The association between health workforce availability and HIV-program outcomes in Côte d’Ivoire

Table 3 Non-parametric regression for effects of healthcare worker category and HIV-program outcomes

	Number of HIV-positive cases identified			Number of HIV-positive cases newly initiated on ART			Viral load suppression rate¹
	β	(IQR)	p	β	(IQR)	p	β	(IQR)	p
Cadre and patient load adjusted models
Clinical	− 0.37	(− 0.82, 0.43)	0.798	− 0.36	(− 0.83, 0.46)	0.879	− 0.00	(− 0.02, 0.01)	0.930
Pharmacy	− 3.1	(− 6.42, 1.60)	0.329	− 3.03	(− 5.97, 1.64)	0.358	− 0.05	(− 0.06, 0.01)	0.678
Laboratory	7.02	(3.53, 8.08)	0.000	6.41	(3.31, 7.41)	0.000	− 0.02	(− 0.02, − 0.01)	0.389
Management	− 0.51	(− 1.68, − 0.04)	0.689	− 0.74	(− 1.85, − 0.29)	0.617	0.02	(0.01, 0.05)	0.599
Lay Workers	0.42	(− 0.32, 2.11)	0.284	− 0.36	(− 0.38, 2.09)	0.268	0.06	(0.04, 0.07)	0.037
Support Staff	− 1.07	(− 1.96, − 0.14)	0.323	− 1.21	(− 2.00, − 2.09)	0.255	− 0.03	(− 0.07, − 0.00)	0.411
Cadre, patient load, region, facility type and health system level adjusted models
Clinical	0.04	(− 0.93, 1.00)	0.621	0.04	(− 0.89, 0.95)	0.581	0.06	(− 0.01, 0.12)	0.474
Pharmacy	− 2.08	(− 4.24, − 0.01)	0.082	− 2.03	(− 3.87, − 0.10)	0.091	0.19	(− 0.03, 0.27)	0.500
Laboratory	3.84	(2.00, 7.48)²	0.000	3.60	(1.74, 6.74)³	0.000	− 0.03	(− 0.06, 0.01)	0.824
Management	0.12	(− 1.37, 3.27)	0.477	0.10	(− 1.35, 3.21)	0.467	− 0.11	− (0.03, 0.24)	0.657
Lay Workers	1.92	(0.85, 2.88)²	0.001	1.93	(0.86, 2.82)³	0.002	0.20	(0.17, 0.28)	0.057⁴
Support Staff	− 1.03	(− 1.83, − 0.10)	0.091	− 0.94	(− 1.71, − 0.05)	0.089	− 0.21	(− 0.31, − 0.13)	0.190

The bold represents values with a p-value less than 0.05
β Regression Coefficient, IQR 25th, 75th centile values
^aViral load suppression proportions are expressed as percent in this table
^bThese cadres had non-linear associations with the number of patients identified, the quadratic term coefficients are as follows: Laboratory (β = 0.03, IQR:0.02–0.05, p < 0.001), Lay worker (β = 0.09, IQR:0.05–0.11, p < 0.001)
^cThese cadres had non-linear associations with the number of patients initiated on ART, the quadratic term coefficients are as follows: Laboratory (β = 0.03, IQR:0.02–0.05, p < 0.001), Lay worker (β = 0.09, IQR:0.05–0.11, p < 0.001)
^dThe lay worker cadre had a non-linear association with viral suppression (β = 0.004, IQR:0.003–0.005, p = 0.082)

ISSN: 1478-4491