BACKGROUND: A comprehensive Primary Health Care approach includes clear referral and continuity of care pathways. South Africa lacks data that describe Intermediate Care (IC) services and its role in the health system. This study aimed to describe the model of service provision at an IC facility and the role it plays in the continuity of care in Cape Town. METHODS: Sixty-eight patients (65% Response Rate) were recruited in a prospective cohort design over a one-month period in mid-2011. Patient data were collected from a clinical record review and an interviewer-administered questionnaire, administered at a median interval between admission and interview of 11 days to assess primary and second ary diagnosis, knowledge of and previous use of Home Based Care (HBC) services, reason for admission, demographics and information on referring institution. A telephonic interviewer-administered questionnaire to patients or their family members post-discharge recorded their vital status, use of HBC post-discharge and their level of satisfaction with care received at the IC facility. A Cox regression model was run to identify predictors of survival and the effect of a Care-plan on survival. Seventy staff members (82%) were recruited in a cross-sectional study using a self-administered questionnaire to describe demographics, level of education and skills in relation to what they did for patients and what they thought patients needed. RESULTS: Of the 68 participants, 38 % and 24% were referred from a secondary and tertiary hospital, respectively, and 78% were resident of a higher income community. Stroke (35%) was the most common single reason for admission at acute hospital. The three most common reasons reported by patients why care was better at the IC facility than the referring institution was the caring and friendly staff, the presence of physiotherapy and the wound care. Even though a large proportion of the IC inpatients had been admitted in a health facility on the year preceding the study, only 13 patients (21%) had used a Community Health Worker (CHW) ever before and only 25% (n=15) of the discharged patients had a confirmed CHW visit post-discharge. The presence of a Care-plan was significantly associated with a 62% lower risk of death (Hazard Ratio: 0.380; CI 0.149-0.972). Notably, 46% of staff members reported performing roles that were outside their scope of practice and there was a mismatch between what staff reported doing and their actual tasks. In addition, of the 57 patients that could be traced on follow-up 21(37%) had died. CONCLUSION: Patients and family understood this service as a caring environment that is primarily responsible for rehabilitation services. Furthermore, a Care-plan which extends beyond admission could have a significant impact on reducing mortality. IC services should therefore be recognised as an integral part of the health system and it should be accessed by all who need it.