Introduction
The methods of meeting a naval emergency—a rapid increase in personnel aboard ship, lengthened periods in the forward areas, sacrifice of personnel space in favor of ordnance—all place considerable stress on human performance. One way in which these stresses may manifest themselves is by increasing the number of incapacitations from disease and accidents. Consequently, a quantitative knowledge of this effect may help to determine to what extent increasing the fire power of a vessel at the expense of its personnel is in the long run a profitable venture. This is a question of some practical importance, and, in addition, the same quantitative information is of theoretical interest insofar as it sheds light upon the mechanism whereby stress takes effect on a human population. For these reasons it was thought worth while to undertake a study of the effect of combat, overcrowding, and weather on the incidence of certain diseases and accidents aboard several Fleet vessels.
That the rates of incidence of certain groups of diseases, particularly respiratory, are affected by climate and weather is widely held to be an obvious fact; but as in the case of other “obvious facts,” the proof is not so simple and the literature dealing with this subject is voluminous. In his studies of climatic factors, C. A. Mills[1] visualizes the body as an engine which runs rapidly in a cool, dry environment and sluggishly in a hot humid environment whether or not these environments are due to season or geography. A beneficial result of rapid, vigorous function, such as might occur in a cool climate, is the adequate maintenance of defense against infectious disease; a deleterious result is a “wearing out” process which eventually leaves the body open to metabolic and circulatory afflictions. On the other hand, in a tropical climate one might expect a susceptibility to infectious disease but, conversely, a relative freedom from heart failures and so on. This state is complicated, according to Mills, by the fact that certain regions of cool or temperate climates, and some tropical zones, are subject to cyclonic storminess (rapid changes in temperature, humidity, and barometric pressure) which in some as yet unexplained way precipitates infectious diseases. The balance between the influences of these climatic factors would determine in any region the rate of incidence of a group of diseases. Without necessarily accepting Mills’ views literally, we may adopt them as a working hypothesis and examine them in the light of existing data. When this is done, it may be said that much evidence points to weather as one of the causative factors in respiratory disease, but quantitative study of this relationship has hitherto been impaired by the following difficulties:
(a) The populations compared have differed socio-economically as well as in being subject to different climates.
(b) Practically all large studies are based solely on mortality rates. Consequently, non-fatal infections have not been subjected to extensive study, and the time interval of the rate of incidence can therefore not be made small enough to follow rapid weather changes, else the rates themselves will be too small. Finally, diagnosis of the cause of death may be unreliable.
(c) It is very difficult to dissociate the obvious weather variables from the other unknown factors which may exhibit a seasonal variation and which conceivably could be causative.
It was thought a priori that the analysis of naval data would largely circumvent all of these difficulties, for aboard naval vessels the “socio-economic factors” are about as uniform and constant as could be hoped for, admission rates of uniformly and well-diagnosed patients are available, and the weather variables change through wide ranges practically in experimental fashion without reference to season. To some extent these expectations seem fulfilled even in the present incomplete study.
When we pass from infectious respiratory diseases to other incapacitations which reduce the active complement aboard naval vessels, we get next to no help from the existing literature. The effect of seasonal variations on industrial accident rates has been studied in a few instances, but the results have little transfer value.
So far as the authors are aware, there is available no large population work on the incidence of gastrointestinal and skin disease as a function of weather. The same appears to be true regarding the naval medical effects of overcrowding and of combat, although the exigencies of the war just passed certainly suggest the importance of such an inquiry. It was with this in mind that the authors embarked upon this study.
Data Used in the Present Report
Lest there be special differences among vessel types with regard to the general health and safety of personnel, an attempt was made to select from various types; thus two battleships (the Washington and the Idaho), two heavy cruisers (the Wichita and the Indianapolis), two light cruisers (the Santa Fe and the Richmond), two destroyers (the Caldwell and the Bancroft), and one escort carrier (the Nassau) were chosen. Moreover, in the case of the BB’s, the CA’s and the CL’s one of the two ships was comparatively new and the other comparatively old. In all cases, the ships had had considerable combat experience in both the cold northern waters and the hot, humid southwest Pacific waters. The histories of these vessels were followed from January 1, 1941 (or from the date of commission in the case of new ships) to December 31, 1944, by means of their routine official records.
