[Development of a control method for sample distributors].
The use in routine of distribution automations in Blood Bank laboratories should imply regular testing of the equipment by users. This work was carried out on Microlabs as well as on an IEMS. We used a coloured solution (Alizarine Red in pH9 buffer). We applied the Beer-Lambert law, knowing that any volume modification induces a change in optical density. After testing the linearity of the photometers used in the laboratory, we set up 4 distribution programmes so as to simulate classic distribution of diluted or undiluted serum: 1) for large volume dispensing, 2) for small volume distribution in a small amount of buffer, 3) for distribution of small volumes in large buffer volumes, 4) a specific IEMS programme. Results, expressed as coefficients of variation (CV) for each volume, varied from 9.8 to 0.7% according to the programme. These results show that large volume distribution does not pose any problem (CV < 5%). When dilution occurs, the error observed in serum distribution adds up to that of the thinner, and the smaller the two volumes, the greater is the error. It therefore appears necessary to find the best compromise between serum volume and thinner volume to obtain a CV < 5%. This error may appear negligible in semi-quantitative tests, but it may be important in quantitative tests. Lastly, this programme can easily be adapted in routine to monitor the dispensers efficiently.