Markov modelling of viral load adjusting for CD4 orthogonal variable and multivariate conditional autoregressive mapping of the HIV immunological outcomes among ART patients in Zimbabwe

Table 4 The joint effects of time-varying CD4 cell counts levels and orthogonal CD4 cell counts variables on the viral load transition rates Markov model

Transition Rate λ_ij	Baseline	Log-linear β_ij		Hazard exp(β_ij)		Time-varying CD4 cell counts levels CD4_{states, k}
	λ_ij(VL, 0)	CD4_{states, k}	\( {P}_{CD4(k)}^{\ast } \)	CD4_{states, k}	\( {P}_{CD4(k)}^{\ast } \)	State 1	State 2	State 3
State 1 to 2 λ₁₂	0.0286	0.0434	−0.3383	1.0443	0.713	0.0305	0.0321	0.0337
State 1 to 3 λ₁₃	0.0025	− 1.0072	− 0.3336	0.3652	0.7163	0.0167	0.0059	0.0021
State 1 to 4 λ₁₄	0.0082	0.0288	0.0247	1.0292	1.0243	0.0078	0.0081	0.0083
State 1 to 5 λ₁₅	0.0159	− 0.7155	−1.1045	0.4889	0.3312	0.1774	0.0422	0.0206
State 2 to 1 λ₂₁	0.0161	− 0.3926	− 0.2849	0.6762	0.7521	0.0351	0.0237	0.0161
State 2 to 3 λ₂₃	0.0008	0.2065	0.6052	1.2293	1.8315	0.0004	0.0005	0.0006
State 2 to 4 λ₂₄	0.0057	0.0725	−0.2381	1.0752	0.7881	0.0056	0.0061	0.0066
State 2 to 5 λ₂₅	0.0305	− 0.7458	− 0.9312	0.4744	0.3941	0.3372	0.0759	0.0361
State 3 to 1 λ₃₁	0.0266	− 0.9855	−0.7571	0.3733	0.4691	0.1926	0.0728	0.0275
State 3 to 2 λ₃₂	0.0035	− 1.4092	−2.8306	0.2443	0.0589	0.1359	0.0324	0.0077
State 3 to 4 λ₃₄	0.0272	− 0.0811	0.3475	0.9221	1.4155	0.0255	0.0237	0.0221
State 3 to 5 λ₃₅	0.0343	− 0.5176	0.0947	0.5959	1.0993	0.1273	0.0457	0.0273

model information: −2*log-likelihood = 6930.71; Akaike’s information criterion = 7002.71

ISSN: 1742-4682