefully controlling for potential confounding factors and exploring possible interactions.