Skip to main content

Table 4 Regression discontinuity results for critical path indicators; 2010–2015

From: Using the critical path method to rollout and optimise new PMTCT guidelines to eliminate mother-to-child transmission of HIV in Zimbabwe: a descriptive analysis

Indicator P-value for Sktest of normality Slope of linear curve before introducing intervention (β1) Slope of linear curve after introducing intervention (β2) Decision
P-value* Slope (95% CI) Slope (95% CI)
ANC Bookings 0.3434 0.001 (−0.005–0.006) 0.310 (0.220–0.401) β2 > β1 > 0
HIV Testing 0.1542 −0.001 (− 0.004–0.002) 0.029 (− 0.023–0.812) a
AZT Prophylaxis 0.8391 0.142 (− 0.002–0.030) 0.271 (0.108–0.443) β2 > β1 > 0
CD4 Testing 0.0560 0.008 (0.001–0.015) 0.168 (0.040–0.296) β2 > β1 > 0
Mothers’ ART 0.0024 0.005 (−0.003–0.014) 0.160 (0.036–0.282) β2 > β1 > 0
Infant ARVs 0.3617 0.007 (0.002–0.012) −0.064 (− 0.144–0.017) a
EID 0.1502 0.010 (− 0.002–0.022) 0.309 (0.107–0.511) β2 > β1 > 0
  1. * Indicator meets normal distribution criteria if P > 0.05 or sample is large (> 3000 in all indicators, Table 3)
  2. a 95% CI of β1 and β2 overlapping or β2 < β1 hence no difference in slope between the two segments of the indicator’s curve, before and after the start of interventions