Study designDiagnostic performance study requirements and dataset layout.

Requirements

  • A qualitative or quantitative variable of the measured values or indications (positive/negative) of the diagnostic test.
  • A qualitative variable indicating the true state (positive/negative) of each subject.

Dataset layout for a single test

Use a column for the true state variable (State), and a column for the test measured values (Test); each row has the values of the variables for a case (Subject).

Subject (optional) State Test
1 Diseased 121
2 Diseased 118
3 Diseased 124
4 Diseased 120
5 Diseased 116
6 … …
7 Healthy 100
8 Healthy 115
9 Healthy 102
10 Healthy 98
11 Healthy 118
… … …

Dataset layout for paired tests

Use a column for the true state variable (State), and a column for each of the test measured values (Test 1, Test 2); each row has the values of the variables for a case (Subject).

Subject (optional) State Test 1 Test 2
1 Diseased 121 86
2 Diseased 118 90
3 Diseased 124 91
4 Diseased 120 99
5 Diseased 116 89
6 … … …
7 Healthy 100 70
8 Healthy 115 80
9 Healthy 102 79
10 Healthy 98 87
11 Healthy 118 90
… … …  

Dataset layout for independent tests

Use a column for the true state variable (State), a column for the factor (test or group indicator), and a column for the test measured values (Test); each row has the values of the variables for a case (Subject).

Subject (optional) State Test/Group ID Test result
1 Diseased 1 121
2 Diseased 1 118
3 Diseased 1 124
4 Diseased 1 120
5 Diseased 1 116
6 Healthy 1 100
7 Healthy 1 115
8 Healthy 1 102
9 Healthy 1 98
10 Healthy 1 118
11 … … …
12 Diseased 2 86
13 Diseased 2 90
14 Diseased 2 91
15 Diseased 2 99
16 Diseased 2 89
17 Healthy 2 70
18 Healthy 2 80
19 Healthy 2 79
20 Healthy 2 87
21 Healthy 2 90
… … … …

Dataset layout for a single binary test

Use a column for the true state variable (State), and a column for the test indications (Test); each row has the values of the variables for a case (Subject).

Subject (optional) State Test
1 Diseased P
2 Diseased P
3 Diseased P
4 Diseased N
5 Diseased P
6 Healthy N
7 Healthy N
8 Healthy N
9 Healthy N
10 Healthy N
… … …

Frequency form dataset layout for a binary test

Use a column for the true state variable (State), and a column for the test indications (Test) and a column for the number of cases (Frequency); each row has the values of the variables and the frequency count.

State Test Count
Diseased Positive 3
Diseased Negative 1
Healthy Positive 0
Healthy Negative 5

Frequency form dataset layout for 2 independent binary tests

Use a column for the true state variable (State), a column for the factor (test or group indicator), a column for the test indications (Test) and a column for the number of cases (Frequency); each row has the values of the variables and the frequency count.

State Test Group/Test ID Count
Diseased Positive 1 3
Diseased Negative 1 1
Healthy Positive 1 0
Healthy Negative 1 5
Diseased Positive 2 24
Diseased Negative 2 2
Healthy Positive 2 2
Healthy Negative 2 30

Frequency form dataset layout for 2 paired binary tests

Use a column for the true state variable (State), a column for each of the test indications (Test 1, Test 2) and a column for the number of cases (Frequency); each row has the values of the variables and the frequency count.

State Test 1 Test 2 Count
Diseased Positive Positive 3
Diseased Negative Positive 1
Diseased Positive Negative 24
Diseased Negative Negative 2
Healthy Positive Positive 0
Healthy Negative Positive 5
Healthy Positive Negative 2
Healthy Negative Negative 30
Available in

From the Statistical Reference Guide for Analyse-it 6.24.0: https://analyse-it.com/docs/user-guide/diagnostic-performance/study-design

Statistical Reference Guide v6.24.0