🚨 Respiratory Stratified Analysis
Dataset Summary¶
| Group | N | Percentage |
|---|---|---|
| With O2 Support | 60 | 26.3% |
| Without O2 Support | 168 | 73.7% |
| Total | 228 | 100.0% |
Analysis Variables¶
Outcome: time_to_fof\n\nPredictor: time_to_first\n\nStratification: o2_device_at_36_weeks\n\nConfounders: gestational_age_weeks, birth_weight_grams, medicalcomplexity
Interaction Analysis¶
Respiratory Support Interaction Testing\n\nTests whether feeding relationships differ by respiratory support status.\n\nUnadjusted model:\n- Interaction coefficient: -0.094\n- Interaction p-value: 0.3523\n- Model R²: 0.070\n\nAdjusted model (including confounders):\n- Interaction coefficient: -0.083\n- Interaction p-value: 0.4074\n- Model R²: 0.114\n\nModel diagnostics warning: High multicollinearity detected. Residuals may not be normally distributed. \n\nStatistical conclusion: No significant interaction (p = 0.4074)
With O2 Support Group Analysis¶
Sample size: 60 infants\n\nAnalysis includes: Optimal timing identification and slope comparison
With O2 Support: Optimization Analysis¶
Unadjusted quadratic model:\n- Sample size: 60\n- R²: 0.045\n- Linear coefficient: 0.595 (p = 0.1057)\n- Quadratic coefficient: -0.005 (p = 0.1164)\n\nNo optimal found in data range (unadjusted)\n
With O2 Support: Adjusted Model¶
\nAdjusted quadratic model:\n- Sample size: 60\n- Adjusted R²: 0.056\n- Adjusted quadratic coefficient: -0.002 (p = 0.5351)\n\nNo optimal found in data range (adjusted)
With O2 Support: Slope Comparison¶
Slope comparison using median (no optimal found) (57.6 days):\n\nInteraction p-value: 0.1564\n\n
With O2 Support: Slope Results¶
| Group | N | Slope | SE | p-value |
|---|---|---|---|---|
| < 57.6 days | 30 | 0.13459 | 0.183585 | 0.469579 |
| ≥ 57.6 days | 30 | -0.253575 | 0.197144 | 0.208893 |
Without O2 Support Group Analysis¶
Sample size: 168 infants\n\nAnalysis includes: Optimal timing identification and slope comparison
Without O2 Support: Optimization Analysis¶
Unadjusted quadratic model:\n- Sample size: 168\n- R²: 0.042\n- Linear coefficient: -0.104 (p = 0.5696)\n- Quadratic coefficient: 0.003 (p = 0.1861)\n\nUnadjusted optimal: 20.7 days\n
Without O2 Support: Adjusted Model¶
\nAdjusted quadratic model:\n- Sample size: 168\n- Adjusted R²: 0.053\n- Adjusted quadratic coefficient: 0.004 (p = 0.0633)\n\nAdjusted optimal: 56.9 days
Without O2 Support: Optimization Plot¶
Quadratic model fit with optimal point marked.
Without O2 Support: Slope Comparison¶
Slope comparison using adjusted optimal point (56.9 days):\n\nInteraction p-value: 0.9708\n\n
Without O2 Support: Slope Results¶
| Group | N | Slope | SE | p-value |
|---|---|---|---|---|
| < 56.9 days | 150 | 0.0979199 | 0.0825746 | 0.237588 |
| ≥ 56.9 days | 18 | 0.090033 | 0.217715 | 0.684703 |
Summary: Answers to PI's Key Questions¶
Question 1: Do infants with O2 support have different feeding relationships?\nNO - No significant interaction\n\nQuestion 2: Do they show different statistical patterns?\nYES - One group has optimal timing, the other does not\n\nQuestion 3: Is optimal timing different between groups?\nYES - With O2: No optimal found, Without O2: 56.9 days