Analyzing Ebola – Is it spreading at exponential rate?

By Gregory Piatetsky, @kdnuggets, Oct 4, 2014.

Ebola is a major global health threat in 2014, but how quickly does it spread? Various news organizations report that the number of cases is increasing exponentially, but there were only a few thousand cases (not millions like for flu), so I decided to investigate from primary data.

I found reports from

• WHO (World Health Organization), www.who.int/csr/disease/ebola/situation-reports/en/
• and CDC www.cdc.gov/vhf/ebola/outbreaks/2014-west-africa/index.html

 
and extracted data for the 3 most affected countries - Guinea, Liberia, and Sierra Leone.

Of course, the reported numbers should be considered only a lower approximation, since reporting on Ebola is very difficult. The heroic health care workers in West Africa, who collect the data on the ground, face not only the disease (many have already died) but also attacks from scared villagers who think the health care workers spread Ebola. CDC estimates that for every case reported in publicly available case counts, an additional 1.5 cases are not recorded.

More advanced and geographic analysis is possible, as well as looking at lagged time series (since Ebola typically takes about 3 weeks to fully manifest itself), but even our initial analysis is quite illuminating.

First, we see that the disease progression is markedly different in the 3 affected countries.



Fig 1. Total Ebola cases in West Africa, as of Oct 1, 2014

Ebola is clearly spreading the fastest in Liberia, where the number of cases has doubled in about last 25 days. In Sierra Leone, it took about 28 days.

In Guinea, where according to WHO, Ebola started in Dec 2013, we see the earliest cases. However, it seems that the growth rate is slowest there. In fact there was no growth in reported cases in Guinea during July 2-20, with the number of cases stuck at ~ 410, although perhaps this was due to faulty reporting. The growth in reported Ebola cases has resumed in late July, but slower than in the other 2 countries. Still the number of cases has doubled in the last 43 days.

Next graph shows the total number of deaths. Liberia is much harder hit than others.



Fig 2. Total Ebola deaths in West Africa, as of Oct 1, 2014

Finally, we look at the death rate, which again is markedly different - highest in Guinea, lowest in Sierra Leone. One hopeful sign is that % of death in Sierra Leone seems to be decreasing.



Fig 3. Total Ebola deaths in West Africa, as of Oct 1, 2014

Here is the data in the table form.

Date	Guinea Cases	Guinea deaths	Liberia cases	Liberia deaths	Sierra Leone cases	Sierra Leone deaths
1-Oct	1199	739	3834	2069	2437	623
28-Sep	1157	710	3696	1998	2304	622
23-Sep	1074	648	3548	1830	2021	605
21-Sep	1022	635	3280	1677	1940	597
19-Sep	1008	632	3022	1578	1813	593
14-Sep	942	601	2710	1459	1673	562
13-Sep	936	595	2407	1296	1620	562
7-Sep	861	557	2081	1137	1424	524
5-Sep	812	517	1871	1089	1261	491
25-Aug	648	430	1378	694	1026	422
20-Aug	607	406	1082	624	910	392
18-Aug	579	396	972	576	907	374
16-Aug	543	394	834	466	848	365
13-Aug	519	380	786	413	810	348
11-Aug	510	377	670	355	783	334
9-Aug	506	373	599	323	730	315
6-Aug	495	367	554	294	717	298
1-Aug	485	358	468	255	646	273
30-Jul	472	346	391	227	574	252
27-Jul	460	339	329	156	533	233
23-Jul	427	319	249	129	525	224
20-Jul	415	314	224	127	454	219
17-Jul	410	310	196	116	442	206
12-Jul	406	304	172	105	386	192
8-Jul	409	309	142	88	337	142
6-Jul	408	307	131	84	305	127
2-Jul	412	305	115	75	252	101
30-Jun	413	303	107	65	239	99
20-Jun	390	270	51	34	158	69
16-Jun	398	264	33	24	97	49
10-Jun	376	241	13	11	117	19
8-Jun	372	236			89	7
3-Jun	344	215			81	7
30-May	291	193			50	6
27-May	281	186			14	5
21-May	258	174
10-May	233	157
3-May	231	155
29-Apr	221	146
20-Apr	208	136
16-Apr	197	122
10-Apr	157	101
5-Apr	151	95
1-Apr	127	83
31-Mar	122	80
30-Mar	112	70
27-Mar	103	66
25-Mar	86	60
24-Mar	86	59