Rationale and Objectives: The key element for the integration of health networks is an effective system of reference and counter reference, understood as a mechanism for mutual referral of patients between different levels of service complexity. Thus, the aim of this study was to identify and analyze the reference system for the referral of patients diagnosed with tuberculosis in the city of Santa Cruz do Sul, Brazil. Methods: The study was conducted by PET Health Surveillance/Tuberculosis at the Reference Unit for Tuberculosis in the municipality (RU-TB). We conducted a descriptive survey study by reviewing the records of patients who underwent treatment for tuberculosis between the years 2007 and 2011. Data were tabulated in Excel and analyzed using SPSS 18.0. Results: A total of 239 medical records were evaluated and of these, 157 (65.7%) had no reference document or any type of referral; 53 (22.2%) had a medical prescription as a way of referral and 29 (12.1%) had the offi cial reference and counter-reference document. Of the 49 documents (20.5%) with information on the health unit that made the referral, 19 (38.8%) patients were referred to less complex units and 30 (61.2%) referred to greater complexity units. The time interval between the referral date and the reference and the start of treatment was 23 days, ranging from 0 to 296 days; the median was 4.5 days and mode was 1. There were no records related to counter-reference. Conclusion: These data show the importance of systems of reference and counter-reference for health networks, and that errors, such as failing in fi lling out these forms by health care professionals, can have an effect on health system users. KEYWORDS: Tuberculosis. Primary Health Care. Access to health services.