15 Jul 2020
ASCOT blog: A COVID-19 effect on scientific research
Written by Professor Jason Roberts, Chief Investigator, AustralaSian COVID-19 Clinical Trial and Sue Anderson, Consumer Representative, AustralaSian COVID-19 Clinical Trial Steering Committee.
‘The magnitude of this pandemic, which has touched virtually everyone in the world, clearly deserves a commensurate evaluation; an honest evaluation.’
- WHO Director-General, Tedros Adhanom Ghebreyesus.
The world, as we alive now know it, is in uncharted territory. COVID-19 has spread inexorably to every part of our planet. Governments, leaders, and health care policy professionals are engaging with health care providers, business and social experts to curb the march of fatality and disease. Economic and broader threats to communities are emerging daily. To face down this menace, all have responded with unparalleled speed.
The rapidity with which all have mobilised is obviously laudable. Everybody who can, everybody who has any role to play, have rallied with a grim determination and unwavering commitment to minimise the spread of COVID-19, save, and minimise the damage to lives, and livelihoods.
Clinicians at the coalface of health provision have formed hundreds of groups globally to study drug therapies and devised treatment strategies to both improve outcomes for patients and stem the spread of the virus.
The results of these COVID-19 studies are subjected to the most robust analysis possible by team members. However, in our current scenario, time is of the essence, and this has implications.
The effect on trials by the early release of data
Unlike the rest of evidence-based medicine, and because there is so little existent evidence and expertise from which effective treatment guidelines can be developed, the COVID-19 bar for evidence is lower than normal.
Time pressure, combined with the imperative to share knowledge which might positively inform and save lives, has led to a new norm of hasty interpretation of traditional publication of study findings. These findings are being communicated in the form of pre-publication and even media release.
Media releases with the barest of data from large randomised clinical trials (RCTs), albeit from highly respected groups, are now influencing treatment guidelines globally. This highlights the current desperation for prospective data where bias is minimised, leading to interpretations of highly flawed, ‘hypothesis-generating’ retrospective observational studies as ‘hypothesis-testing.’
This situation has been encapsulated perfectly by the case study of hydroxychloroquine, a drug with strong in vitro activity against SARS-CoV-2 (1). On the basis of preliminary non-clinical data, hundreds, if not thousands of clinicians and triallists rushed to incorporate hydroxychloroquine into treatment guidelines as a standard therapy, or devised large and expensive RCTs on the basis of the hope provided by the pre-clinical studies.
Overseas health care staff were reportedly using hydroxychloroquine as a prophylaxis agent whilst they were caring for COVID-19 patients. There were suggestions that Australians were acquiring hydroxychloroquine for themselves, their family and friends in the event the pandemic situation deteriorated further (2). Subsequent media excitement reached a crescendo with US President Trump stating his belief in its potential as a treatment and prophylaxis agent. The broad response to Trump’s support for hydroxychloroquine was resoundingly negative.
Further low-quality studies (high risk of bias) ensued, including one that was first published and then retracted, from The Lancet. This study, surprisingly, indicated that hydroxychloroquine was associated with a massive increase in mortality rate (3). Later low-quality studies (high risk of bias) showed completely the opposite effect (4). These inform the latest published data which now supports further prospective RCT use of hydroxychloroquine. However, these RCTs are unlikely to occur. The equipoise for clinicians with hydroxychloroquine appears unlikely, particularly given the scant media release data from RECOVERY (5) and published findings of pharmaceutical-industry sponsored studies with remdesivir (6).
Clinicians await full results of the large RCTs that have tested hydroxychloroquine, RECOVERY and SOLIDARITY. They want real data to be published and to be externally scrutinised, above what has been done by the data safety monitoring boards. It’s noted that both studies have stopped randomising patients to hydroxychloroquine for lack of efficacy (as opposed to apparent toxicity).
But, equally important in the current COVID-19 media-saturated atmosphere, is the access that prospective patients have to negative articles regarding hydroxychloroquine, decreasing the likelihood they will consent to trial participation.
