Retrolental Fibroplasia: A Modern Parable – Chapter 11

Chapter 11
The Price of Progress

I have focused so closely on the treatment disasters which occurred during the first decades after World War II that I must declare the obvious: the picture was not one of unrelieved gloom. New knowledge resulting from expanded preclinical research gave rise to clinical applications which seem to have passed the crude test of time (although there is no way of knowing whether all of the “ticking bombs” have been detected). There has been resigned, almost philosophic, acceptance (in the medical community and in the community-at-large) of the view that the “accidents” which occurred were the inevitable price that had to be paid for progress. Public support for clinical explorations was whole-hearted in the early years. The success of the Cooperative Study of RLF in ending the blindness epidemic was often mentioned in news accounts as an example (and, in 1953, one of the earliest instances) of the effective use of government monies for extramural clinical research (research not performed directly by federal agencies). The visible return of government investment in the 1954 field trial of the Salk poliomyelitis vaccine-the largest public health experiment ever conducted -added to public awareness and approval. In the late 1950s, requests to Congress for funds to expand the National Institutes of Health extramural research programs (both preclinical and clinical) cited the conquests of RLF and polio as bread-cast-upon-the-water examples. Congress was lavish in its praises and largesse. In the 5-year period between 1955 and 1960, a great deal of federal “bread” was cast into the medical ocean (Table 11-1).

Table 11-1

National Institutes of Health Research Grant Expenditures
and U.S. Gross National Product in 1950, 1955 and 1960*

The large disbursements spurred development of the programs for premature infant care which were described in Chapter 10. At the same time, research activities which were directed at neonatal and, later, prenatal problems expanded remarkably in the United States and abroad following the RLF incident (Table 11-2). There was intensification of background studies in physiology and biochemistry. Knowledge concerning perinatal matters increased immeasurably, and the momentum has continued to the present day. Developments and innovations in the management of pregnancy have been extended and elaborated to include evaluation of the wellbeing of the fetus. Care of the new infant now begins well before birth. The state of health of the fetus can be assessed from measurements in mother’s blood and urine and from small samples of amniotic fluid (obtained by amniocentesis). Genetic defects (and the gender of the fetus) may be determined early in pregnancy by examining cells in amniotic fluid and by chemical analyses of the fluid. The growth of the fetus (especially the head) can be measured accurately by means of serial ultrasound images. The position and function of the supply organ of the fetus — the placenta — can be monitored. The state of maturation of the lung (which determines the risk of developing hyaline-membrane disease) can be assessed from chemical analysis in amniotic fluid, and lung maturation can be accelerated [by administering a hormonal drug (betamethasone) to the mother] when premature delivery is threatened. During labor and delivery the fetal heart rate can be monitored continuously and, from samples of blood taken from the scalp of the fetus, the biochemical status (acid-base and blood gases) can be accurately gauged. After delivery, electronic and biochemical surveillance of the small and sick newborn infant take place in intensive care units under the supervision of highly trained personnel and with the help of all of the panoply of modern life-support “hardware.”

Table 11-2

Citations Concerned with Perinatal Topics and
Total Medical Citations in 1950, 1955 and 1960

From the “inevitable-accidents-as-the-price-of-progress” point of view, the costs of the misadventures along the way must be weighed in the balance. It can be argued that the net gains for American infants have been enormous because of the expansion of activities brought about by the investment of public funds in perinatal research since the mid-1950s. For proof, according to this view, one has only to point to the progressive improvement in outcome of pregnancy and delivery, as measured by intact survival. The same general argument has been made concerning the beneficial effect of expansions in all other fields of modern medicine. And, the formulation has been subjected to increasing criticism from many quarters in recent years. The debates are timely, and, in my opinion, will lead to long-overdue inspection of some fixed dogmas. For example, the “inevitability-of-accidents” premise deserves a close look, and I shall return to a criticism of this belief in the chapters which follow. At this point I wish to take issue with the other propositions in the argument.

