Methods

Figure 1

Table 1
 

Methods

Study population

Patients were recruited to the trial at times when a middle grade or consultant in emergency medicine was available in the department. All patients in whom a blood glucose test was requested during these times were eligible.

Study design

Treatment allocations were prepared in a series of sealed opaque envelopes.

Statistical power

An a priori power calculation suggested that about 30 patients would be required in each treatment group to achieve a 90% power of detecting a 20 mm difference on the visual analogue scale between groups.
 

Figure 1


    Participant flow

Table 1

    Characteristic
    Ear group

    (n=30)

    Thumb group

    (n=30)

    Significance
    Mean age (years)
    52
    54
    P=0.88
    Diabetic
    5
    7
    P=0.52
    Male
    17
    13
    P=0.79
    Previous test
    13
    19
    P=0.12

    Values are numbers of participants, unless stated otherwise