Executive summary

An assessment was carried out of the utilisation and analytical reliability of two dry chemistry pathology analysers (Kodak Ektachem DT60 and Leo HemoCue) in hospital ward side rooms at the Lyell McEwin Health Service, Elizabeth Vale, South Australia.

Tests using the instruments were carried out by resident medical officers (RMOs) at the hospital. The RMOs were provided with extensive training and backup by laboratory staff.

Over a period of thirteen weeks, utilisation of the equipment was low, with only two RMOs performing more than ten tests.

Reasons for low utilisation included continued availability of laboratory services, pressure of other work, disquiet at taking responsibility for results, reluctance to carry out quality control, and a perception by some individuals that performance of diagnostic tests was not part of their normal duties. Some individuals were deterred from performing tests by the need to centrifuge specimens, but this requirement was not a major reason for the low utilisation.

Analytical imprecision achieved by the RMOs was acceptable, as judged by criteria established for the trial, although the levels of imprecision were more than those obtained during the pre-trial evaluation of instruments.

Insufficient results were obtained for a comprehensive assessment to be made of levels of analytical inaccuracy achieved by the RMOs. On the basis of the limited data available from analysis of external quality control material it was found that:

  • levels of inaccuracy achieved by the RMOs for measurements of glucose and urea were acceptable;
  • random biases were shown for measurements of sodium similar to those noted in the pre-trial instrument evaluation. Similar biases were shown for measurements of potassium, and appeared to be operator-related. Analytical inaccuracy achieved for analyses of these analytes was not acceptable;
  • systematic biases, with unacceptable inaccuracy, were obtained for measurements of bilirubin and haemoglobin. Inaccuracy for haemoglobin measurements was satisfactory on the basis of comparing results with those of the coordinating laboratory on split patient specimens.

Requests by RMOs for after hours laboratory tests decreased during the period of the trial, but increased beyond the pre-trial level after ward testing was discontinued.

Hospital administrators should take into consideration the need for training, probable levels of utilisation, requirements for technical back up and benefits of existing laboratory services when planning the introduction of decentralised pathology testing facilities.