Time to tackle unwarranted variations in practice
BMJ 2011; 342 doi: https://doi.org/10.1136/bmj.d1513 (Published 17 March 2011) Cite this as: BMJ 2011;342:d1513
All rapid responses
Rapid responses are electronic comments to the editor. They enable our users to debate issues raised in articles published on bmj.com. A rapid response is first posted online. If you need the URL (web address) of an individual response, simply click on the response headline and copy the URL from the browser window. A proportion of responses will, after editing, be published online and in the print journal as letters, which are indexed in PubMed. Rapid responses are not indexed in PubMed and they are not journal articles. The BMJ reserves the right to remove responses which are being wilfully misrepresented as published articles or when it is brought to our attention that a response spreads misinformation.
From March 2022, the word limit for rapid responses will be 600 words not including references and author details. We will no longer post responses that exceed this limit.
The word limit for letters selected from posted responses remains 300 words.
Dear Editor,
We congratulate BMJ for echoing Wennberg's work. The attention
received is quite stimulating for the Spanish Medical Practice Variation
Atlas project, a nation-wide research initiative born in Spain in 2002
(www.atlasvpm.org). Inspired by the Dartmouth's Atlas, our project
analyzes unwarranted variation in medical practice across the Spanish
National Health System healthcare areas, as well as poor outcomes amenable
to hospital care.
Along this decade of work, we have learnt how similar and yet
different should the analysis of variation in Europe be as compared to the
US. The key insight might be that Wennberg's taxonomy of underlying
factors prevails, but their relative weight in the explanation of
variations needs to be understood in the specific organizational and
political context. This consideration is extremely relevant to fulfil the
ultimate goal of the exercise: to manage these factors with a view of
tackling unwarranted variations. Three key observations built upon our
experience analyzing unwarranted variations in Spain are:
First. Elective surgery is conceived as a preference-sensitive
service in the Dartmouth framework. Our analysis for the Spanish context
reveals that under the preference-sensitive label we can, in fact, find
the three different types of services in Wennberg's taxonomy (effective,
supply-sensitive, and preference-sensitive care). Take, for instance, knee
replacement performed in patients with high, intermediate or low WOMAC
scores -a measure of pain and disability. Whereas intervention on patients
with high scores could be considered appropriate and evidence-based
(effective care), low scores suggest patients who are inappropriately
chosen for surgery (supply-sensitive care). In turn, patients having an
unclear benefit-risk balance might be preference-led. All three components
need to be accounted for in designing the measures to tackle variation.
Second. Cross-sectional studies, particularly those based on ecologic
designs, benefit from observation over time when it comes to identifying
possible underlying factors. In the Spanish case, we have learnt that a
joint analysis of the evolution of both, the rate of utilization and the
evolution of the variance, is needed. Actually, we may observe a reduction
of the variation with a simultaneous increase in the utilization rates
over time. Depending on the type of service, the reduction of this
variation should not be necessarily understood as a symptom of improvement
but as a widespread overuse (e.g. the reduction in prostate surgery or c-
section variation across the territory with increasing rates of
utilization).
Third. Utilization rates vary in a narrower range across healthcare
areas in Spain than in the US, but the most noticeable discrepancy lays on
the lower rates of utilization (e.g. spine surgery rate in 2003 was about
43 per 10,000 Medicare insured vs. 3.4 per 10,000 people aged 65 and over
in the Spanish NHS). This finding, which was also pointed out for the UK
in Mays' comment [1] , is relevant in terms of the actual impact on the
overall healthcare budget. Most important, it underlines the different
organizational and financial incentives built into each healthcare system.
Thus, dealing with unwarranted variation requires a unit of analysis
relevant from the policy point of view, reflecting decision-making
processes. The study of conservative breast cancer surgery rates in Spain
illustrates this point. Up to a 60% of variation across healthcare areas
was explained by the Autonomous Region of reference, where decisions about
screening policy and health services planning and provision are made.
The comparison of the results obtained from the experiences in the
US, Spain and, more recently in The Netherlands and the UK, suggests the
importance of including system idiosyncratic characteristics to adequately
interpret the underlying "causes" of unwarranted variation. Following
these lines the new EU 7th Framework Programme project, European
Collaboration for Healthcare Optimization (ECHO) was implemented last
year. (http://www.echo-health.eu/) The ECHO project studies healthcare
system performance in seven European countries, by eliciting variation in
practice, and accounting for each of the relevant contexts from provider
to country.
[1] BMJ 2011; 342:d1849
Competing interests: No competing interests
Variation in practice reflecting regional variation in disease presentation: Tonsillectomy
Following the publication of Wennberg's article highlighting
variations in practice for elective surgical procedures in both the United
States and English health care systems,(1) a retrospective study of
regional tonsillectomies performed at the James Paget University Hospital
over the past ten years was conducted. The purpose of this investigation
was to identify any annual variation in the number of tonsillectomies
performed and to see if this correlated with annual frequency of disease
presentation in terms of acute tonsillitis and quinsy admitted to hospital
for treatment. The results are displayed in figure 1. A positive
correlation is observed between combined patient numbers admitted into
hospital for treatment of acute tonsillitis and quinsy, and the number of
tonsillectomies performed per annum between 2000/2001 and 2010/2011.
Statistical testing with regression analysis derives a regression
coefficient with a P value of 0.009, indicating statistical significance
to this correlation.
SIGN guidelines on management of sore throat and indications for
tonsillectomy were first published in 1999 and have been adhered to at the
James Paget University Hospital for this decade long retrospective study,
thereby standardising criteria for tonsillectomy.(2) Annual variations in
tonsillectomy performed in this region are therefore dependant on the
incidence of recurrent tonsillitis and quinsy presenting to the
department, the former of which appears to be cyclical in nature (figure
1). On a national scale at any given point in time, regional variations in
prevalence and incidence of disease will therefore result in regional
variations in treatment delivered. Where national guidelines are widely
adopted, any regional variations in practice observed should be
interpreted with caution as they may simply reflect variation in regional
disease presentation rather than non-compliance to evidence based
guidelines.
(1) Wennberg JE. Time to tackle unwarranted variations in practice.
BMJ2011;342:d1513.
(2) Scottish Intercollegiate Guidelines Network. Management of sore
throat and indications for tonsillectomy.
2010.www.sign.ac.uk/guidelines/fulltext/117/index.html
Competing interests: No competing interests