Is Health What We Know or What We Can Afford?

Book Review: 
Fatal Years: Child Mortality in Late Nineteenth Century America.
by Samuel H. Preston and Michael R. Haines

    The turn of the previous century marked a turning point in the rate of mortality in America.   In Fatal Years: Child Mortality in Late Nineteenth Century America  (1991), Samuel H. Preston and Michael R. Haines apply modern statistical analysis to data from the 1900 census and other information to conclude a cause for the change in the rate of mortality, a decline that began in the early 1900’s.  Specifically, they use child mortality figures as a measure of the relative healthiness of various population groups of the time and attempt to identify variables that might describe risk factors.
     Officially collected mortality figures available from the Death Registration Area began to be collected only after 1900 and was limited to information from ten north eastern states and the District of Columbia. Relying on this information alone would give a skewed view of mortality at the time, because it would omit statistics from the more rural and less populated regions of the country. Because the census was taken nationwide, its data could provide a more complete picture of mortality at the time just prior to the turn of the century. 
     Although the census did not collect detailed information on deaths in households, it did include information about the number of surviving children.  The authors develop a probabilistic model to use the census data on the number of surviving children born to a woman in 1900 to estimate child mortality.  The authors assume that the model is accurate because the estimates derived from the method correlate with other assays of mortality.  Information directly provided from census questions was then used to form groups between which mortality differentials could be compared.
     Sub groupings by race, region of the country, urban areas vs. rural areas, husbands’ occupations, wives’ (mothers’) working status, country of nativity, literacy, and questions with indirect income implications were made so that comparisons of mortality differentials could provide a picture of groups at higher relative risk of mortality.  Throughout the book, the authors use the sub groupings to correlate risks to a general socioeconomic or geographic condition.  For example, African Americans had higher mortality than Caucasians; New England, as a region, was unhealthier than the South; cities were associated with higher mortality than rural areas; and except for laborers, a husband’s occupation had little overall effect on mortality.

     Multivariate analysis was used to determine, considering all variables, which ones had the greatest impacts on health.  The authors found that the single most important characteristic in determining mortality was race.  African Americans had the highest probability of early death in childhood and, the authors noted, in adulthood as well.  The assumptions used to create some of the mortality models did not work well in the analysis of black mortality.   Multivariate analysis showed that size of place of residence had the second most profound effect on childhood mortality; urban areas were more unhealthy than rural ones.  Larger urban areas had greater differentials than smaller urban areas. Region of residence ranked next in the analysis.  Northern regions had higher mortality differentials than southern areas; the midwest had the lowest rates of child mortality of the nation.
     The authors assume from their analysis that population density was the common thread linking populations with higher mortality.  Cities and the relatively densely settled rural areas of New England were noted for their high mortality; analysis of homes with boarders also showed the effect of higher child mortality.  The authors assert their belief that infectious disease was the most common cause of death for infants and young children at the beginning of the twentieth century.  Dense population distributions would favor the spread of respiratory and gastro-intestinal diseases in a society that was unaware of the role of microorganisms in those disease processes. 
     The germ theory of disease was not widely understood by doctors, public health officials, or the public at large at the time.  Preston and Haines view the low mortality differentials between literate and non literate mothers, and among occupational groups as evidence that people did not have the knowledge required to protect themselves and their children from disease.  As there were few medical interventions available to treat disease, morbidity and mortality were distributed among the population with more favorable outcomes the consequence of geographic location.  Where conditions did not favor the easy transmission of disease, mortality rates were lower. In America, a family’s wealth, education or social status did not provide significant advantages in avoiding high mortality (with the repeated exception of African Americans).
    Preston and Haines refute theories that improvements in life expectancy during the twentieth century were the results of improved nutrition secondary to increased prosperity; or that improvements in medical interventions in disease treatment caused lower mortality.  The authors mention a competing theory offered by McKeown which also downplays the role of medical treatment, but maintains that improved nutrition was the significant factor in mortality decline.  Fatal Years establishes that although income had a significant effect on child survival, child mortality remained high because Americans had not yet embraced the germ theory of disease.  The authors argue that while Americans began the twentieth century in a state of universal ignorance of germ theory, those groups that had advantages had them by virtue of geographic location and parenting practices that tended to protect children from disease (breast feeding infants for longer periods, for example).  As germ theory became accepted, public health measures were undertaken to stem the transmission of pathogens.  Protected water supplies, sewage collection and disposal, sterilization of milk supplies, and quarantine of infected individuals represented great strides in preventing disease transmission where few interventions were available for treatment.
     They do not fully explain the wide differences in mortality between social classes existing in England and Wales at the turn of the century, however.  If income effects were of little consequence in America, why are these differences associated with wide mortality differentials in England and Wales where, presumably, the same ignorance of germ theory prevailed?  The authors use the opportunity to show that income differentials by occupation in the United States were lower than in England; social class structures were not as highly defined.  But the comparison serves to weaken their previous argument that income effects (as measured indirectly by occupational status) were not the most significant factors in American outcomes. Further, others have argued that the wealth gap of “Guilded Age” America was as high as those seen in Europe up to the turn of the century.
     Improvements in public health activity during subsequent decades were responsible for overall mortality declines, but the authors note that the more educated (and presumably more affluent) of American society enjoyed a faster rate of decline as mortality differentials between classes increased. This they attribute to greater acceptance of a scientific view of disease and changes in parental behavior - hand washing, boiling milk and infant utensils, for example, and not to income effects.