First, the conventional opinion concerning the immediacy of the growth of modern scientific medicine should not be allowed to pass without a challenge. There is some evidence which suggests that this perspective of the recent past may be much too narrow. Price calculated the rate of growth of science as a whole (using numbers of publications and the size of scientific manpower as quantitative indicators). He found that the present exponential rate has been steady for some time:

In 1900, in 1800 and perhaps 1700, one could look back and say that the number of scientists alive is greater than the total number of all previous scientists and most of what is known . . . has been determined within living memory. Scientists have always felt themselves to be awash in a sea of scientific literature that augments in each decade as much as in all times before.

The popular view that the advancement of medical knowledge is solely the result of increased expenditure of public funds for medical research (the notion that new ideas can be purchased) is a worrisome distortion of reality. Indeed, there is reason to suspect, from Price’s studies, that the gains from increased expenditure are limited by the law of diminishing returns. He found an inflationary trend (in science as a whole) during the past few decades in the United States, the U.S.S.R., and less so in the rest of the world. With the marked increase in social status of the scientist and need for his services, competition developed: general salaries rose, and there were automatic increases in research funds and facilities commanded by the prestige and by the cargo cult of modern science. Each increase in prestige produced a pay-off in increased scientific results, but heightened competition raised the stakes for the next round. A recent critique of Soviet science also raised questions about the quality of research as volume increased. One Russian scientist opined, “You can’t make an ocean liner by stringing together a bunch of row boats.”

Additionally, the post hoc argument which implies that most of the improvement in medical outcomes (particularly in perinatal results) in the United States is the result of specific technical interventions overstates the case for causality. In spite of the solidly grounded scientific studies on which many of the present-day developments in perinatal medicine are based, there have been few systematic attempts to evaluate their practical effects. The relative contribution of the technical inflation to secular trends in perinatal mortality rates is unknown. The “in-tandem” effects of social and demographic changes which have occurred in the United States cannot be ignored. For example, significant changes in the distributions of births occurred when there was an overall decline in birth rate in New York City (following the establishment of widely available family-planning services in the years 1966-1970, and legalized abortion beginning in 1970). The fall-off in rates of birth among “high-risk” groups (women whose risk of premature birth was known to be high were they to become pregnant) was particularly abrupt; these groups were

• women at both extremes of the childbearing age-range (the very young and the relatively older women);
• women who had a history of many births;
• unmarried women;
• women whose socioeconomic status was relatively low.

And, there was a marked decline in the births of small babies, especially those weighing 1.5 kg (= 3 lb 5 oz) or less. These shifts were accompanied by an unprecedented fall in infant mortality (the decrease was particularly striking in the years following legalized abortion). A similar experience was reported by Lee and others at Bronx Municipal Center in New York City. They analyzed mortalities which occurred among newborns in this hospital during the years 1966 through 1973. There was a close relationship between annual mortality rate and the proportion of very-low-birthweight babies (under 1.5 kg) born each year. In 1970, two events occurred: intensive care of newborn infants was increased at the hospital (there were significant changes in trained personnel and introduction of specialized equipment), and an active abortion program was initiated. Analysis of the fall in mortality rates which occurred in the unit in the years following the two changes suggested that a relatively steep reduction in the births of very small babies accounted for three-quarters of the improvement; the remaining difference could be attributed either to improved medical care or to improved health status of the infants at birth.

