RESUMEN
Despite the fragmentation of healthcare provision being considered one of the main obstacles to attaining effective health care in Latin America, very little is known about patients' perceptions. This paper analyses the level of continuity of health care perceived by users and explores influencing factors in two municipalities of Colombia and Brazil, by means of a cross-sectional study based on a survey of a multistage probability sample of people who had suffered at least one health problem within the previous three months (2163 in Colombia; 2167 in Brazil). An adapted and validated version of the CCAENA© (Questionnaire of care continuity across levels of health care) was applied. Logistic regression models were generated to assess the relationship between perceptions of the different types of health care continuity and sociodemographic characteristics, health needs, and organizational factors. The results show lower levels of continuity across care levels in information transfer and care coherence and higher levels for the ongoing patient-doctor relationship, albeit with differences between the two countries. They also show greater consistency of doctors in the Brazilian study areas, especially in primary care. Consistency of doctors was not only positively associated with the patient-doctor ongoing relationship in the study areas of both countries, but also with information transfer and care coherence across care levels. The study area and health needs (the latter negatively for patients with poor self-rated health and positively for those with at least one chronic condition) were associated with all types of continuity of care. The influence of the sex or income varied depending on the country. The influence of the insurance scheme in the Colombian sample was not statistically significant. Both countries should implement policies to improve coordination between care levels, especially regarding information transfer and job stability for primary care doctors, both key factors to guarantee quality of care.