The modelling of the transmission of HIV-infection in selected European countries: a Markov chain approach

Stanislaw Maciej Kot, Gdansk Technical University
Jolanta Wojciechowska, University of Gdansk

In the paper, Markow Chains are applied to modelling the transmission of HIV-infections between the following groups ("states")of people: homo/bisexual men (s1), injection drug users (s2), and heterosexual people (s3). At the moment t, new infections are described by the discrete random variable Xt, which takes the state sj with the probability pj(t) = P(Xt = sj), j = 1,2,3. Let the conditional probability pij = P(Xt = sj| Xt-1 = si), i,j = 1,2,3, denotes the probability that a HIV-positive person, being infected at the moment t-1 by someone from the group si, infects the person from the group sj at the moment t. In the paper, a convenient method of the estimation of transition probabilities pij is proposed. Empirical examples are presented for selected European countries, using EuroHIV surveillance data for the period 2000 -2003.