Stop press! After poking the BMC offices, my comments went live today; you can read them here.
A couple of months ago I sent some comments into BMC Health Services Research in response to this paper. The authors claim that hospital competition in the Netherlands CAUSED increases in patient satisfaction. There are 4 major flaws in this paper, causing misleading conclusions to be drawn in an area of intense political activity. Most notably, the causal increase is one-thousandth of the satisfaction scale. Wow! If I was one-thousandth of the way closer to total satisfaction, I’m not sure I’d notice. I’ve never seen a paper make a fuss about a regression coefficient so small that it had to be written in scientific notation! It is statistically significant because there are more than 20,000 patients in the dataset… so what?
I note that one of the authors works for a well-known consulting firm working in re-organisation of secondary care in the Netherlands, but of course I am entirely reassured by the statement that the authors have no competing interests.
NB these represent my professional opinion and are not (necessarily) the view of the University of London or Kingston University.
The Comprehensive Biomedical Research Centre are running courses in the Autumn, free to healthcare professionals in London if you work for Guy’s, King’s & Tommy’s, St George’s, SLAM, or Barts & the London.
Quantitative Research Methods Courses
- Correlation and linear regression
- Quantitative data analysis part 1
- Quantitative data analysis part 2
- Introduction to critical appraisal of quantitative research
- Data entry and management
- Introduction to statistical data management
- Introduction to SPSS
- Quantitative data analysis: multivariable analysis
- Introduction to meta analysis
Qualitative Research Methods Courses
- Introduction to critical appraisal of qualitative research
General Issues in Conducting Research
- Performing observational research – an introduction to different study design options
- Implementing patient and public involvement
- Evaluation of diagnostic tests
- Economic evaluation in healthcare
There is a wonderful article out last year in Science magazine by David Spiegelhalter and two colleagues which shows many different ways that have been tried out to communicate uncertainty (a broad term including sampling error i.e. confidence intervals, but other forms of uncertainty too). Some of them are very inventive. Worth a read!
Here’s the Bank of England’s GDP fan chart as a taster:
If you are one of our postgrad students in the Faculty of Health and Social Care Sciences, my stats surgery on Friday 8 June will now be at Kingston Hill from 12:00-13:30, somewhere in the Social Work department. Send me an email if you want to book a slot.
If you ever want to totally understand confounding in 8 minutes, watch this video! Wonderfully clear explanation from Gordon Guyatt at McMaster.
Length of stay (LOS) data are very skewed and often have genuine outliers, making it a hard situation to model adequately. A common objective is to compare the LOS between different health care providers, after adjusting for some patient characteristics. I have used multilevel negative binomial regression in the past and then got BLUPs for the random intercept as predictions of each hospital’s log discharge rate ratio compared to the overall average. So I was interested to see this new paper just out in BMC Med Res Meth. They compare various approaches on New Zealand ICU data (and I was relieved to see my one included). Worth reading or bookmarking should you ever encounter LOS data.
The Oxford Centre for Statistics In Medicine are running a course on randomised controlled trials (RCTs) over one week in September 2012. Should be a great foundation for anyone aiming to work in trials. The blurb says:
RANDOMISED CONTROLLED TRIAL COURSE: A Guide to Design, Conduct, Analysis, Interpretation and Reporting
17th September – 21st September 2012
Oriel College, Oxford
About the course: This course provides a thorough grounding in the principles and practice of randomised controlled trials (RCTs) for the evaluation of healthcare interventions. It will include talks and practicals to give examples and guidance on the methodology of trials using a problem-based learning approach.
* Clarify the fundamental principles and practice of RCTs
* Demonstrate the optimal methodology to use in RCTs based on considerable practical experience and using extensive examples from the literature
* Examine the critical issues involved in planning, conducting and completing a successful trial
* Provide a basic understanding of statistical input into RCTs per se including analyses frequently used
* Facilitate critical appraisal of published reports of RCTs
Who should attend: The course is aimed at persons planning or actively involved in trials, or individuals interested in furthering their knowledge of trial methodology. Applications are invited from clinical and non-clinical researchers and other professionals allied to medicine.