[Sampling limitations in studies of primary care: comparison of 4 sampling designs]

Different sampling methodologies have been used to avoid the great amount of information which arises in Primary Health Care. The objective of this study is to compare the estimates of four sampling schemes, analyzing variables of utilization, referral activity and diagnostic morbidity, based on data obtained from consultations in a health centre over a one-year period. The percentage of error between each sample estimation and the real value for the whole data set was obtained. Sampling schemes using week periods showed the worst results (40% of error in the estimation of chronic consultations, 18% of error in the percentage of radiology requests and more than 13% in the estimation of morbidity). The findings of this study suggest that some of the sampling schemes usually applied in primary care studies have important limitations, showing the need for more research in this field.