‘Dopesick’; why a critical view of the literature may save your patient’s life

ABSTRACT Introduction Teaching critical literature appraisal is challenging. Providing a compelling clinical context using ‘cinemeduation’ stimulates interest in the topic. Methods After watching the first episode of the mini-series ‘Dopesick’, where the scope, timeline and extent of the problem of opioid abuse are clearly seen, abstracts of the period literature strongly supporting the use of Oxycontin for use in chronic pain are shown. All were published in highly ranked medical journals. A simple paper evaluation structure is suggested. It is PPICOREAD which stands for the following questions: Who Paid for the study? What was the Population studied? What was the Intervention given in the study? What was the Control group used? What was the Outcome and was it of clinical significance to you? Was the trial Registered? Was there an Educational element for you? Was there anything Applicable to your practice? What was the Duration of the trial? Is this duration sufficient to reassure you that the trial is relevant to your practice? Results The very poor quality and dishonest nature of the conclusions of these papers are quickly and easily uncovered in a supervised workshop. The causes for these clear discrepancies are discussed. The devastating consequences are described.


Introduction
Teaching and encouraging critical appraisal of medical literature by using evidence-based medicine (EBM) tools is difficult.Several learner-based EBM challenges have been identified including suboptimal role models, students' lack of willingness to admit uncertainty, a lack of clinical context, and students' difficulty mastering EBM skills [1].Other challenges are lack of time for teaching the topic and lack of requirement for EBM skills in healthcare management [2].Many physicians have suboptimal knowledge of EBM and low levels of practising it [3].For medical specialists in training (or residents), specific barriers towards implementing EBM were a lack of interest by the staff or even staff disapproval of EBM [4].For general practice (GP) trainees, barriers identified to the use of EBM were lack of time, knowledge and skills, negative attitude, and external factors including poor trainer motivation and lack of formal training in EBM skills [5].Hence the need for novel approaches to teaching and stimulating interest in this important topic.Creating explicit learning moments in which EBM decision making is discussed, can improve shared knowledge and can also be useful to unveil tacit knowledge derived from mindlines [6].Mindlines are the unstated internal cognitive processes that doctors use when coming to a decision on how to manage patients [7] and have the potential to expand EBM's conceptual toolkit to produce richer forms of 'evidence-based' knowledge [8].When developing mindlines, the skill to critically appraise research articles and guidelines is considered necessary for the successful transformation of clinical knowledge [9].Also, those using mindlines recognise the need to take account of human and social contexts in their approach to evidence [10].
The Oxycontin story provides a compelling clinical context where profit-driven dishonest manipulation of the literature by unscrupulous parties clearly happened and resulted in widespread misery and loss of life.Simple evaluation of the supporting literature by healthcare professionals at the time that the drug was being promoted would have clearly shown the poor design and totally inappropriate conclusions of the studies.This would have saved many lives.Using movies has been accepted worldwide as a tool to help students learn medical professionalism and critical thinking and is called 'Cinemeducation' [11].I describe the use of the first episode of the mini-series 'Dopesick ' [12] to stimulate interest in and some mastery of the basic skills of EBM and critical literature appraisal in practice.

Methods
Episode 1 of the mini-series is shown to the class.Students are then asked what stands out for them.
Feedback from the class usually includes the following: • They are impressed by the caring attitude and work ethic of Dr Samuel Finnix the family physician.• Many also believe that as this is a dramatisation, it must not be completely true.
It is appropriate at this point to highlight some important parts of the film, illustrated here by quotations.These are summarised in slide 5 of supplement 1 (a PDF of the accompanying PowerPoint presentation which can be found online).

Consequences in reality
The students are then shown a series of slides summarising the serious consequences of overuse of Oxycontin.These are slides 6 to 14 of supplement 1. References are given in slides 38 to 40 of supplement 1.These include the US General Accounting Office report to Congress [14], an editorial from the American Journal of Public Health entitled 'The Promotion and Marketing of Oxycontin: Commercial Triumph, Public health Tragedy' [15], a report from the LA Times 2016 called 'You want a description of hell?Oxycontin's 12 hour problem' [16], American Medical Association Journal of Ethics opinion piece called 'How FDA failures contributed to the Opioid Crisis' [17], Lancet editorial entitled 'Opioid overdose crisis: time for a radical rethink' [18], the alarming increase in drug overdose deaths since 1999 [19], the fact that the US Food and Drug Administration (FDA) were forced to explain the timeline of 'selected FDA activities and significant events addressing substance use and overdose' in the context of approving Oxycontin for use [20] and finally news that the Sackler family paid $6 billion to settle Purdue opioid lawsuits [21].

Medical literature
Now the class is convinced that the film is a true representation of the facts.So, the next task is to look at what literature was being used by Purdue Pharma at the time to promote Oxycontin.There are six papers published between 1980 and 2000, with a systematic review in 2003.
Details of these are summarised in supplement 1 (slide 41; Oxycontin literature references).
Slides 16 and 17 in supplement 1 show the journals these papers were published in along with impact factors (IF).For comparison purposes, the IF of some commonly read journals (slide 17) are shown in red.It becomes rapidly clear that just because the papers were published in 'highly ranked journals' does not automatically mean that their message can be accepted uncritically.