Further to issues with hydroxychloroquine, a secondary and recent perpetuation of clinician uncertainty centres on lopinavir/ritonavir, seemingly removed as an ongoing treatment intervention in RECOVERY, with scant data made available from brief media releases (5).
The challenges faced and implications
The global scenario presents major challenges for clinicians, clinical trialists and consumers in the community.
- For clinicians, can proposed treatments be altered based only upon sparse data issued in a media release? Or changed according to a colleague’s feedback of what they have communicated with investigators in a large RCT?
- For triallists, should the treatments in current or future planned studies be modified without knowing the limitations of media release data?
- What too of patient characteristics, and the rapid-effect implications of policy and healthcare decisions made in the midst of a pandemic?
- For consumers, how should they interpret the inconsistent messages emerging in all forms of media? What impact does the media have on patient willingness to participate in a trial? And how might uninformed or erroneous media releases affecting consumer choice to participate in trials be empathically addressed by clinicians?
There are clear implications for RCTs from present-day challenges. Obviously, studies cannot work without their potential benefits being clearly conveyed to, understood and agreed upon, by a wide patient populace.
And, to ensure the best outcomes for patients, more good trials are needed to provide certainty of best treatments.
Good, strong trials, by nature, are an in-depth, rigorous exploration of a hypothesis. They are exhaustive. Those who work on them demonstrate a doggedness of personality, a serious resolve to see something through to its end, to see it to fruition, to craft objective outcomes concisely charted and resolved according to statistical probability.
Yet, while the work of such clinicians has never been more needed, it is perhaps, now, never been more compromised.
What’s to be done?
Current circumstances require an entirely new level of trial flexibility, of ability to pivot and adapt are essential, which can only be achieved through highly sophisticated, highly adaptive, trial design.
Trial design must still be ‘stress tested’ however, to assess for strengths and weaknesses of hypothesis and delivery at every pressure point.
Impediments to study must be addressed clearly and without bias, identifying real knowledge gaps and strategically filling the void around small amounts of knowledge currently available. All new, and newly informed knowledge must be communicated quickly, clearly and concisely to clinicians, trialists and consumers alike.
COVID-19 media coverage will continue and continue to impact. RCT design and delivery. Consumer knowledge of COVID-19 affiliated drugs and treatments will remain at levels previously unknown for studies. While not necessarily a negative, it does mean that consumer engagement, and population sampling both at trial outset and throughout, is critical to determine initial and ongoing patient willingness to engage in any trial. Real-time knowledge of consumer numbers likely to consent to trial participation as a population percentage, then necessarily informs the patient numbers required for trial success.
For, with all the best clinician intentions in providing the best levels of care, if patients and non-trial clinicians ‘aren’t buying into what’s being sold,’ if their values, beliefs and knowledge set don’t consistently align with the proposed outcomes of a trial, no trial can possibly succeed.
- Yao X et al, In Vitro Antiviral Activity and Projection of Optimized Dosing Design of Hydroxychloroquine for the Treatment of Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2). Clin Infect Dis. 2020 doi: 10.1093/cid/ciaa237
- Coombes I et al, Principles of ethical prescribing for self and others: hydroxychloroquine in the COVID‑19 pandemic. Aust Prescriber 2020; 43: 76-7
- Mehra M et al, RETRACTED: Hydroxychloroquine or chloroquine with or without a macrolide for treatment of COVID-19: a multinational registry analysis. Lancet 2020; org/10.1016/S0140-6736(20)31180-6
- Arshad S, et al Treatment with Hydroxychloroquine, Azithromycin, and Combination in Patients Hospitalized with COVID-19. Int J Infect Dis 2020; org/10.1016/j.ijid.2020.06.099
- Press release available at https://www.recoverytrial.net/results
- Beigel J et al, Remdesivir for the Treatment of Covid-19 - Preliminary Report. NEJM 2020; doi: 10.1056/NEJMoa2007764