The test characteristics of a biased or ignorant diagnostician

Table 3 Examples for the influence of bias and ignorance on test outcomes

Scenario	Bias or ignorance			Test matrix			Biased test matrix			Outcome
Scenario	Dx+ (%)	Dx− (%)		Dx+ (%)	Dx− (%)		Dx+ (%)	Dx− (%)		Dx+ (%)	Dx− (%)
Unbiased tester
Dx+	100	0	T+	80	5	T+	80	5	p	50	50
Dx−	0	100	T−	20	95	T−	20	95	PPV	94	6
Biased tester
Dx+	0	100	T+	80	5	T+	20	95	p	50	50
Dx−	100	0	T−	20	95	T−	80	5	PPV	17	83

	Expert Dx+ (%)	Expert Dx− (%)		Dx+ (%)	Dx− (%)		Dx+ (%)	Dx− (%)		Dx+ (%)	Dx− (%)
Informed tester
Amateur Dx+	100	0	T+	80	5	T+	80	5	p	50	50
Amateur Dx−	0	100	T−	20	95	T−	20	95	PPV	94	6
Ignorant tester
Amateur Dx+	60	30	T+	80	5	T+	54	32	p	50	50
Amateur Dx−	40	70	T-	20	95	T-	46	69	PPV	63	37

T+ and T− represent a positive or negative test result, respectively; Dx+ and Dx− represent a positive or negative diagnosis, respectively; p & PPV represent pre-test probability & positive predictive value

ISSN: 1472-6947