- Aim: The aim of the present study was to compare implementation rates by primary care physicians of geriatric assessment recommendations given in various assessment settings. Methods: We compared Model "OCGAU", an Outpatient Comprehensive Geriatric Assessment Unit where there was no direct contact between the geriatrician and the primary care physician with three "Clinic" models of in-clinic geriatric assessment: Model "Clinic A-2007" in which the primary care physician participated in the assessment, Model "Clinic A-2013" where there was no contact with the primary care physician, and Model "Clinics B-2013" where the primary care physician participated in a staff meeting with the geriatrician in the clinic. Sub-groups of "OCGAU" model were comprised of patients referred to the geriatric unit by primary care physicians of patients included in three "Clinic" models. Results: Model "OCGAU" included 240 patients, Model "Clinic A-2007" 107, Model "Clinic A-2013" 127, and Model "Clinics B-2013" 133. The patients in Model "OCGAU" were older (mean age 83.2±6.2 years) than in "Clinic" models where the mean age was 79.7±6.5, 81.5±6.1, and 80.7±6.5, P<0.001). More recommendations were given per patient (6.4) in the Model OCGAU than in the "Clinic" models (range 1.9-3.9, P<0.05), but the implementation of recommendations by primary care physicians was lower in Model OCGAU (48.9%) than in "Clinic" models (range 56.9%-71.8%, P<0.005). Conclusions: Although more recommendations were made in the geriatric unit, the implementation rate was lower. This indicates the need for organizational changes, in particular improving communication between the geriatric staff and primary care physicians.