What method?

7 posts / 0 new
Last post
Nucleus
Nucleus's picture
What method?

I´m currently reviewing results of physician experience in number of cases on patient outcome.  I have a group of 39 physicians who treated from 1 to 29 cases and their results in the table below.  
Can anyone comment on the best way to present this data?
 
 

Physician

No of cases

Survived to admission

Survival to admission%

1

29

17

59%

2

21

12

57%

3

15

3

20%

4

14

6

43%

5

12

9

75%

6

11

4

36%

7

11

4

36%

8

11

3

27%

9

10

5

50%

10

9

3

33%

11

8

2

25%

12

6

2

33%

13

5

3

60%

14

5

1

20%

15

5

1

20%

16

5

1

20%

17

5

1

20%

18

5

1

20%

19

4

0

0%

20

3

2

67%

21

3

2

67%

22

3

2

67%

23

3

0

0%

24

3

1

33%

25

3

0

0%

26

3

2

67%

27

3

1

33%

28

2

0

0%

29

2

1

50%

30

2

1

50%

31

2

2

100%

32

2

0

0%

33

1

1

100%

34

1

0

0%

35

1

1

100%

36

1

1

100%

37

1

1

100%

38

1

0

0%

39

1

1

100%

 

232

97

42%

Satinder
Satinder's picture
Hello Nucleus

Hello Nucleus
This data raises a question that what  disease are patients ithe ncluded suffering from?
In case the disease is cardiovascular, diabetes or tumor then stage of entry of the patient will be a key
factor.
At the same time physician handling them will be of the key importance too, as the senior or expert may be handling those in advanced stages and hence have higher rates of associated mortality or survival ( could be either of the 2 depending on diseases)
 

Nucleus
Nucleus's picture
This study is about cardiac

This study is about cardiac arrests, where the outcome is traditionally not reported based on the underlying morbidities but only the overall survival.  The question about physicians is about how the experience in number of cases affects their results.
Should the group be split by some number in more or fewer than 5 cases our would another form of data presentation be more appropriate?

hilltrekker
hilltrekker's picture
I think you can reprsent the

I think you can reprsent the data graphically and these days it has become very easy to do it with the computer.
For stastical purposes readymade stat program like instat can be tried.

Monu
Monu's picture
You try with basic

You try with basic biostatistics tools and also student-t test. Or if you are very new and want to do it fast try some statistical softwares like SPSS or even yo can do it on a MS-Excel worksheet. This is not very complex as have only one variable.

Nucleus
Nucleus's picture
Thanks for helpful comments.

Thanks for helpful comments.
I created a scatter plot in excel that gives a nice graphic presentation of the data.  Can anyone comment on what these numbers the plot gives really mean and how to interpret them?

y = -0.0033x + 0.2317
R² = 0.0063

Monu
Monu's picture
Nucleus,

Nucleus,
You really did a great job. Many times I used to use sophisticated techniques but the answer comes up many times very easily using simpler methods this happened with you also. Excel is the simplest and cosmopolitan data analysis tool available and yougot your answer through that. Thats really good. I am happy to know this.