Morris and coworkers evaluated several possible causes for the decline in infant mortality seen throughout the United States during the late 1960s. They found a systematic shift in the distribution of births during this period from relatively high-risk categories to those of lower risk for infant mortality (Fig. 11-1). The changing maternal-age/birth-order distribution of births since 1965 accounted for about 27 percent of the reduction in infant mortality rate that occurred over the next 12 years in the United States. The contribution of organized Maternal and Infant Care Projects (funded by the U.S. Children’s Bureau beginning in 1965) was also examined. Under the most generous assumptions, only a small fraction of the decline in mortality could possibly be attributed to the success of this well-intentioned effort to deliver high-quality services to high-risk mothers and infants. From other (sporadic) studies, there is reason to suspect that other factors (e.g., increased use of contraception for spacing pregnancies, increased economic and social support during pregnancy, increased emphasis on the importance of adequate diet in pregnancy) also may have contributed to the changes in pregnancy outcome in recent years.

Despite the continuing trend of decreased mortality rates, the United States, as it has for many years, has a relatively low standing in international comparisons. The lag is frequently quoted in support of the need for increased supply of personnel, facilities and equipment for intensive technical supervision of perinatal activities. In this connection, Hinds examined the relationship between infant mortality rate and the number of health care workers available (per capita) for the years 1966 through 1970 in 28 developed countries of North America, Europe, Asia, and Oceania. Lower rates of mortality were more closely related to numbers of auxilliary health workers (nurses and midwives) than to the numbers of physicians available. St. Leger also explored the relationship between infant mortality and the provision of doctors (making an adjustment for the variations in affluence in 18 developed countries, as measured by gross national product per capita — Fig. 11-2). The strong positive correlation seen in his figure (countries with the greatest number of physicians per capita had the highest infant mortality) is just as provocative (and as suspect) as the oft-cited relationship between increased technical interventions and survival.

There are also questions about the relationship between interventions and the “quality” of survival. A community study conducted in Newcastle upon Tyne, England, recorded the functional abilities in the first decade of life of 13,203 children born in the years 1960 through 1962. It was found that even two of the most lethal perinatal adverse factors which can be directly modified by medical actions (breech delivery and prolonged delay in establishing regular respiration) appeared to have only trivial effects on the subsequent neurologic state of the children. Any effects which they may have had on the intelligence or the behavior of the survivors were explained by an association with two groups of much more potent determinants: biologic characteristics and social factors. In the first category, variations in birth weight seemed to have the most important effect: low birthweight was the characteristic most commonly associated with recognizable forms of brain damage. The second group of factors, whose effects far outweighed those of the perinatal ones, comprised the social attributes. The least specific, and yet the most powerful of these, was the child’s social class of origin as determined by father’s occupation. The cultural and general environmental implications of social class appeared to be of paramount importance in determining outcome. In a large U.S. study, the perinatal correlates of intellectual performance were examined in 26,760 4-year-old American children who were born in the years 1959 through 1965. It was concluded that perinatal circumstances made a minor contribution to outcome as compared with the major effect of social factors. Low socioeconomic status of the family (as measured by educational attainment and occupation of the head of household, and by family income) was associated with a relatively high frequency of retarded development in the preschool children. The findings of these two large-scale studies were consistent with the conclusions in two smaller surveys reported in 1967 and 1970; again, social class variables had a much greater effect on intellectual development than did the presence of perinatal complications.

