1 Teaching a statistical program in emergency medicine research rotations: Command- driven or click-driven? Muhammad Waseem, MD Research Director Emergency.

1

Teaching a statistical program in emergency medicine

research rotations: Command-driven or click-driven?

Muhammad Waseem, MD

Research Director

Emergency Medicine

Lincoln Medical & Mental Health Center

Bronx, New York

2

Disclosure

• Financial disclosures: None

3

Background

• Stata is a command-driven program

• It is a general-purpose statistical software package used by people of different backgrounds and professional disciplines

• Most Stata users are non-physicians

4

Background

• No matter which book you choose or which course you attend….

• Since Stata is used by people in many fields, most training programs offered are geared toward programmers and non-physicians

5

Background

• Although Stata has simple commands, these may be difficult for non-programmers to use

• Generally, physicians are familiar with “clicking on” rather than writing commands

6

Background

• Research is an important part of curriculum of emergency medicine residency training

• A Research Project is necessary for graduation of EM residents

7

Background

• In order to fulfill this requirement EM residents have a mandatory research rotation

8

Background

• ACGME fully accredited 3 year Emergency Medicine Residency Training Program

• 33 EM residents (11 / year)

9

EM Research Rotation

• Formulating a research question

• Basic research design

• Protocol writing and submission to IRB

• Data Collection

• Basic statistical concepts

10

EM Research Rotation

• Data recording

• Data entry

• Data analysis

11

Background

• First step is to organize the raw data into a format which is compatible with Stata

12

Objectives

• Does introduction of a Statistical program (such as Stata) improve the accuracy of data recording and data entry?

13

Methods

• Design: Observational

• Setting: Emergency Department of an Urban Teaching Hospital

• Participants: Emergency Medicine Residents

14

Methods

• Research Elective– Two blocks of 14 days each

– 1-2 per month (1:1 learning)

15

Methods

• Emergency Medicine Stata Manual describing basic operation of Stata created with version 10

16

Methods

• This manual was developed keeping in mind the needs of in-training Emergency Medicine Residents

• EM residents may have some basic knowledge of statistics but no knowledge of Stata– Emphasis on data entry and data import

17

Emergency Medicine Stata Manual

• Just an introduction designed for first time users

18

Methods

• Teaches EM residents the essential skills for successful data entry and data import to Stata

• Purpose is to reduce the number of incorrect entries and make the data set usable to Stata

• Hands-on experience using Stata

19

Methods

• Strengths– Tailored to the needs of a specific audience

(EM residents)

– Discusses only simple and basic operation of Stata

20

Methods

• Weaknesses– Tailored to the needs of a specific rotation

(EM residents research elective)

– Does not teach residents detailed statistical analysis

– Does not discuss programming

21

Methods

• Keep it simple

• Utilize pull-down menus (rather than commands)– No discussion on commands

22

23

Methods

• Getting started

• Stata screen

• 4 Windows, 9 Tabs

24

Methods

• 4 Windows– Review

– Variables

– Results

– Command

25

Methods

• 9 Tabs

• File, Edit, Prefs, Data, Graphics, Statistics, User, Window, and Help

26

27

Methods

• Basic commands with pull-down menus

Describe and summarize data

• Data input and output

Data Import

Saving files

28

Methods

• Describe data

• Data editor

• Data browser

29

30

31

Conclusions

• Basic operation of Stata can be learned quickly by EM residents

• Teaching Stata to the EM residents expanded their research related knowledge

• Providing introduction of Stata prior to data entry improve accuracy of data recording and facilitates data analysis

32

Limitations

• Observational study– No standardized method

• Small sample

• Single center study

33

Implications

• It is necessary to develop a training program that addresses the needs of physicians

34

Acknowledgement

• Biostatistics Course (2009) Weill Medical College of Cornell University New York