Table 2 Description of parameters for combined models of pre- and post-treatment data

β ₀

Pre-treatment mean intercept

β ₁

Pre-treatment mean slope

U ₀₀

Pre-treatment intercept subject-specific random effect variance

ρ

Correlation between pre-treatment intercept and slope subject-specific random effects

U ₁₁

Pre-treatment slope subject-specific random effect variance

σ

Standard deviation of residual error term for each measurement, shared by pre- and post-treatment parts of model

κ _pre

Scale parameter for pre-treatment fBM process

H _pre

Hurst index for pre-treatment fBM process

ϕ₁ model

These parameters relate to the long-term maximum value of the response variable after treatment initiation

At1₁,At1₂

Intercept and slope terms in relationship with \(u_{i}^{+}\) for patients treated within 6 months of seroconversion

At2₁,At2₂

Intercept and slope terms in relationship with \(u_{i}^{+}\) for patients treated beyond 6 months but within 1 year of seroconversion

A₁,A₂

Intercept and slope terms in relationship with \(u_{i}^{+}\) for linear or NCS models ^a

A₃,A₄

Third and fourth coefficients for NCS models ^a

ϕ₂ model

These parameters relate to the rate of recovery of the response variable after treatment initiation

Bt1₁,Bt1₂

Intercept and slope terms in relationship with \(u_{i}^{+}\) for patients treated within 6 months of seroconversion

Bt2₁,Bt2₂

Intercept and slope terms in relationship with \(u_{i}^{+}\) for patients treated beyond 6 months but within 1 year of seroconversion

B₁,B₂

Intercept and slope terms in relationship with \(u_{i}^{+}\) for linear or NCS models ^a

B₃,B₄

Third and fourth coefficients for NCS models ^a

P

Residual variance for long-term maximum (ϕ_1:i) not explained by \(u_{i}^{+}\)

κ _post

Scale parameter for post-treatment fBM process

H _post

Hurst index for post-treatment fBM process

d f _pre

Degrees of freedom parameter for pre-treatment stochastic process

d f _post

Degrees of freedom parameter for post-treatment stochastic process

ρ _Moran

Correlation parameter for latent scaling variables of pre- and post-treatment stochastic processes