Objective:To describe intensive care unit (ICU) facilities in Sri Lanka; to describe the pattern of admissions, case-mix and mortality; compare patient outcome against the various types of ICUs; and determine the adequacy and standards of training received by medical and nursing staff.Materials and Methods:Observational study of multidisciplinary (general) and adult speciality ICUs in government sector hospitals.Results:Hospitals studied had 1 ICU bed per 100 hospital beds. Each bed catered to 70-90 patients over a year. Death rates were comparable in each level of hospital/ICU despite differences in resource allocation. Fifty to 60% of patients had their original problems related to medicine, while only 35% - 45% were surgical. Thirty two percent of medical patients and 15% of surgical patients died. More than 90% of ICUs had a multi-monitor for each bed. Seventy seven percent of ICUs had one or more ventilators for each bed. Arterial blood gas (ABG) facilities were available in 83% of ICUs. There were serious inadequacies in the availability of facilities of 24 hour physiotherapy (available only in 36.7%), 24 hour in hospital Ultra Sonography (22.4%), electrolyte analyser in ICU (54.2%), haemodialysis / continuous renal replacement therapy (HD/CRRT) (41.7%), and Echocardiography. Medical Officers’ training was anaesthetics dominated as opposed to a multidisciplinary training. There was a severe shortage of critical care trained nurses.Conclusions:Only limited evolution has taken place in intensive care over the past 5 years. The reasons for higher death rates in medical patients should be investigated further. Moving towards a multidisciplinary approach for training and provision of care for ICU patients is recommended.