    Were that view that applicable to 1991, the year Fatal Years was published, one might argue that wealth is less important to child survival than parental behavior; that universal access to health interventions regardless of economic circumstance would have less impact on child mortality or overall health of a population than education in and adoption of healthy practices by individuals.  While the public health initiatives that were undertaken over the past century had benefits extending to all members of society regardless of whether they accepted the underlying principal of germ theory, educated individuals reaped the greatest benefits because they adapted their behaviors to put into practice methods to improve their own health.  The authors’ position suggests that when information exists that can improve health, there will be groups in society that will accept and put it into practice, and others who will ignore it with the consequence of poorer health outcomes.  Comparing the situation at the start of the century with today, they write:  “In place of sharp differentiation now commonly associated with behavioral differences among classes were important variations in mortality according to factors over which individuals had little or no control” (p209). 
    The debate over improved access to health care by the economically disadvantaged that became a national issue in the late 1980’s and prompted some states to expand Medicaid coverage to more children (Minnesota in 1988 became the first state to increase child health coverage) continues today.  Those wishing to find “objective” support for positions against the increased coverage might favor the authors’ view.      
     It is questionable that conclusions drawn from analysis of data describing America of a century ago can be used to describe the environment today.  While there was little curative power in medicine then, medical science has made genuine progress in its ability to cure and treat illness.  Infectious diseases were a primary cause of death in 1900 America, many of which are now curable with drug therapy.  However treatment is not a function of individual initiative and behavior as much as of availability, which is a function of individual wealth.  The position that ill health is primarily due to poor individual behavioral choices is misguided.
     Samuel Preston is the Frederick J. Warren Professor  of Demography at the University of Pennsylvania; Michael Haines is the Banfi Vinters Professor of Economics Colgate University.  As demographers, they rely heavily on statistical analysis in constructing their story of child mortality in 1900 America.  Life expectancy tables from the period and the rich descriptive data of America from the 1900 census along with a model whereby child mortality estimates could be extracted, allow the data to be analyzed in ways aggregate mortality figures cannot.  There is little probing into turn of the century life beyond the numbers, however.  More historical monologue might have explained the picture of life at the time and provided a context against which the reader could better interpret the statistical data.  Could the statistics have been interpreted differently with better historical background?  Preston and Haines note that while life in cities was associated with high mortality, the ten largest cities at the time actually had better mortality rates than the next largest cities.  Without investigating, they assume that this was because these largest cities began public health efforts earlier.  However, these cities also experienced the earlier growth of “suburban” areas which allowed wealthier residents to leave the congested cores; they may also have been the first to be connected to the public water and sewer facilities.  Statistics may show differentials, but a deeper understanding of history might provide better insights into interpreting the results.
     The authors also note repeatedly that African Americans had the highest rates of child mortality and, generally, the lowest life expectancy of any group in the United States.  Despite the glaring disparity, they do not explore the issue in the text.  Again, reliance on statistics alone may not be adequate to understanding all of the mechanisms related to mortality.  Greater historical perspective may have shed more light on the reasons for the differences shown by the statistics, particularly when the high child mortality differentials exist to this day.  While they dismiss the historical, racist presumption that genetic frailty was operative, they did note that Blacks were paid less than Whites in the same occupations, had restricted access to education, and were not included in many public health initiatives of the time.
     While the present may be built upon the foundations of the past, we must be careful about extrapolating conclusions about the past to the present without considering intervening history.  No one would argue that germ theory did not represent a great tool for advancing human longevity and continues to be essential in maintaining health; nor that individual behavior continues to have a tremendous impact on health.  While medical practice may not have had a measurable impact on mortality in 1900 America, it does play a major role today.  Improved access to medicine today may be the most effective means at further reducing mortality.  The modest impact of the income variable on health at in 1900 is not necessarily true in twenty-first century America.



                   
























Reference
    Fatal Years: child mortality in late nineteenth-century america (1991).
Preston, S. H. & Haines M. R.. Princeton University Press. Princeton, NJ

About Pamela Lazorko