The outcome of the smallest and most immature infants cared for in neonatal intensive care units since 1970 (which roughly marks the beginning of the “modern technologic era”) has been watched with much apprehension. It was important to know whether a fall in mortality among these babies could be achieved without paying a high price in terms of handicapped survivors. Drillien reported that between two 5-year periods (1948-1952 incl. and 1955-1960), the survival rate of Edinburgh babies with birthweights of 1.36 kg (3 lb) or less had increased from 17 to 30 percent; the change was attributed to advances in perinatal care over the span of 12 years. Unfortunately, the increase in the survival rate seemed to have taken place at the cost of an increase in the proportion of moderately or severely damaged children among the survivors: the handicapped rate rose from 32 percent in 1948-1952 to 56 percent in 1955-1960. It was gloomily predicted that as the survival rate of infants of very low birthweight improved with the introduction of new life-saving methods, an increasing number of damaged children would survive. A series of reports from University College Hospital, London, set out to refute this disturbing assumption. Improved survival depends on the anticipation of biochemical and other abnormalities which the smallest infants are unable to withstand, it was noted. These abnormalities-such as oxygen-lack, low blood glucose, and elevated concentrations of the toxic yellow pigment, bilirubin-are all capable of damaging the brain, the English observers argued, and their prevention should lead to an improved outlook for mental development among the survivors. With these thoughts in mind, they reported (1971) the developmental progress of a group of infants who weighed 1.5 kg (3 lb 4 oz) or less at birth and who had been treated more vigorously than in the past and with the newly developed techniques. The survey suggested that the outlook had improved: among 72 surviving infants only 5 were abnormal and 4 were classified as “doubtful” (when examined at ages ranging from 9 months to 4 1/4 years). Further reports from the London group have argued that there is continuing improvement in the prognosis for the infants in the smallest and most immature ranks. And, the favorable outlook has been attributed to intensive methods of care. Several follow-up observations from other neonatal intensive care units have also presented encouraging accounts, although many of the children were still very young at the time of the reports.

An ongoing survey of the eye changes among infants in an intensive care nursery in Vancouver, British Columbia, was begun in January 1968 by McCormick. Between the years 1968 and 1976, he examined 2031 infants and he noted a sharp rise in the numbers of eye-damaged infants beginning in 1974. This increase was associated with a general increase in the number of infants admitted to the Canadian special care unit, with a considerable increase in the number of small infants transported from hospitals not equipped to care for sick newborn infants, and with an apparent rise in the survival rate of the smallest premature infants.

The interpretation of wholly descriptive surveys in medicine is difficult. The information has been particularly slippery in the perinatal field of inquiry because of the “built-in” problem of population selection and the practical difficulties of accounting for the effects of events which occur in the long interval between discharge from the hospital and the time of formal evaluation of the outcome. And, the interpretive difficulties have mounted in recent years. The demographic composition of the population of infants born alive, as I indicated earlier, has been changing rapidly. The development of regional programs for perinatal care has resulted in an increasing distortion in the selection of the population of infants in intensive care units (for example, one-half of the children in the original University College Hospital follow-up survey had been born in other institutions and were transferred to the specialized unit for care). In addition to these issues which raise problems in interpretation, there are questions introduced by changing policies in the vigor with which life-support measures are applied to infants who are severely compromised (e.g., severe malformations, prolonged asphyxia, suspected hemorrhage into the brain). To emphasize once more, there are unavoidable (but major) problems associated with the long interval which must elapse between the events in the perinatal period and the time when a reliable assessment of outcome can be made. In addition, the social environment in which children are reared has changed rapidly over time. Despite all of these difficulties in interpreting the practical effects of the new childbirth technology, surprisingly few doubts have been expressed concerning the favorable claims which have been made. I do not wish to imply that the rescue efforts have been fruitless, but what is missing is some quantification (from formal tests of specific interventions and an accounting of the multivariate influences which affect outcome). Difficult as it is to apportion the effects of the numerous changes which have been occurring, it seems mindless to ignore the complexity of the situation. And, in the face of the uncertainties, I believe it is unconscionable to lobby for increased investment in efforts focused on technologic escalations. The point is that the relation between outlay and outcome cannot be settled by the ex post facto evidence presently available. Furthermore, there is now a healthy skepticism about the cost-effectiveness of medical programs. Questions can no longer be put off by the kind of simplistic cause-and-effect assurances which sufficed in the past.

