Dr. Dublin

In res Dr, Armstrong's Suggestions for the investigation
of Heart Disease in Relation to Predisposing Causes

 

The topic might be discussed in either one of the
two following ways:

1. By some attempt at numerical correlation measure-
ments.

26 By general descriptive discussion.

Discussion of (1). The method (1) has been tested and it has been recommend-
ed that no reliance be placed on correlation coefficients obtained by comparing
time series of heart disease and certain other diseases, with an arbitrary lag
introduced by trial in such manner as to produce the highest possible correle-
tion coefficient. It does not seem desirable to reopen this question, the chief
objection to this method being that the sequelae of such diseases as rheumatism,
whooping cough, etc. are so spread over a series of years after the prime cause
that no correlation study can be expected to effectively detect the true cause
and effect relation which undoubtedly exists; any correlation coefficients

that may be found by this method are very apt to be due to something quite dif- ©
ferent than the relation that is really being investigated.

Discussion of (2) We have good material on which to base a discriptive
discussion, gathered together in our work sheets for the book on mortality.

A really good descriptive discussion is valuable even though it may not involve
correlation or similar technique. Dr. Armstrong, in his memorandum of August
i3th expressed interest particularly in the trend of heart disease fatalities
at ages 30 to 60. As the work sheets now stand, we cannot give data on precise-
ly this age group but we can give data at 35 to 65. The sheets before me cover
ae Organic heart disease; b. Endocarditis and valvular heart disease;

ce Chronic myocarditis; andd. Angina pectoris. :

Among the items enumerated above the case of chronic
endocarditis and valvular heart disease is peculiar in showing for the period
1921 to 1930 a marked decrease at ages 35 and over, whereas all the other causes
listed show increases, and in some cases very great increases.

The effect of the influenza epidemic seems to be clear.
ly visible in all those series which we have been able to carry from the year
1911 to 1929, There is a sudden drop in deathrates for the year 1919 as com.
pared with the year 1918, followed by a tendency to return to the old level or
to exceed it, except in the case of chronic endocarditis and valvular heart
disease.

 
a £2

: As to the relation between heart disease and certain
predisposing causes, these sheets, of course, give no clue. A discussion of
this topic can be conducted along the lines of information gathered from
general experience. This, it is true, would not in itself bring forward any-
thing very novel, but it would quite properly form part of the discussion
of information contained on the work shects relating to heart disease.

: If it is desired to have data expressly for the group 30
to 60 years, mentioned by Dr- Armstrong in his memorandum of August 13th
some additional computation would have to be made.

"Alfred ’3."Lotka’*’

Sept, 25,1931