Skip to main content

Table 1 Study characteristics of included studies (n = 119)

From: Sample size requirements are not being considered in studies developing prediction models for binary outcomes: a systematic review

Characteristic

n

%

Study type

  

Development only

100

84.0%

Development with external validation

19

16.0%

Data source

  

Existing data

75

63.0%

Prospectively collected cohort

23

19.3%

Registry data

11

9.2%

Case control study

5

4.2%

Cross-sectional

4

3.4%

Randomised controlled trial

1

0.8%

Clinical specialty

  

Oncology

20

16.8%

Cardiovascular related

19

16.0%

Gastroenterology and hepatology

13

10.9%

Infectious diseases

9

7.6%

Neurology

9

7.6%

Obstetrics and gynaecology

9

7.6%

Endocrinology and metabolic disorders

5

4.2%

Respiratory disorder

5

4.2%

Trauma and Orthopaedics

5

4.2%

Intensive Care Unit

4

3.4%

Radiology

3

2.5%

General population

3

2.5%

Urology

3

2.5%

Haematology

3

2.5%

Surgical patients

2

1.7%

Rheumatology

1

0.8%

Geriatric medicine

1

0.8%

Paediatrics

1

0.8%

Emergency medicine

1

0.8%

Ophthalmology

1

0.8%

Psychiatry

1

0.8%

Not specified

1

0.8%

Prediction type

  

Prognostic

88

73.9%

Health event*

30

34.1%

Mortality

19

21.6%

Complication

16

18.2%

Health status**

14

15.9%

Treatment related

8

9.1%

Readmission

1

1.1%

Diagnostic

31

26.1%

Disease presence

31

100.0%

  1. *Health event = presence of a disease or occurrence of an even at a given time point
  2. **Health status = regression or progression of patient health