The impact of user fees on health services utilization and infectious disease diagnoses in Neno District, Malawi: a longitudinal, quasi-experimental study

Table 3 Results from sensitivity analyses which restricted the control group to MOH health centres offering free services through study period and geographically separated from CHAM user fee introducing/removing centres

	Total outpatient attendances	Total new malaria diagnoses, under 5 s	Total new malaria diagnoses, over 5 s	New confirmed HIV+ patient, aged 15–49
Periods 1 and 2: Introduction of user fees^a
Estimated change (%)	−70	−22	−56	−52
95 % confidence interval	[−92, 8]	[−78, 177]	[−86, 40]	[−77, 1]
P-value	0 · 060	0 · 66	0 · 14	0 · 051
Number of centres	8	8	8	8
Number of months	35	35	35	35
Periods 2 and 3: Removal of user fees^b
Estimated change (%)	345	170	261	-^c
95 % confidence interval	[136, 737]	[13, 545]	[130, 467]	-
P-value	0 · 001	0 · 031	<0 · 001	-
Number of centres	8	8	8	8
Number of months	27	27	27	27

The estimated change represents the average difference in monthly attendances or diagnoses associated with the introduction or removal of user fees. Standard errors were clustered at the health centre level
^aEstimated using data from the period July 2012 to June 2015
^bEstimated using data from the period August 2013 to October 2015
^cThere were two few new confirmed HIV+ patients treated at the centre which removed user fees to analyse

ISSN: 1472-6963