We recommend that you upgrade to the latest version of your browser.

Statistical guidance

Statistical analysis is a crucial tool in medical research. It is used to organize information and provide data based decision support. In other words, it tells the story hidden within the data.

Fingre på tastatur.
 Illustration photo. Colourbox.com.

 

Knowledge and understanding of data analysis should be present in both the planning, implementation, and finalizing stages of a study.

About Section of Biostatistics

The section is part of the Department of Research and a vital component of the research support offered at SUH (Stavanger University Hospital). Our guidance service in statistics and methodology is available to everyone engaged in research at SUH. The general procedure for inquiries regarding statistical advice is to submit a short description of the project and of the guidance sought. Alternatively, we have drop-in guidance (“Statistisk poliklinikk”) Mondays 1-3 pm. The principle of empowering researchers to learn the methods used is a key goal.

In collaboration with the University of Stavanger, we offer two courses in medical statistics: the introductory course TN907 (covering descriptive statistics, tests, analysis of variance, linear regression) and the advanced course TN908 (logistic regression, survival analysis, mixed models).

Western Norway Regional Health Authority (“Helse Vest”) has a regional competence center located at Haukeland University Hospital that provides advice on experimental design, data collection, data processing, and statistical analysis. This service is available to all researchers in health trusts in Helse Vest, but researchers at SUH are encouraged to rather contact us.

Section of Biostatistics has offices in “Forskertua”, i.e. on the 2nd floor of Jan Johnsens gate 4 (former Maurtua kindergarten).

General inquiries regarding courses and guidance can be directed to the section leader, Ingvild Dalen.

There is a wealth of literature, software, and websites that are valuable tools for conducting statistical analysis of medical data. Below is an overview of statistical resources recommended by Section of Biostatistics.

Statistical Resources

Statistical Literature

A wide selection of reading material related to statistics is available on the market. Several textbooks serve as introductions to the field and others delve into specific topics. Additionally, the market offers guides and textbooks for statistical analysis software.

General Medical Statistics and Epidemiology

  • Odd O. Aalen (ed). "Statistiske metoder i medisin og helsefag," 2nd edition. Gyldendal akademisk, 2018.
  • Marit B. Veierød, Stian Lydersen, Petter Laake (eds). "Medical Statistics in Clinical and Epidemiological Research." Gyldendal akademisk, 2012.
  • Douglas G. Altman. "Practical statistics for medical research." Chapman & Hall, 1991.
  • Per Magnus, Leiv S. Bakketeig. "Epidemiologi," 4th edition. Gyldendal akademisk, 2013.

In-depth Coverage of Statistical Programs

  • Julie Pallant. SPSS Survival Manual: A Step by Step Guide to Data Analysis using IBM SPSS. 7th edition. Open University Press, 2020.
  • Andy Field. Discovering Statistics Using IBM SPSS Statistics, 6th edition. SAGE Publications Ltd, 2024.
  • Alan C. Acock. A Gentle Introduction to Stata. 6th edition. Stata Press, 2023.
  • Andy Field, Jeremy Miles, Zoë Field. Discovering Statistics Using R. SAGE Publications Ltd, 2012.
  • Ewen Harrison, Riinu Pius. R for Health Data Science. CRC Press, 2021.
  • Babak Shahbaba, Biostatistics with R: An Introduction to Statistics Through Biological Data. Springer, 2012.
  • Edward Curry, Introduction to Bioinformatics with R: A Practical Guide for Biologists, 1st edition. Chapman and Hall/CRC, 2020. 

Software

Helse Vest has licenses for several types of software for storage and processing of collected data. The choice of programs depends partly on personal preference and partly on the nature of the study. Biostatisticians at SUH have extensive experience with programs such as SPSS, STATA, and R, as well as familiarity with SAS, MATLAB, and Python.

SPSS

IBM SPSS Statistics is user-friendly with good-quality manuals. Contacting Helse Vest IKT support provides access to all SPSS statistical package programs at no extra cost. Helse Vest's course catalog offers an introductory course in SPSS. SPSS has an intuitive interface with point-and-click features and quick menus for desired commands. Additionally, commands can be generated as syntax within the program, allowing for storage and repetition.

Additional packages with extended features can be added to SPSS. For example, the Stats Power package can be used to determine necessary sample size and test power. AMOS is another add-on package suitable for working with structural equation models, including confirmatory factor analysis.

