Figure 1. The estimated number of people killed due to eaten BSE mad cows.
Örjan Hallberg
Hallberg Independent Research
www.hir.nu
e-mail: oerjan.hallberg@swipnet.se
pdf-version (Jan 2002, not updated); An early version was submitted to and rejected by The Lancet in June 6, 1997. It may be of interest to compare the early predictions with recent outcome.
Updated 2009-01-12
The life distribution of BSE infected cattle has been used in combination with an estimated percentage of cattle born in specific years since 1982 that were actually infected by BSE. The estimate was modeled to give a good fit to actual measured number of cattle that have died from BSE. The maximum infected level was around 7% of those born in 1987. The delay to BSE breakout gave a maximum around 1993. The human BSE peak is expected around 2000-2001 if the incubation time for man is similar to that of cattle..
The amount of slaughtered cattle as a function of their age was then used together with the life distribution of infected BSE cows to arrive at the total number of eaten ‘mad cows’. Obviously also a large number of just infected BSE cows have been eaten but for this analysis the correlation to dead people was optimized to these ‘half dead’ cows. This analysis showed that around 51000 seriously sick cows have been eaten in England over the years. A similar analysis to count all infected cattle ends up in 1.17 million eaten infected cows.
The life distribution for an infected person was initially assumed to be similar
to the cows in the absence of other information. Today the statistics might
indicate a delay compared to the life distribution of infected cows. The relation
between the number of eaten mad cows and human deaths was optimized (by varying
only one constant) to give the best fit to actual measured numbers in 1994-1996.
Initially (June 1996) it was estimated /1/ that 78 people were expected to die
until the year 2002 with a peak of 19 cases in 1998. In September 1996 a new
estimate arrived at total 160 dead people based on the numbers being reported
at that time. The actual outcome at the end of 2002 was 121 deaths. Figure 1
shows the estimate based on nvCJD reports up to January 8, 2002. This ends up
in 150 killed up to 2006. The reported nymbers are continuously updated in Figure
1.
Figure 1. The estimated number of people killed due to eaten BSE mad
cows.
There are some factors that may make great difference to the size of the human disease numbers predicted. For instance, should the incubation period for humans not be the same as for cattle but much longer then the number of human deaths would be much greater. Another factor to consider is the confidence levels involved. Several other methods have been used to predict variant CJD case numbers in the UK /2,3/ but these admit that we know neither the level of infectivity needed to infect a human, nor the time at which the current cases seen became infected. As a result they predict the levels being between close to zero and possibly several million. Since this model was defined in September 1996 the reported number of dead nvCJD-people has followed the initial prediction very well (Figure 1). If the incubation time for human cases is prolonged we will see larger numbers than earlier expected.
The calculations do not rely on any particular way of spreading the disease. If the age distribution of the nvCJD victims are normal one should suspect the food chain. If the cases predominantly are young people one might look for other ways like contaminated vaccines or possibly damaged brain blood barriers due to heavy EMF exposure. In that case more detailed analysis of raw data should be done.
Figure 2. The accumulated number of nv CJD is levelling off according to a normal distribution function..
The total number of nvCJD was calculated as a function of the median incubation time to best fit the outcome so far. Figure 3 gives the total number Vs. median incubation time. Figure 4 gives the accumulated number for the case the median incubation time is 7,4 years and sigma is 1,8 years.
Figure 3. Estimated number of victims Vs incubation time to fit current data.
Figure 4. The reported and calculated accumulated number of cases for a lognormal incubation time t0,1%=2,29 years and sigma = 0,18 decades. In a normal distribution the closest fit was median time 8,16 years and sigma 3,2 years. The best fit was achieved with the lognormal distribution..
Figure 5. The graph gives the annual total number of CJD cases in England according to Department of Health.
Figure 6. This graph gives the annual number of reported BSE-cases (on cows) in Europe outside England.
Figure 7. The history of BSE cases in England as it was modelled in 1996.
It is still too early to exactly determine the incubation time from this data but the total number of nvCJD will be low. If it is possible to determine the incubation time more precisely by other means, figure 3 could be used to estimate the total number. The outcome presented in Figures 1 and 5 indicates that the incubation time is close to the originally estimated value; 7-8 years. Note, the last few months numbers have been higher than in last year. Is a second wave coming? The first wave came exactly as predicted but is is something promoting a second and much larger wave? Keep an eye on this site to follow the numbers.
The increasing number of BSE cases in Europe should be cause for concern.
1. Ö Hallberg, e-mail communication to S Dealler, 1996-06-24
2. Dealler SF and Kent J, British Food Journal 1995; 97: 3-18
3. Dealler SF, House of Commons Select Committee on Agriculture and Health.
April 1996, Hansard, UK
4. P Oscarsson and Ö Hallberg, "EriView 2000 - A Tool
For The Analysis Of Field Statistics", Proc. ESREL 97, Lisbon, June 1997,
ISBN 0-08-042835-5.