Healthy Skepticism International News

January 2002

Background for a research proposal

January/February 2002 Vol 20 No 1/2

By Peter R Mansfield

Healthy Skepticism International News

Introduction

Healthy Skepticism is moving into teaching and research. This edition presents the backgrounder for a proposal to investigate the teaching critical appraisal of promotion to change attitudes towards drug promotion and thus improve prescribing.  The following commences with a brief overview of the Australian general practice context but the trends are similar in many other countries. After that is a discussion of some risk factors for suboptimal prescribing and the analogy of cardiovascular disease as an aid to thinking about those risk factors and the types of research that are needed.  Finally we report the early results of some teaching of critical appraisal of promotion.

We would appreciate any comments on this backgrounder. 
Please send comments to .(JavaScript must be enabled to view this email address)

Home: www.healthyskepticism.org

The Australian general practice context

The quality and economy of prescribing by Australian general practitioners (GPs) is highly variable.  Looking at costs as an example, in calendar year 1993 the median prescription cost per consultation for “full time” vocationally registered GPs after omitting the lower and upper 5% of the distribution was $9.26. The interquartile range (the middle 50% of cases) was $6.87-$11.73 and the adjusted range was $2.57 - $25.64.[1]

One of the main causes of this variability is differences in case mix. However responses to case vignettes (eg National Prescribing Service case study 5) show wide variations GPs responses to identical “patients”. [2]  It appears that different GPs may diagnose similar patients differently and/or treat the same diagnosis differently.  Consider the example of antidepressants. Firstly, GPs vary in the tendency to diagnose various forms of distress that do not meet the criteria for major depression as “depression” rather than use other concepts.  Secondly, having diagnosed “depression” GPs vary in the tendency to prescribe an antidepressant alone vs non-drug treatment alone vs antidepressant plus non-drug treatment vs no treatment.  Thirdly, having decided to prescribe an antidepressant GPs vary in the tendency to prescribe newer more expensive drugs vs older cheaper alternatives.. [3] 

Some of the newer drugs are medically superior and sometimes even more cost effective than the older alternatives but many are not.  For example COX2 inhibitors are not more effective and appear to have a higher total rate of severe adverse effects than older non-selective NSAIDs. [4]  Other example include newer vs older drugs for asthma, diabetes, hyperlipidemia and hypertension.

In financial year 1992-93 the Pharmaceutical Benefits scheme cost taxpayers $1.52 billion and in 1993-94 it cost taxpayers $1.80 billion. [5];[6]  By financial year 2000-01 the cost to taxpayers had risen to $3.82 billion, an increase of 19.9% on the previous financial year.  The growth in prescriptions volumes was only 7.2% so the main driver of growth is the continuing trend for doctors to change to prescribing newer more expensive drugs. [7]  There is increasing concern that this growth rate is not sustainable so there is now a consensus that change is necessary but there is no consensus on what type of changes would be best. [8-10] 

On 8 March 2002 Australian Prime Minister John Howard announced there would be major policy changes regarding growth of the PBS.  He said the Government had to ensure “proper constraints and proper disciplines because the cost of prescribing these drugs on the so-called free list can grow exponentially if we’re not careful”.  Reportedly “the Finance Department is pressing the Health Department to introduce measures over four years to slash yearly spending on the Pharmaceutical Benefits Scheme (PBS) by $1 billion a year by 2006.” [11]

Pharmaceutical promotion is a significant cost driver and is “unambiguously bad for welfare”

A major study by Dutch economists of product level data sets for the promotion (detailing, advertising and direct mail) and sales of prescription pharmaceuticals in the Netherlands, for every month in the 1994-1999 period was published in March 2002.  The data covered eleven therapeutic markets; together these markets covered 58% of all markets for prescription pharmaceuticals when measured by total sales and 55% when measured by marketing expenditure.  They were able to produce a multiple regression model that predicted sales each month with an R squared of 0.87 (ie accounts for 87% of the variation) based on recent sales and recent expenditure on promotion alone. They reported that “marketing expenditures have a statistically significant, positive effect on sales: a 1% marketing increase results in 0.30% more sales (p =0.00). Of this sales rise, approximately 35% is compensated by sales decreases of competing products, the remaining 65% is due to market growth.”  However they did find evidence of the law of diminishing returns.  Because their study was observational the possibility that the causation relationship may not be as simple as they suggest should be considered.  The economists also concluded that pharmaceutical promotion also leads doctors to be “less sensitive to prices when deciding which pharmaceutical he [or she] should prescribe. This is unambiguously bad for welfare.” [12]

Risk factors for suboptimal prescribing

Observational evidence suggests that two of the many independent risk factors that account for variability in prescribing are:

1)    Exposure to pharmaceutical promotion. (Behaviours such as allowing more visits by drug company representatives, attending more meetings arranged by drug companies, reading more publications funded by drug companies rather than reading high quality medical journals and independent drug bulletins.)

2)    Positive attitudes towards pharmaceutical promotion. (Beliefs such as “Drug company representatives are a useful source of information”, “Pharmaceutical advertising rarely includes misleading claims”, “Information from drug companies helps me to improve my prescribing”.)

Eleven relevant studies from other countries that Prof Joel Lexchin and I am aware of have found that higher levels of exposure and more positive attitudes towards pharmaceutical promotion are both associated with lower quality and/or economy of actual prescribing or prescribing surrogates (eg choice of drug in response to a case vignette). [13-23]  A 3 page tabular summary of these studies written by Joel is available on request.  The best of these studies by Berings et al (1994) studied 128 GPs in Flanders.  They found that the number of visits from commercial representatives and the estimated utility of commercial information plus year of graduation accounted for 26% of the variation in the tendency to prescribe benzodiazepines. [22]  Responses to National Prescribing Service Case Study Number 5 suggest that these findings also apply to Australian general practitioners except in the former study older doctors prescribed more benzodiazepines but in the NPS data older doctors were less positive about inappropriate new drugs.  The NPS data suggests that important risk factors for unjustified beliefs about the appropriateness of new drugs included in order of significance:

1)    More recent graduation.

2)    Stronger belief that “My prescribing is improved by information from drug companies.”

3)    Belief that a lower % of drug advertisements contain potentially misleading claims.

4)    More frequent attendance at drug company sponsored meetings.[3]

A useful analogy

The current state of the art is analogous to that point in medical history when doctors had observed a relationship between pulse pressure and early heart attacks.  The corresponding concepts are described in table 1.

Table 1 Using the analogy of early heart attack as a way to think about suboptimal prescribing.

The aim of the proposed research program is the equivalent of describing some of the mechanisms of hypertension, inventing a sphygmomanometer, and a thiazide and then showing that it works to reduce the rate of early heart attacks.  Positive side effects of the thiazide may include reduced “smoking” and improved “diet and physical activity levels” but I am less optimistic of achieving impact via such changes.

View/Hide References

HS Int News index

Page views since 15 March 2010: 5368

Comments

Our members can see and make comments on this page.