In contrast to land, the diurnal variation of temperature and humidity on the ocean is small, so that it is possible to use the temperature and humidity at any one time of day as representative of the day. Accordingly there were recorded, from the navigational logs each day, the noon (zone time) dry bulb temperature and absolute water vapor pressure,[2] as well as the highest and lowest barometric pressures for the twenty-four hour period.
For each day, there was recorded from the war diary the combat status of the ship, namely, (a) sailing under normal wartime conditions, (b) about to attack or be attacked, or (c) actually in contact with the enemy.
For a measure of overcrowding, the ratio of actual complement to design complement was decided upon. This ratio was computed monthly. The so-called design complements are calculated on the basis of standard volumes of space to be allotted per man, and are fairly uniform throughout the larger vessels of the fleet, both old and new.
Of the many causes which may commit a man to sick bay for a time in excess of twenty-four hours, certain ones may be expected to be subject to weather, combat, and overcrowding. Among these are the respiratory infections, gastrointestinal infections, skin infections, and accidents other than those due to misconduct. Within any one category there are of course diseases with different etiologies, but they have enough in common so that it can be expected that gross environmental changes will affect them in the same way. Through December, 1944, it was regulation to report monthly the number of men discharged from sick bay, together with their diagnoses (NMS Form F). Since the average length of stay in sick bay is short (of the order of three days), and since successive months compensate for each other, the number discharged each month is approximately equal to the number admitted each month. This assumption has been made in the discussion which follows. The number of cases in each category was noted and recorded in the conventional manner as “cases per thousand.”
Before proceeding to discuss the analysis, it is instructive to examine the ranges of conditions to which the crews of these vessels were subject.
As was to be expected on the basis of selection, all of the ships cruised in a large variety of weather conditions without reference to season. In the course of three or four years all records exhibit many instances of rapid change and also several months of steady conditions. Compared with the weather experienced in ordinary civilian life, that met by naval personnel is, of course, extreme. Steady (monthly average) air temperatures occurred in the range of 30-90° F. The rates of change of temperature which seem to be of medical interest are of the order of 0.4-8.0° F. per day. Water vapor pressures ranged from 0.15 to 1.2 inches of mercury. Mean barometric pressures ran from some 28.9 to 30.0 inches of mercury, with frequent barometer jumps of one inch of mercury in the same day. It was found for all ships, that when the temperature was high, the water vapor pressure was high and the barometer was quite steady. Conversely, in cold weather, the water vapor pressure was low and the barometer fluctuated widely. This association of the weather variables is, of course, not necessarily present on land.
Aside from anti-submarine actions, combat was infrequent during 1941-1942. In the years 1943-1944, however, the number of days in which a ship was either in combat or immediately anticipating combat, was substantial. In order of decreasing number of days, the ships were: Santa Fe (128 days), Bancroft, Washington, Indianapolis, Wichita, Caldwell, Nassau, Idaho, Richmond (23 days). The longest consecutive period in combat was 23 days for the Santa Fe.
The factor of overcrowding (ratio of actual complement to design complement) ran an erratic course in most ships, and ranged from 0.7 to 2.2 among the nine. However, there was observed a steep increase aboard each ship when the average factors for 1941-1942 were compared to the average factors for 1943-1944. The older the ship, the more overcrowded it was, but there was no discernible difference in overcrowding of large ships as compared with small ones. Either the number of ships was too small or the effect was completely masked by the relationship with age of the ship.
Analysis and Results
From the information described above, composite charts for each ship were prepared plotting the incidence rates of the various diseases and accidents, the weather variables, and the factor of overcrowding against the calendar day.[3] Combat and alerted days were also recorded on the same chart. These graphs were first examined for parallelism between extreme fluctuations in the operational conditions and coincident fluctuations in the disease and accident rates. Next, for each category of incapacitation a comparison between steady extreme conditions was made. In the case of temperature, the average monthly incidence over cold months was compared with the average monthly incidence over hot months. In the case of overcrowding, the average monthly incidence during uncrowded years (1941-1942) was compared with the average monthly incidence during crowded years (1943-1944). Also, the ships were arranged in order of the average monthly incidence over the crowded years and this order compared with the average measure of overcrowding during these years. Finally, the arrangement of ships in order of incidence rates was compared with an arrangement in order of number of combat days plus days of alert. In making these comparisons, care was taken to see that either “all else was approximately equal” or that the other operational variables had been demonstrated to be ineffective on the basis of the initial examination of the graphs.
