By Robin Blythe and Nick Graves
In mid-2017, an American physician, Dr Paul Marik, released the results of a single-site, observational study on vitamin C in sepsis care. The cocktail of IV vitamin C, thiamine, and hydrocortisone, three cheap and low-risk ingredients, produced spectacular results, reducing mortality in the treatment group of over 30% and significantly reducing rates of acute kidney injury.
As with any vitamin C miracle cure, readers were sceptical. Drastically reduced mortality and morbidity for the low, low cost of $499.99 – ask your Dean of medicine today! Images of gleaming-moustachioed snake oil salesmen come to mind, promising the undeliverable. Issues such as sample size and selection bias were seen as indicators that we must wait for better evidence – namely, multi-centre randomised control trials on a large scale. This is the default position in medicine, usually for good reason, as it is better to do nothing than cause harm to patients. These secondary studies are now underway, but they present a significant opportunity cost: do we wait a year (or more) for results while passing up a treatment that could potentially save 30% of sepsis patients, of which there are over 1.5m annually in the USA alone?
This is a difficult scenario for most hospital administrators. On the one hand, it would be bad practice to jump at every new result, and history suggests the bundle may be another wild goose chase. On the other, with no documented risks from treatment and for around $500/patient, we could potentially save hundreds of thousands of lives.
We argued in a recent paper (link: https://f1000research.com/articles/7-500/v1) that compared to waiting for better evidence, we have enough information to fund the vitamin C bundle on a trial basis. We presented a conservative scenario in which the bundle was substantially less effective than purported compared to standard care. We found despite our conservative estimates, we were potentially passing up saving a significant number of lives and over $32,000 per patient in savings. In a single hospital in the US, the bundle would only have to work in 2% of patients to represent a cost saving to the organisation.
Miracle cures pop up in medicine fairly often, and very few stand up to scrutiny. However, there is no reason that in low-risk, low-cost situations, hospitals could not fund interventions on a trial basis. This would contribute to the available evidence and potentially save dollars and lives. We believe this method of modelling offers a viable alternative to clinical activity in situations of uncertainty.
