Ebola Vaccine Trial Begins In Congo As Cases Surge

The official account says infections are accelerating and a vaccine trial is beginning to meet the emergency. The data - such as it has been given to us - says only that both statements were issued within the same twenty days, by the same institution, the World Health Organization, with no case count attached to either. One of these is a fact. The other is an inference dressed as one, and I should like to see the register before I applaud the intervention.

Let us take the claim as reported. Ebola is spreading faster in the Democratic Republic of Congo than it was; a trial vaccine is now to be deployed against it. This is the sequence a committee wishes to hear - threat rising, remedy arriving, the machinery of relief visibly turning. I do not dispute that such sequences sometimes occur honestly. I dispute that this one has yet earned the word “accelerating,” because acceleration is a rate, and a rate requires two points in time and a count at each. I have been given one window - twenty days - and told the trend within it is upward. Upward from what baseline, counted by whom, using what case definition? A suspected case in one district register is not always a suspected case in another; I learned that lesson counting fever deaths at Scutari, where “diarrhoea” meant something different on every ward’s slate, and the sum of the wards meant nothing at all until the definitions were forced into agreement.

Here is the deeper irregularity, and it is the one that ought to trouble a statistician more than any missing case count. The vaccine trial is being reported as the response to the acceleration. But a trial, by its nature, is not a response to an emergency - it is a question being asked of one. A response has a target enrollment, a control arm or its ethical equivalent, an endpoint defined before the first dose is given. If the WHO trial has these features, the announcement has not told us so, and their absence is not a small omission. It is the difference between “we are testing whether this vaccine works” and “we are administering something so that a rising number can be met with a visible gesture.” The dying man does not care which of these is true. The policy maker deciding whether to fund the next twenty days of this operation very much should.

Consider the alternative reading that a fair-minded observer would raise against me: that in an active outbreak, waiting for a clean baseline before acting is itself a form of negligence dressed as rigor, and that WHO field teams have limited days and limited staff to produce the tables I am demanding. I grant the force of this. I did not wait for a perfect ledger before pulling filthy linen out of the Barrack Hospital. But I did not stop counting once I started acting, and that is the distinction I am pressing here. Speed of response and honesty about the denominator are not in tension. They are both disciplines of the same clock. An organization capable of mobilizing a trial in the Congo within twenty days is entirely capable of publishing, alongside it, the daily case tally by district that would let the rest of us judge whether “accelerating” means a doubling or a rounding error.

What the data before me actually supports is this: an outbreak is underway, its trajectory is asserted rather than shown, and a trial vaccine is being introduced into the middle of that asserted trajectory rather than alongside a documented one. I would want the enrollment figures, the district-level case curve for the full twenty days, and the endpoint the trial is measuring, set beside each other in a single table before I permitted myself either alarm or relief. Until that table exists, what we have is not evidence of acceleration. It is a vaccine trial beginning in the dark, and a press release standing in for the lamp.