Abstract conclusions
Excerpts from the abstracts which could have been used by Purdue Pharma representatives to promote Oxycontin are as follows: • 'Addiction rare in patients treated with narcotics' (Porter and Jick 1980 [13;] New England Journal of Medicine; Impact Factor (IF) 176.1; research letter)

How to evaluate a paper rapidly: PPICO READ
Next the class is introduced to PPICOREAD to highlight the areas to consider when reading a paper critically.This is summarised in Box 1.
Payment: Arguably the most important element of a study is its funding source.If the sponsor is set to gain financially from a positive outcome of the study and affirmation of the product or drug tested, then considerable bias is established.This is most likely to manifest itself as a 'positive spin' on the results, talking up an otherwise lacklustre result.
Population: Was the population studied similar to the type of person you treat in your practice?If you are a GP for example, results of studies of a highly selected population attending a tertiary specialist service may not be applicable to your practice.
Intervention: Is this available and affordable to your patients?For example, physical activity or a simple drug like paracetamol may well be.However, a new expensive monoclonal antibody or biological agent that has to be given regularly by infusion may not be.What are the potential adverse effects of the new drug or intervention?Have they been fully listed and discussed?
Control: Was a control population used?If so, were they matched to the intervention population?Was allocation properly blinded or was there a risk of bias?Were dropouts accounted for in the intervention and control groups?
Outcome: Is this of relevance to you? Absolute outcomes such as death, stroke or myocardial infarction are good.Surrogate outcomes such as lowering of cholesterol or blood glucose are not as relevant as they may not necessarily make the patient feel better or live longer.Composite outcomes combine a number of possible outcomes together as the treatment is not good enough to have a significant effect on any one outcome.Was the outcome of clinical as well as statistical significance.For example a drop of 2 mm Hg in blood pressure in a large trial may be statistically significant but at an individual patient level such a small drop is not.Ideally the result should be expressed as absolute risk reduction (ARR) which will allow ready calculation of the number needed to treat (NNT) to achieve this result.Beware studies that only express their results as relative risk reduction (RRR) or reduction in odds ratio (OR) as these can be misleading BOX 1.How to evaluate a paper rapidly PPICOREAD P: Who paid for the study?Did they stand to gain from the study's findings?P: What was the population studied?Is it similar to your patients?I: What was the intervention used?Is this something simple you can promote such as taking more exercise or otherwise e.g. a drug that may have unwanted effects?C: What was the control used?Was the control population similar to the intervention population?Was blinding adequate?O: What was the outcome?Was this actual, composite or surrogate?Was the analysis Post-Hoc or pre-planned?Did the authors quantify the differences using Absolute Risk Reduction (ARR), Number Needed to Treat (NNT), and compare with Odds Ratio (OR) or Relative Risk Reduction (RRR) R: Was the trial registered (from 2005 approximately; https://www.who.int/clinical-trials-registry-platform)?E: What was the educational impact for you personally?Did you learn something new?A: How applicable are the findings in your practice?D: What was the duration of the trial?Does this reflect everyday practice?For example, the study duration is weeks only, but you would expect to use this drug for many months or years.
and do not allow ready calculation of NNT.Look out for any unusual statistical tests which are run to achieve this result.Another trick is 'post hoc' analysis where the data is gone over repeatedly with the intention of showing a positive result in a group which was not considered in the initial design of the study.Such outcomes should be considered with considerable scepticism.
Registered: From around 2005 it has become mandatory to register trials before they start.Several platforms exist one of which is run by the World health Organisation (WHO) (https://www.who.int/clinicaltrials-registry-platform).This should be clearly marked in the abstract.It helps to ensure that trials with negative results are not suppressed and left unpublished.This is also known as publication bias.
Education: Did you learn anything new that is of potential value to you?For example, regular physical activity makes people live longer.
Applicability: Can you apply this knowledge easily in your everyday practice?For example, advising patients on the benefits of physical activity.
Duration: What was the duration of the trial?Does this reflect everyday practice?For example, if the study duration is weeks only, but you would expect to use this drug for many months or years.
Using PPICOREAD it is obvious that: • Despite being of 'high quality' design and being published in prestigious journals with high impact factors, each study is of very poor quality.• This is probably due to them being funded and written by Purdue Pharma (except for the Porter and Jick [13] research letter in the New England Journal of Medicine).The latter is a letter of interest rather than a research letter.• Reading the abstract alone for each study is insufficient.The entire paper needs to be reviewed.• The PPICOREAD formula is a quick and easy way to critically analyse the papers.• Finally, the students are shown the systematic review published in 2003 by Chou et al [22] in the Journal of Pain Management.This is the only paper that was funded independently.It clearly shows the poor quality of the 5 earlier Oxycontin papers (Chou et al 2003 [22]).

Conclusion
This 'cinemeducation' lesson is a very effective way of convincing students of the importance of critically evaluating medical papers.It is fun and interactive.By engendering feelings of anger and subsequent empowerment, it stimulates interest in the topic.Students are given a tool -PPICOREAD -during the session that they can use to rapidly evaluate any paper.

•
This leads to the next point which is that an independently conducted systematic review is usually very helpful in evaluating the current literature on a topic.Slide 35 shows how pharmaceutical companies are scrambling to get a share of the lucrative global obesity drug market.Slide 36 shows how the pharmaceutical company Novo Nordisk is producing Liraglutide, a 'blockbuster' drug for obesity trade named 'Saxenda'.Recently they have been suspended as a member of the Association of the British Pharmaceutical Industry (ABPI) for two years because of 'serious breaches' of the organisation's code of practice.The company failed to disclose clearly its sponsorship of a weight management training course for healthcare professionals which included positive information on its obesity drug liraglutide.This is shown in slide 36.Slide 37 is from a recent article from the Drugs and Therapeutics Bulletin.It states that 'among 270 US physician authors of clinical guidelines, 74% received payments from industry, but only 37% disclosed financial conflicts of interest'.Therefore, ongoing vigilance and critical appraisal are essential.
Students then say that this all happened a long time ago -'surely things are different now?' Unfortunately, this does not seem to be the case.I illustrate this in slides 35-37.