Table 5 Tanzania—Two-segmented Poisson GEE Model Log(Uptake) parameter estimates (AR(1) working correlation) using all available 26 datapoints (9 pre-, 17 post) (excluding intervention roll-out phase)
Control Arm | Intervention Arm | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
Parameter description | Parameter estimate (Log Uptake) (αi) | SE | 95%CI | p-value | Parameter description | Parameter estimate (Log Uptake) (αi) | SE | 95%CI | p-value | ||
LL | UL | LL | UL | ||||||||
Pre-Intervention phase (Pre-Step 1) | Pre-Intervention phase (Pre-Step 1) | ||||||||||
 Baseline Level, (α0) | 4.858 | 0.214 | 4.440 | 5.277 |  < 0.0001* |  Excess level, (α4) (vs Control) | -0.610 | 0.254 | -1.108 | -0.112 | 0.016** |
 Trend (α1) | 0.0176 | 0.018 | -0.018 | 0.054 | 0.34* |  Excess trend (α5) (vs Control) | 0.011 | 0.025 | -0.038 | 0.060 | 0.66** |
Post-Intervention phase (Post Step 6) | Post-Intervention phase (Post Step 6) | ||||||||||
 Excess level (α2) ( vs control grp pre-intervention baseline level) | 0.472 | 0.154 | 0.171 | 0.773 | 0.021
|  Excess level (α6) ( vs Control) | 0.388 | 0.302 | -0.203 | 0.980 | 0.198¥ |
 Excess trend (α3) (vs control grp pre-intervention trend) | 0.0016 | 0.019 | -0.037 | 0.040 | 0.94
|  Excess trend (α7) (vs control) | -0.053 | 0.027 | -0.107 | 0.0003 | 0.053¥ |
- *p-value < 0.05 indicates that uptake baseline levels and/or trends during pre-intervention phase, for the control facilities, is significantly different from zero
-
p-value < 0.05 indicates that the post-intervention phase uptake level and/or trend, for the control facilities, is statistically significant (vs pre-intervention phase) - **p-value < 0.05 indicates that the pre-intervention uptake level and/or trend, for the CaPSAI facilities is significantly different from the Control group facilities) (i.e. excess pre-intervention change ( is significant)
- ¥p-value < 0.05 indicates that excess uptake level and/or trend in Intervention arm during post intervention phase is statistically significantly different from the excess change in the Control arm (The p-value tests the overall effectiveness of the CaPSAI on changing the level and/or trend in contraceptives uptake, after all known pre-intervention and other confounding factors in both groups have been accounted for
p-value < 0.05 indicates that the post-intervention phase uptake level and/or trend, for the control facilities, is statistically significant (vs pre-intervention phase)