SPSS Overview

STATA

STATA – Data Analysis and Statistical Software is well-documented, has its own syntax for commands, and includes some point-and-click features. STATA's syntax is more user-friendly and efficient compared to other statistical programs. The program offers more options than SPSS and allows for more advanced statistics, including robust methods such as robust regression models and robust standard errors. STATA is available in Helse Vest's software portfolio with a limited number of licenses, which can be expanded upon request.

STATA

R and RStudio

R is a free, open-source, command-line program with a wide range of packages for visualizing and analyzing various types of datasets. R offers more advanced statistical capabilities compared to SPSS, similar to STATA. New statistical methods are readily available in R. Different user-developed program packages must be downloaded and installed locally. RStudio provides a more user-friendly interface for running R.

For bioinformatics and analysis of genomic data, there are many established program packages with good tutorials available on Bioconductor and CRAN.

R
RStudio

MATLAB

MATLAB is typically used for (non-statistical) mathematics and has excellent matrix computing functions. MATLAB is also a good option for image processing. The program is available in the software portfolio for employees at SUH but requires a license, which must be purchased individually. Additionally, several available packages must be purchased separately.

MATLAB

Python

Python is a free programming language available at SUH, often used with Anaconda to simplify programming (similar to RStudio for R). Python is not a statistical program but is used for more complex analyses in genetic research, artificial intelligence, image analysis, or for creating custom programs or apps. Many modules/libraries with software are available for specific applications. The program is free but requires programming knowledge. Several good data analysis packages, such as Bioconda, are available.

Python

Other Programs

There are several other programs suitable for statistical analyses beyond those mentioned above. Statistica is easy to use and produces visually appealing graphs. EpiData is a freely available WHO database on public health, epidemiology, and medicine worldwide, suitable for clinical and epidemiological studies. EPI Info is designed for creating questionnaires, building databases, and performing simple statistical analyses tailored to epidemiological studies. S-Plus has a programming language similar to R but is not open source and is not available in Helse Vest's software portfolio.

Some programs are only available at Helse Bergen, including SAS, Maple, MedCalc, SigmaPlot, Graphpad Prism, and StatXact. Researchers at Helse Stavanger can request access to these programs via Helse Stavanger IKT.

SAS
S-Plus (S+)
Statistica
Maple
MedCalc
EpiData
EPI Info
Sigmaplot
GraphPad Prism
StatXact

Choosing the Right Statistical Program

To determine which statistical program is suitable for a study, one should consider all the characteristics of the program. Below is a table outlining the pros and cons of each statistical program (as perceived).

Program

Pros

Cons

Availability

SPSS

Easy to learn and process datasets

Limited robust methods, limited flexibility and quality in graphical representations, limited selection of more advanced methods

Helse Vest

STATA

Robust methods, user-friendly

-

Helse Vest

R

Open-source, almost unlimited statistical analyses, free

Steep learning curve, varying documentation quality, problematic with very large datasets

Helse Vest

MatLab

Matrix computation, mathematics, image processing, machine learning

Limited wrt general statistics, expensive

Helse Bergen and Helse Stavanger; licenses must be obtained for each user

Python

Artificial intelligence, creating small programs or apps, good for very large datasets, Data Science

Steep learning curve, fewer statistical program packages than R

Helse Vest

Statistica

Beautiful graphs, easy to use

Relatively expensive

Helse Vest

Overview of available analyses in SPSS, STATA, R, MatLab, and SAS

Web Resources

Several websites contain relevant literature and guides for specific statistical programs. or tools useful in statistical analysis.

Series of papers on medical statistics

  1. JAMA Guide to Statistics and Methods
  2. Medical statistics (biomedcentral.com)
  3. Statistical notes in BMJ

Online help for statistical programs

  1. OARC Stats – Statistical Consulting Web Resources (ucla.edu) (highly recommended!)
  2. CrossValidated
  3. SPSS Tutorials
  4. Quick-R

Other statistical tools

  1. VassarStats
  2. StatPages
  3. OpenEpi

 

Data Collection and Storage Solutions for Research Data

Contact

Biostatisticians: 

Ingvild Dalen, PhD (Head of section)

Email: ingvild.dalen@sus.no

 

Anastasia Ushakova, PhD

Email: anastasia.ushakova@sus.no

 

Hanne Brit Hetland, MSc

Email: hanne.brit.hetland@sus.no

 

Jan Terje Kvaløy, PhD

Email: jan.t.kvaloy@uis.no

 

Bjørn Henrik Auestad, PhD

Email: bjorn.auestad@uis.no

 

Bioinformatician:

Marie Austdal, PhD

Email: marie.austdal@sus.no

 

Last updated 1/29/2024