(1) The influence of weather. It is obvious at first inspection that weather exerts a far greater influence on respiratory diseases than on the other types of incapacitation. The incidence of respiratory infections throughout a few years can be best described as a constant (roughly 5 or less cases per thousand men) interrupted here and there by “spurts”[4] or peaks ranging from 15 to 30 or more cases per thousand men. There were 38 such spurts distributed among the nine ships. Thirty-one of these were associated with a drastic temperature variation. The variation consisted in passing from high temperature to low temperature or vice versa, and the total rise or fall in temperature was of the order of 25-30° F. This essential feature seems to be the rate of change with which some sort of adaptation fails to keep pace (Figure 1). The association between these spurts and weather appears to be high; in only five cases were there similar variations without sharp increases in the rate of respiratory infections.
The factor of contagion by contact with land was not controlled in this study, but certain considerations make it appear unlikely that contagion alone could account for these outbreaks. For example, the study covers 351 ship-months, and on the average a vessel is at sea without contact with land for consecutive periods of one month; thus in 351 ship-months there were approximately 351 contacts with land. With an occurrence of 38 spurts, this means that had contagion been the sole cause, it must have been ineffective 313 times (89 per cent). This would seldom be regarded as the attribute of a causative factor. Furthermore, epidemiological experience suggests that the introduction of susceptibles into a population is far more dangerous (from the point of view of initiating a new wave of infections) than the introduction of infected individuals. One might look, therefore, to substantial increases in complement as a contributory factor. When this was done it was found that 19 of the 31 spurts associated with weather changes coincided with fair increases in complement. These results lead to the tentative conclusion that certain types of weather sequence very probably condition spurts in respiratory infections, and that their action appears to be reinforced by the addition of new men on board.
When infection rates during steady cold months are compared with rates during steady hot months, it is found that the former are consistently, though only slightly, higher.
In contrast to respiratory infections, gastrointestinal disorders show no spurts, and the average monthly incidence shows no relation whatever to combat or to overcrowding. In six of the nine ships the average incidence was very slightly higher during months of sustained hot weather than during months of sustained cold weather.
The association of rapid change in temperature with changes in the incidence of skin diseases occurred three times, of which the best example is the year 1944 aboard the Indianapolis (Figure 2). Most ships do not show this effect, and the Richmond shows a greater incidence during colder temperatures. It should be noted that these results refer to skin infections severe enough to require sick bay admission. Milder afflictions such as heat rash show a marked temperature dependence. The rate of incidence of skin infections averaged around 5 per thousand; spurts around 10.
The accident rate shows a pronounced spurt character, but the spurts are associated with various operational conditions. The highest degree of association noted (6 out of 11 spurts) was with increases in personnel. These spurts are of the order of 5-10 per thousand whereas the average rate of incidence is about 3 per thousand. In comparing steady temperatures there appears to be a somewhat greater incidence in hot than in cold weather.
(2) The influence of combat and alerts. Despite the great range in number of combat days during 1943-1944, the operation of this factor over a long period of time is completely imperceptible in all four categories of incapacitation. There was observed a very slight association between combat and spurts in the accident rate. However, it must be concluded from the present information that whatever fatigue, nervousness, or sleeplessness arising from being in the danger zone is completely masked by the influence of other operational conditions.
(3) The influence of overcrowding. Increases in complement may condition increases in disease or accident rates in at least two ways: namely, by introduction of new individuals susceptible to infectious sources already aboard, and by limiting the berthing, sanitary, or messing facilities in either a temporal or a spatial sense. Substantial increases in personnel arc associated with spurts in the incidence of skin diseases. No such effect is observed in the case of gastrointestinal diseases. Six of eleven spurts in the accident rate are also associated with rapid increase in personnel, suggesting that newcomers on board may be hazards to safety.
There was no detectable long term effect of overcrowding when the uncrowded years, 1941-1942, were compared with the crowded years, 1943-1944, with due regard for changes in other conditions.
