Objective: To compare the comprehensive care plan assessment (Minimum Data Set, MDS) of vision with the clinically evaluated visual status of nursing home residents to determine the accuracy of the vision care plan. Design: Descriptive study. Setting: One publicly funded and four privately owned long-term care nursing facilities. Participants: 151 nursing home residents undergoing vision examination at the nursing home request. Mean age of 81, range 34 to 97 years. Main Outcome Measure: The charted MDS assessment for vision patterns was compared with the MDS plan which would have been formulated based on direct clinical vision examination. Determinations were made based on the three areas required by MDS: visual acuity (VA), peripheral vision, and presence of prosthetic visual devices. Conclusions: The MDS care plan for vision and actual visual status agreed in only 34% (52/151) of persons. The Vision (VA) subsection overestimated VA in 41 % and underestimated VA in 11%. Concordance for VA was extremely poor (kappa 0.176). In no participants did the MDS indicate a visual field defect, whereas 16.5% (25/151) did show visual field restriction on examination. Care plan assessments also failed to note the use of visual prosthetic devices in 17% (26/151). These results indicate that MDS care plan assessment for vision is often inaccurate and may result in the lack of appropriate vision care being triggered through resident assessment protocols (RAP). © 1995 American Academy of Optometry.