Evidence Live—Dangerous ideas for the future of evidence based healthcare

evidence_2015Evidence Live 2015 is underway.

Once again there is an international line up of speakers to stimulate thought debate and action. We would like you to consider throughout the conference dangerous ideas for the future of evidence based healthcare. Closing the gap between evidence and clinical practice remains a weighty issue to solve. To improve on the current status quo we need new, radical, and innovative ideas. These important ideas would ensure practitioners are equipped with the totality of evidence and the right tools to fully inform patients about the benefits and harms of effective, or in many cases, ineffective interventions. We are calling these ideas dangerous not because they are likely to cause problems, but because they can often be true and therefore might provide solutions and shorten the gap of translating medical research into clinical practice

Here are some dangerous ideas. We will be adding to these over the next couple of days. 

Dangerous Innovation: The opportunity for all members of the public to plan and participate in health research by 2020—Amy Price and Amanda Burls

Evidence informed health includes patients who successfully self-manage complex healthcare. ThinkWell invites the public to self-manage research involvement. We are committed to helping people mould research agendas, contribute to study design, and participate in health research.

PLOT-IT (Public Led Online Trials—Infrastructure and Tools) is a ThinkWell project to help the public prioritise, initiate, design, and participate in methodologically robust randomised trials investigating research questions of interest to them.

PLOT-IT represents disruptive innovation that turns traditional research methods on their head with the capacity to generate new relevant evidence to inform people’s health choices. PLOT-IT more than simply involves patients, we set out to put the public at the very heart of health research and challenge the barriers that prevent people from taking control of their own health and research choices.

By helping the public and patients initiate and participate in health research ThinkWell contributes to generating knowledge, starts to redress the imbalance in the research agenda and empowers participants to gain confidence and competence for making informed research decisions.

Curious or want to collaborate? Learn more at http://ithinkwell.org

EBHC studies that check for NCD harm from vaccines or studies that check for beneficial effects from infections are non-existent—Wouter Havinga

Vaccines rule and cohorts on non-communicable diseases (NCD) harm from vaccines have not been done in the developed world. Similarly, NCD data are not systematically collected in connection with vaccines schedules in developing countries.

However, is the burden of disease due to “chronic” (life long) NCDs such as asthma, allergies, IBD, ADHD, ASD, etc, in the developed world now greater than that from the prevented “acute” (transitory) infectious diseases? Vaccine effectiveness studies have been confounded by good hygiene, nutrition, and less over crowding.

Critical research is needed to investigate whether infectious diseases actually prevent NCDs.

Comparative studies of vaccinated and unvaccinated cohorts are needed.

Vaccines affect the microbiome which in turn affects immune responses. How have vaccines changed the commensal flora and NCD prevalence?

Urgent research is now needed in developing countries that are currently targeted by the vaccine industry. Without systematic cohort studies that check for NCD harm from vaccines, important data will get lost.

This omission might well create an even bigger market for the pharmaceutical industry following vaccinations, for treatments of chronic conditions.

Dangerous idea—Carl Heneghan

Publication of all trial documents in full whilst the trial is in process.

We have observed considerable problems with the reporting of clinical trials. In an era where comparative effectiveness relies on the validity of trials and the full disclosure of the benefits and harms of a drug there is an urgent need for transparency.

Therefore my idea is once a trial has received ethical approval the full protocol would be published, and as the trial progresses the statistical analysis plan and protocol amendments would also be published prior to data analysis. A final report would then be published along with the data files. Thus providing a complete transparent record of the study in one place for future researchers to access and analyse.

Dangerous idea—Ben Goldacre

Treatments are currently licensedand even recommended by NICEon the basis of small, brief trials, in ideal patients, often measuring surrogate outcomes. Instead, where there is outstanding uncertainty about effectiveness, all newly licensed treatments should only be available on the NHS in the context of a pragmatic, open-label, randomised trial; compared against the current standard of care; with real world outcomes measured at very low cost from routinely collected electronic health record data, and patient reported outcomes online; until such time as there is clear evidence of superiority, or at least non-inferiority with a preferential side effect profile. Similarly, wherever there is uncertainty about which is the best of two commonly prescribed treatments, there should be routine randomisation and follow-up. Through this, we can turn the NHS into a machine that constantly tests, learns, adapts, and optimises care, creating a gift of evidence to the world.

EBM must reform or be irrelevant to mainstream medicine—Huw Llewelyn

EBM focuses mainly on the role of RCTs in establishing provisional treatment efficacy and also restricts itself to an idiosyncratic form of diagnosis based on Bayes rule by multiplying prior odds by likelihood ratios. On the contrary, transparent diagnostic reasoning usually involves formulating differential diagnoses and working through them. This cannot happen in an evidence-based way without clear treatment indication criteria based on disease severity, which are combined under an umbrella term to create evidence-based diagnostic criteria. Furthermore, the most important “evidence” is the group of symptoms, signs, and test results obtained from the individual patient to make each prediction. This patient’s “evidence” is called the “particular” evidence in logic and has to be matched with “general” evidence from research on many patients. It is also important to combine two types of thought: non-transparent which is quick and used by most experienced doctors, and transparent reasoning, which is how diagnoses and decisions are explained to patients, relatives, doctors and other health professionals using evidence. Unless existing EBM reforms itself, then it will continue to be called “rubbish EBM” and this will undermine progression to “real” EBM. This approach is already taken by the Oxford Handbook of Clinical Diagnosis.

The n-of-1 deep analysis take off—Pierre La Rochelle

A lot of day to day clinical decisions are taken on the base of simple assessment by the clinician and his patient where the causality acuity is weak. The n-of-1 trails have a high accuracy in causality assessment.
Important clinical decisions are based on this poor quality judgment in spite of high stakes: different treatment costs or requirement, no real alternative, future potential risks, etc.

The n-of-1 trials are a promising tool for clinicians but very few use them mainly because their impact use have not been well studied compared to usual care.

My idea would be to study the use of this trial design to a very large scale to get powerful data: to determine the magnitude effect and to permit some sub-group analysis.

Run-in phase of large RCTs could become a large laboratory of n-of-1 trials: placebo and ethic acceptation could be more easily acquired, short term outcomes, and long term outcomes could be studied simultaneously or independently.

Run-in phases of large RCTs should be offered to searchers as n-of-1 trial laboratory testing. At term, n-of-1 trials could become a requirement to take important decisions in specific high stakes situations.