An examination of the reproducibility of three tests for identifying patients at risk of diabetic foot complications is reported. One thousand and one patients attending a general diabetic outpatient clinic at the Royal Liverpool University Hospital underwent screening using the Semmes-Weinstein monofilaments, the biothesiometer, and palpation of pedal pulses. Two hundred and fifty-nine patients who were found to have a deficit on at least one of these tests were given a second appointment where the tests were repeated. Two hundred and twenty-nine patients attended for their second appointment. The expected range of variation in readings for the biothesiometer was 5 log hertz, sufficient variation to make the biothesiometer unacceptable as a screening instrument. The range of reproducibility for the four pedal pulses was from 68% to 81%, also an unacceptable level of variation. Over 85% of the results for the Semmes-Weinstein monofilaments were the same on the first and second measurement. Of the three instruments examined in this study, only the Semmes-Weinstein monofilaments have sufficient reproducibility in the clinical setting for them to be recommended for general use as a screening test for risk of diabetic foot complications.