Table 1.
Crude and stabilized, truncated IPW estimates of the (cumulative) risk differences (RDs) for GFR worsening at 2 years. Each cell describes estimates of the risk difference between two exposure groups identified by the row and column labels. SE and p stand for standard error and p-value, respectively. The 95% confidence intervals are provided in parentheses next to the point estimates.
| RD = row minus column | met+sul | sul | met | ||
|---|---|---|---|---|---|
| no therapy | crude |
−0.0524 (−0.0844; −0.0204) SE=0.0163, p=0 |
0.0016 (−0.0271;0.0302) SE=0.0146, p=0.91 |
−0.0088 (−0.0216;0.0039) SE=0.0065, p=0.17 |
|
| linear A1c |
0.0406 (−0.0096;0.0907) SE=0.0256, p=0.11 |
−0.0033 (−0.0371;0.0305) SE=0.0172, p=0.85 |
−0.0137 (−0.0335;0.006) SE=0.0101, p=0.17 |
||
| non-linear A1c |
−0.1192 (−0.1962; −0.0423) SE=0.0393, p=0 |
−0.009 (−0.0459;0.0279) SE=0.0188, p=0.63 |
−0.0087 (−0.0287;0.0114) SE=0.0102, p=0.4 |
||
| super learning |
−0.0747 (−0.1451; −0.0044) SE=0.0359, p=0.04 |
−0.016 (−0.0528;0.0208) SE=0.0188, p=0.39 |
−0.0194 (−0.0419;0.003) SE=0.0115, p=0.09 |
||
|
| |||||
| met | crude | −0.0436 (−0.0767; −0.0104) SE=0.0169, p=0.01 |
0.0104 (−0.0195;0.0404) SE=0.0153, p=0.5 |
||
| linear A1c | 0.0543 (0.0021;0.1066) SE=0.0267, p=0.04 |
0.0105 (−0.0257;0.0467) SE=0.0185, p=0.57 |
|||
| non-linear A1c | −0.1105 (−0.1885; −0.0326) SE=0.0397, p=0.01 |
−3e−04 (−0.0383;0.0377) SE=0.0194, p=0.99 |
|||
| super learning | −0.0553 (−0.1276;0.0171) SE=0.0369, p=0.13 |
0.0034 (−0.0365;0.0433) SE=0.0204, p=0.87 |
|||
|
| |||||
| sul | crude | −0.054 (−0.0959; −0.0121) SE=0.0214, p=0.01 |
|||
| linear A1c | 0.0438 (−0.0151;0.1028) SE=0.0301, p=0.15 |
||||
| non-linear A1c | −0.1102 (−0.1939; −0.0266) SE=0.0427, p=0.01 |
||||
| super learning | −0.0587 (−0.1363;0.0188) SE=0.0396, p=0.14 |
||||