Association between malaria control and paediatric blood transfusions in rural Zambia: an interrupted time-series analysis

Table 4 Association between malaria control scale-up over time and paediatric blood transfusions - interrupted time series

	Dependent variable: paediatric blood transfusions
Independent variables	(1)	(2)
Monthly trend 2000-2003	−0.60***	−0.52***
	(0.12)	(0.10)
Difference in monthly use 2004–2008 (relative to 2000–2003)	−10.58***
	(3.35)
Monthly trend 2004-2008	−0.02
	(0.04)
Difference in monthly use 2004–2006 (relative to 2000–2003)		−19.06***
		(3.56)
Monthly trend 2004-2006		0.34***
		(0.10)
Difference in monthly use 2007–2008 (relative to 2000–2003)		−27.47***
		(6.56)
Monthly trend 2007-2008		−0.14
		(0.22)
Constant	56.59***	55.12***
	(7.62)	(7.45)
Observations	108	108
Mean of dependent variable	22.79	22.79
Wald Chi-Squared	2734.85	3740.52

*Significant at 10%; **significant at 5%; ***significant at 1%. Coefficient standard errors are noted in parentheses.
Regressions use an autoregressive-moving-average model with 2 lags and robust standard errors. We control for the month of the year in all regressions to account for seasonal correlations. For column 1, monthly trend 2000–2003 represents the rate of change by month for the outcome variable from 2000–2003. Monthly trend 2004–2008 represents the rate of change by month for the outcome variable from 2004–2008. Difference in monthly use in 2004–2008 represents the average difference in the outcome between 2000–2003 period and 2004–2008 period. For column 2, monthly trend 2000–2003 is defined as in column 1. Monthly trend 2004–2006 represents the rate of change by month for the outcome from 2004–2006. Monthly trend 2007–2008 represents the rate of change from 2007–2008. Difference in monthly use in 2004–2006 represents the average difference in outcome between 2000–2003 and 2004–2006. Difference in monthly use in 2007–2008 represents the average difference in outcome between 2007–2008 and 2000–2003.

ISSN: 1475-2875