The aim of the study was to quantify the incidence of illness and treatment behaviour in relation to CD4 count, age, and gender among a cohort of persons living with HIV/AIDS in Thailand. 464 participants with a CD4 count between 50 and 550 cells/mm3 were followed up for 12 months. Multiple Poisson regression was used to model the adjusted incidence rate ratio of illness and care seeking at different levels. The incidence of morbidity and treatment pattern were significantly different among participants with different CD4 count, age and gender. For example, morbidity incidence was significantly higher among participants with CD4 count of less than 200 cells/mm, among female participants, and participants aged 35 years or over. Females made significantly higher use of hospital ambulatory care and private clinics than males and males made significantly more use of private pharmacies. The potential opportunity cost of not providing ART to these different groups can be estimated and used to inform further economic evaluation and policy decisions on whether to provide ART at all and which patient groups to prioritise.