The second Ebola virus disease for 2018 continues to unfold in the Democratic Republic of the Congo. Nearly two weeks have passed since the official announcement of the outbreak by the DRC Ministry of Health, and this is an interesting time as human, health, education, communication and countermeasure resources have all been deployed. In the second week, we’ve gained a more solid idea of the number of cases and how they are continuing to occur which in turn lets the experts gauge the scale and pace of the outbreak and its trajectory.

I thought I’d have a look at some of the early trends in the publicly reported numbers. They of course change frequently and the best interpretations come from those on the ground.

I have a theory that I previously called the control gap, may be useful to highlight differences between outbreaks but also how well controlled a particular outbreak might be.

Explaining the Control Gap: North Kivu Province and Ituri Province

If we subtract the confirmed case total from the suspected case total each report, what do we see? What if we subtract confirmed cases from probable cases? I’ve plotted both of those differences in the graph above (Figure 2).

If we have a lot of data, the control gap is easier to interpret when read over time and alongside the trends in suspected, probable and confirmed cases and deaths.

Suspect|Confirmed case difference

A suspect case is one that could be due to infection by the pathogen of interest, but for which sample testing hasn’t been completed yet. If testing cannot be completed, a suspect case is one that didn’t meet other criteria permitting its classification as a probable case.

If the control gap difference is large there could be a few reasons, but generally speaking, they all hint at a loss of control over the direction of the outbreak. When the difference is big we have a high number of suspected cases but relatively few confirmed cases. We might be looking at:

• samples not being collected in a timely fashion
• samples not being transported to the laboratory in a timely fashion
• the laboratory having trouble keeping up with the processing, testing, confirmation and reporting of results in a timely fashion

However, just because the difference is small doesn’t mean everything is under control.

The difference could be small when there is a high number of suspected cases and a high number of confirmed cases. This would still suggest an outbreak running out of control though.

So the best use of the control gap – at least over just a small number of reports – is to see the change from large positive gap to a near-zero or negative gap. Under these conditions, it’s likely a lot of samples have just been tested and either reported as positive or discarded because they were negative. This at least means the lab side of things is probably running well.

Suspect|Probable case difference

A probable case of a disease is one where testing could not be conducted for some reason, but exposure, travel, contact, signs and symptoms or other circumstances were such that disease was more than likely caused by the pathogen of interest.

This control gap may grow large, but for different reasons. These can include:

• cases dying without samples being collected
• testing does not yet exist or is failing to detect the pathogen 
• deaths are occurring amid an outbreak that is producing so many samples that the laboratory cannot keep up, leaving diagnoses to be made solely on clinical grounds

The Control Gap: Comparing Outbreaks

Another possible application of the control gap is to compare two outbreaks. There is a rare opportunity to do that within the same country with the back-to-back Ebola virus outbreaks in the DRC this year.

We can see from Figure 3 that there was a very different pattern compared to the current EVD outbreak.

The control gap was almost always negative in Équateur. This was, at least mathematically speaking because suspect EVD case numbers were lower than confirmed case numbers throughout almost all of the 2-month outbreak.

Only in that first week were there signs that things were out of control. You’d expect less control early on in an outbreak of any disease.

The latest outbreak in North Kivu and Ituri Provinces has had a sudden turnaround in the latest DRC MOH update. A lot of suspected cases were cleared up – 22 were dropped from the total; the biggest change in suspected case counts of this outbreak to date.

But before I’d say this change in the control gap from a positive to a negative value was a sign of actual control, I’d like to see many more days of reports.

Differences and similarities between two DRC EVD outbreaks

There are some differences and similarities between the two outbreaks that both interest and worry me:

1. When the Équateur outbreak was initially confirmed, there had been 21 total cases (including 17 fatal and 13 confirmed). When this latest outbreak was announced, there were 26 probable cases (20 fatal, 4 confirmed)
2. So far there have been 8 EVD cases in health care workers, compared to the 7 (2 fatal) for the Équateur outbreak.
3. Vaccinations started around day 11 in the Équateur outbreak and around day 8 in the latest outbreak
4. We still have 36 suspect cases outstanding at the 2-week mark. The Équateur outbreak never had more than 14 suspect cases after week 1. It went for approximately 54 days.
5. Today’s clean out of suspected cases didn’t dismiss them all as negative. There were 9 new confirmed cases; the equal biggest single-report rise in confirmed cases of the outbreak.   
6. Over the past two days, a second province has gone from having 2 probable deaths due to EVD to 4  EVD deaths with 7 confirmed EVD cases and 4 suspect cases. The Equateur Province outbreak was kept within a single province. 
7. The African Great Lakes region is a conflict zone with millions struggling to find reliable food sources (what food choices would you make in order to feed your family), millions of mobile internally displaced and refugee people, ongoing sexual assaults (bad at all but also for a virus that can transmit via sex) and cholera (a disease hard to differentiate from EVD early on) as well as other viral haemorrhagic fevers which place extra burdens on laboratory resources 

The coming days and weeks seem likely to have many stories to tell.