Comment
It is felt that although a rigorous analysis may invalidate some of the conclusions reached here, the use of naval records for studies of this kind holds much advantage over use of vital statistics of land populations, for it eliminates the complications of season and socio-economic status. In several instances the present conclusions fit in well with the work of others, and this tends to argue in their favor. Thus, it has been found that, as with influenza epidemics, certain characteristic weather sequences seem to trigger outbreaks of respiratory disease. It is also agreed with Mills that apart from socioeconomic status and season, respiratory disease is primarily a cold weather phenomenon. Because of the close association between low temperature and barometric fluctuations, it was impossible to check Mills’ claim regarding the predisposing effect of pressure fluctuations per se. The slightly increased incidence of gastrointestinal diseases and accidents during hot humid weather was expected, and the absence of an effect of serious skin disorders is not especially surprising. The interesting effect of sudden increases in complement on respiratory and skin diseases is in accordance with what is known about epidemic diseases, and the analogous effect on the incidence of accidents is not unreasonable. With the influence of combat and overcrowding, however, the parallelism between the observed and the expected ceases. It was confidently thought that both disease and accidents would show indirectly the effect of long tours of combat, for such tours naturally mean loss of sleep, irregularity of meals, and considerable tenseness. Furthermore, it appeared likely that all categories would show the effect of long term overcrowding because of the resulting shortages of facilities. Neither of these expectations was borne out in the present study.
The foregoing results have certain practical implications regarding the efficiency of the vessel. If the number of men immobilized in sick bay jumps from 5 per thousand to 30 per thousand, this means that on a battleship some 50 additional men are hospitalized; on a destroyer, 9 additional men. Since these could be any men, their inactivity may well imperil the efficiency of the ship. It appears from this preliminary study that the operational conditions capable of causing these incapacitations are predictable; thus preventive steps could and should be taken in advance of the conditions. Among the predictable conditions, as we have seen, are sharp weather changes (as in steaming from arctic to temperate waters in four or five days) and the taking aboard of drafts of new men. We note also that the weather-susceptible infections are a purely practical argument in favor of proper heating and ventilation installations. On the other hand, it may be noted (with reserve) that neither long term overcrowding nor long combat sequences seem to increase the rates of incidence for the diseases and accidents studied here—in direct contradiction to what would be expected.
The following table summarizes the findings:
| Operational Condition | ||
Weather | Combat | Overcrowding | |
Respiratory disease | Characteristic sequences of weather are strongly associated with “spurts” of these diseases. Greater incidence in steady cold weather than in steady hot weather. | No observable effect. | Substantial increase in personnel are associated with “spurts” of these diseases. No long term effect from overcrowding. |
Gastrointestinal disease | Slightly higher incidence during months of steady hot weather. | No observable effect. | No observable effect. |
Skin diseases* | Very slight association with hot weather. | No observable effect. | Substantial increases in personnel are associated with “spurts” in these diseases. No long term effect from overcrowding. |
Accidents | Slightly higher incidence during months of steady hot weather. | Very slight association. | Substantial increases in personnel are associated with “spurts” in the rate of incidence. No long term effect from overcrowding. |
* Cases of heat rash which do not require admission to sick bay are not included in this category.
After attending both the University of California and Harvard University, Lieutenant Morales was commissioned in the Naval Reserve in 1943. After preliminary training he was stationed at the Naval Medical Research Institute for two years, and then served as Research Observer on the U. S. S. Washington. At present he is instructor in Biophysics at the University of Chicago.
Lieutenant Tarver attended the University of Georgia, and then was trained in laboratory methods at the Division of Laboratories and Research of the New York State Department of Health. Commissioned in 1943, she served thereafter at the Naval Medical Research Institute, where she is now Research Assistant.
[1] Mills, C. A., Climatic factors, In Medical Physics, Chicago, The Year Book Publishers, 1944.
[2] This is a measure of the rate at which water tends to re-enter the body. The rate at which water tends to escape from the body is roughly a constant. The net rate is therefore the difference between these two. Net rate of escape is therefore measured by the absolute water vapor pressure of the air, and can be taken as proportional to the amount of cooling afforded by perspiration.
[3] In the case of quantities which were measured only monthly, the entries were made arbitrarily as of the fifteenth of each month.
[4] The definition of a “spurt” is necessarily partly arbitrary, but can be made quantitative. The criterion decided upon here was that the leading side of such a spurt must have a slope in excess of four cases per thousand per month for each 25 days, and must have a height in excess of the average for the year.