SERUM DIAGNOSIS OF TYPHOID FEVER 387 



by the course observed clinically, and b\\ post-mortem examinations 

 of fatal cases confirmed the result of the test.^ The true value of 

 Widal's test can indeed only be estimated by comparing in this 

 way the number of cases in which the test and clinical observations 

 at an earh^ stage prove to be correct, by the subsequent course and 

 termination of the disease. Negative reactions before the seventh 

 day are not always reliable, and this, in some degree, limits the 

 utility of the test. But Delepine has maintained that of 6000 

 cases diagnosed in his laboratory 97 to 98 per cent, of the tests 

 proved correct. In certain sets of cases it rose to 99 per cent.^ 

 But it may be said that in general it is now conceded that from 95 

 to 97 per cent, of all cases of typhoid fever give the reaction at 

 some time oi* other in the course of the disease.^ The improve- 

 ments recently effected in the diagnosis test have served to make 

 it more accurate, and therefore more reliable. Its application 

 to individual cases of typhoid fever is now a common practice, and 

 in most well equipped Public Health Departments in this country' 

 arrangements are made for prompt serum diagnosis in doubtful 

 cases of enteric. But we may go even further and say that the 

 applicability of the Widal reaction in milk-borne epidemics has 

 been proved. One of the best known instances in which this 

 diagnostic test was used was in the milk-borne outbreak of enteric 

 fever in Clifton in 1897. So graphic and explicit is the state- 

 ment of Dr Davies, the Medical Officer of Health, who applied 

 the test in this outbreak, that we may quote his own words : — * 



"Between the dates 19th and 21st October 1897 I found 

 myself in this position : I knew definitely of five cases of undoubted 

 typhoid, all apparently in the second week, if not later, of their 

 illness ; at the same time I had heard of a very wide prevalence of 

 reputed ' Influenza,' the distribution of which, as it were, in 'selected' 

 houses, struck me as but little characteristic of this disease. For 

 instance, one college house was seriously affected, whereas adjacent 

 houses of the same class on each side were absolutely untouched. 



" I was also aware of the indeterminate nature of the onset of 

 ' typhoid ' fever, in which, as Newsholme has shown, nearly 8c 

 per cent, of the cases are not notified until the second week or 

 later. 



"In addition, the habitual exemption of Clifton from prevalence 



^ Jour. Sanitary Imtitute, 1898, No. i. 

 ^ Brit. Med. Jour., 1901, vol. i., p. 832. 

 ^ Ibid., 1901, vol. ii., p. 596. 



* Trans, cf the Epidemiological Society of London, vol. xvii., 1898, 

 pp. 97-101. 



I 