I indicated above that there is some evidence which suggests that the importance of perinatal events is small in comparison with the effects of socioeconomic circumstances during the first few years of child development. Sameroff and Chandler conducted a critical review of the available evidence on this subject. They pointed out that almost all perinatal follow-up studies have proceeded on the assumption that the developmental course of the human infant is a linear chain of efficient causes and invariant effects; particular characteristics of either the child or his parents should, therefore, predict the ultimate course of growth and development. And failure to predict later pathology has been taken as a mandate to initiate a new search for the elusive cause. On the other hand, an alternative developmental model regards the process of human evolvement as a more circuitous one, in which linear chains of causality are rare. According to the latter interpretation (the organismic model), children and their environments are undergoing regular restructuring: early perinatal factors that have enduring consequences are assumed to do so because of persistent influences acting throughout the life-span. This explication suggests that acute events in the perinatal period should have minimal long-term effects. It is assumed that “self-righting” influences are powerful forces which are disposed to the promotion of normal human development; protracted developmental disorders are found only in the presence of equally distorting influences. From this point of view, predictive failures in the study of developmental disorders are not attributable to missing links or other simple constructions. Instead, the defects are assumed to be related to inadequate knowledge regarding the complex of mutual influences that operate between the child and his environment. Together, these serve to dissipate or to amplify the effects of earlier developmental insults. To gain predictive validity from the perinatal influences one must take into account the maintaining environment. Sameroff and Chandler introduced the phrase “continuum of caretaking casualty” to characterize the circumstance which may arise when a disadvantaged family with poor emotional, physical, and economic resources must cope suddenly with the consequences of the birth of a premature or asphyxiated infant. This crisis is likely to put a burden on the parents’ limited caretaking abilities. The chance that such a child will receive close and attentive care is further jeopardized by his influence on family interaction. The infant’s disturbance in alertness, vigor, sucking, or stability of sleep/wake cycles has a cumulative and disrupting effect on the development of a smooth child-parent relationship. And, there is evidence to make the thesis of the “continuum of caretaking casualty” a fairly reasonable one. Neurologic signs present at 1 year of age in infants who are asphyxiated at birth tend to persist in children who grow up in disadvantaged environments. The outlook for infants (with similar birth histories and abnormal signs at 1 year) appears to be different if they are reared in socially advantaged settings: in many of these fortunate children, the signs disappear and many function normally by the age of 10 years. Werner and her associates analyzed the outcome of 1000 live births and concluded that perinatal complications were related to later physical and psychological development only when combined with and supported by persistently poor environmental circumstances. Moreover, the biologically vulnerable children constituted only a small proportion of the number who were not functioning normally at age ten. Ten times more children had problems related to the effects of poor early environment than to the effects of perinatal stress.

The studies cited by Sameroff and Chandler suffer from the same maddening shortcomings to which I have alluded. Nonetheless, there is at least as much evidence to support their interpretations as there is for the arguments concerning the singular effects of perinatal events. It would be shortsighted, Baum has recently noted, to ignore the influences of the child’s environment on his development after a complicated course in the newborn period. I do not wish to exaggerate more than is necessary for effect, but to the extent that resources and efforts are focused exclusively on technologic support in pregnancy and delivery, I must say that planning vision is disturbingly limited. Doctor Laura Nader was moved to comment on the current situation at the conclusion of a conference in 1975 entitled “Critical Issues in Neonatal Intensive Care.” Questions should be raised, she noted, about who benefits economically from neonatal intensive care: the companies that produce the life-support machinery, the doctors who work at this labor, insurance companies, the hospital, the parents, the families of the newborn infant, the baby? How has our society come to be spending so much time and money on neonatal intensive care, she asked, without similar attention to born healthy but later not-so-healthy, deprived children? These provocative questions remain unanswered.

Infant-mortality risk categories by mother’s age at different ordinal numbers-of-birth. Beginning in 1965 (through 1972), a shift occurred in the distribution of births in the United States: from the high-risk age/birth-order categories toward the low-risk “cells.” This shift was sufficient to account for about 27 percent of the reduction in U.S. infant mortality.

Relationship between infant mortality and doctors provided per 10,000 population in 18 developed countries examined by St. Leger. Mortality is expressed as the “residual” difference after making an allowance for gross national product per capita in these countries (1970 data).

Return to Contents Page