View report page

1956) it was about 50-54 per cent. From figures calculated over the winter of 1956-57,
a winter in which there was no undue 'external' influence of cold or epidemics, apart
from the fourth period smog, the proportion of deaths in the over 70's was remarkably
consistent month by month, beginning at 48 per cent, in October, rising to 53 per cent,
in November and remaining at around the 50-53 per cent, level for the remainder of the
winter. From Figure 5 we can see that, taking the moderate smog of the fourth period
as our base, a very cold smog has more effect on old people, vide period 2, and the
intense smog of the third period has more effect still. It is conjectural how much of
the effect to be seen in the very cold spell following this smog is a delayed effect and
there is yet another weakness in that, as will be seen, the rise in the proportionate
mortality of the over 70's preceded the smog, even allowing for the fact that we are
using seven day moving averages.
The smog of
Jan., 1956
In Figure 6 therefore has been shown the seven day moving average for all deaths,
deaths of persons aged 70 years and over and persons aged under 70 for the period 24th
December, 1955 to 26th February, 1956 (covering the third period which is generally
admitted to be a smog even though there may be doubts about the others) with the
object of trying to grapple with the question of the elusive base line from which to
assess increased mortality due solely to smog. Mortality in the over seventies was rising
sharply before the impact of the smog indicated by the vertical lines on the curves at
30th December (30th December because the seven day moving average centred on
this day is influenced by 2nd January, the first day of the fog), but in the under seventies
it was falling (with a resultant slight rise for total deaths) which accounts for the
proportionate rise in the over seventies in Figure 5. How far these rises would have
progressed without the intervention of smog is a matter of conjecture. On the assumption
that, for total deaths, the rise was the beginning of a ' normal winter rise ' in
mortality a line has been drawn from the point of impact of the fog to link up with
the steep rise in February (shown by dotted line). Again, whether without the very cold
spell the line would have taken this course we do not know—to obtain a satisfactory
seasonal average is impossible because of the shifting incidence and severity of fogs,
cold, epidemics, and the greater the number of years over which any average is
calculated the more different does the pattern of mortality become. In the prevailing
circumstances deaths above this line can reasonably be regarded as attributable to smog
alone and in total they amount to 480 as compared with Logan's (1956) estimate of 440;
the two estimates are, however, built up on difference premises—Logan took as his
datum-line the average number of deaths (131) over the ten days 25th December to
3rd January, any deaths over this line were attributed to smog but he stopped counting
'excess' deaths at 14th January whereas the 480 'excess' deaths extends to 24th January
on the moving average, or 27th January in fact. For the over seventies it is difficult to
know where to draw any similar line; working backwards from the end of February
the logical line would seem to link up with the 24th December at around 70 deaths
daily and on this basis the area above the line represents 640 excess deaths in old people.
On the other hand, if the datum-line is drawn horizontally from the point of impact
of the smog at the level of 90 daily deaths then the excess becomes 370. In the preceding
and following year a similar rise occurred following Christmas and the same kind of
rise is suggested from the figures of weekly death registrations in four out of six preceding
years. This post-Christmas rise appears to occur in deaths from cardio-vascular causes,
the main cause of death in old people. If we assume that the total for all ages is about
right we are left with a choice of a rise of 640 among old people offset by a fall of 160
in younger ones which would mean raising the trend line for them to start at 60 deaths
a day; or starting our trend line for old people at the 90 level and leaving the young
as they are. On balance the latter course seems more justifiable bearing in mind the
post-Christmas phenomenon mentioned earlier.
Total mortality and deaths from appropriate causes for all ages during the two
months covered by Figure 6 are shown, by weeks in Table VI.
206