Society for Pediatric Research Presidential Address 1997: Multiculturalism in Research

Good morning. Friends, colleagues, and guests, welcome to the presidential plenary session of the Society for Pediatric Research (SPR). I am pleased to have this opportunity to address you and have chosen as my topic,“multiculturalism in research.”

The word multiculturalism crept into the popular vocabulary about 10 years ago to describe a philosophy that advocates incorporating into any activity people from various backgrounds to gain the benefit of the unique knowledge and skills that each person brings from his or her experience. This term, as originally coined, implies specifically that we should strive to include individuals of different race, ethnicity, and gender. Although attainment of this goal is vitally important in all spheres of work, including research, I do not intend to address multiculturalism along these lines today. Rather, I wish to extend this concept to teamwork by scientists from different cultures, with diverse expertise, from molecular and cellular biology, through the physiology of animal models and clinical studies of patients, to computer simulations and population epidemiology (Fig. 1).

Figure 1

Multiculturalism in research.

Although applying the term multiculturalism to research may represent a new twist of semantics, it is certainly not a new concept. Others have used phrases such as bench-to-bedside, translational, or integrative biomedical research to describe this type of endeavor. My major reason for referring to this phenomenon as multiculturalism, as opposed to choosing one of the other terms, is that it emphasizes the personal, not just the professional, aspect of the teamwork. For me, the opportunity to work with highly talented individuals of varied interests has proven incredibly fulfilling. The collaborations and friendships arising from my research often have made otherwise boring days into adventures and have kept me going at times when shrinking budgets, hospital mergers, and an ever-growing administrative burden have seemed overwhelming. Although real challenges confront our community in both research and clinical medicine, I hope that sharing my enjoyment of multiculturalism will provide an alternative and more attractive view of a career in clinical investigation to those who may have heard a less optimistic message previously.

I will begin with some personal experiences, move on to describe what I consider an important relationship between multicultural research and the SPR, and finally highlight the broader implications of multicultural research in medicine.

Almost 25 years ago, I applied for a pediatric residency, expressing in my interviews a desire both to care for patients and to be a scientist, but not really having the foggiest notion as to what I was talking about. Not long thereafter, I found myself in the emergency department at the Children's Hospital of Philadelphia, evaluating a young child with a high fever and diffuse aches. She appeared remarkably well despite a temperature of 40°C, and my examination uncovered no definite focus of infection. My senior resident felt that the child's pain was more pronounced in the region of her left hip and recommended some laboratory studies. The pursuit of septic arthritis led up a blind alley, but added a white blood cell count of 18,000/mm³ to the picture. Despite the unexplained fever and the leukocytosis, we assumed the child had a viral infection. Eighteen hours later she returned with purpura fulminans. State-of-the-art intensive care proved futile, and she died in septic shock just as her blood culture grewNeisseria meningitidis.

The next day, I struggled with the effect my clinical decision had had on her fate. At about that time, Klein et al. published with his colleagues a landmark article showing that approximately 5% of young, febrile children with apparent viral syndromes had, in retrospect, bacterial infections of the bloodstream, primarily with Streptococcus pneumoniae or Haemophilus influenzae⁽¹⁾. Subsequently, Friedman and Fleisher⁽²⁾ reported the occurrence of occult meningococcal bacteremia in children who recovered after receiving antibiotics as outpatients, without developing any overt signs of meningococcemia. These observations suggested that children did not progress directly from perfect health to sepsis or meningitis, but rather developed these serious infections after an initially asymptomatic bacteremia.

Armed with my newfound knowledge about the role of bacteremia in pathogenesis, I proceeded to think about a series of questions stemming from my encounter during internship with the meningococcus. First, I wondered whether any aspect of my patient's appearance should have provided a clue to the subsequent outcome. Addressing this issue required research in the clinical culture. Using the Yale Observation Scale, developed by McCarthy and colleagues⁽³⁾ to quantitate the degree of illness in children, Teach et al.⁽⁴⁾ evaluated young, febrile infants. Although the groups with and without bacteremia had significantly different mean scores, the marked degree of overlap(Fig. 2) precluded the designation of a cutoff point that would offer sufficient accuracy for diagnosis.

Figure 2

Yale observation score (YOS) in children with and without bacteremia. (Reproduced with permission from Teach et al., Journal of Pediatrics 126:877-881. 1995⁽⁴⁾.)

The next question from my patient concerned the interpretation of her leukocytosis. To gain insight into this area, first Jaffe and Fleisher⁽⁵⁾ and then Kuppermann et al.⁽⁶⁾ used the technique of receiver operating characteristic curve analysis, a methodology developed by the military to study the detection of missiles by radar operators. The curves that we generated from these clinical studies verified earlier reports of elevated white blood cell counts in children with occult bacteremia and demonstrated that a level of 15,000/mm³, yielding approximately a sensitivity of 80% and a false positive rate of 25%, would provide the greatest diagnostic accuracy. Additionally, use of logistic regression analysis showed that a combination of the white blood cell count, temperature, and age was a better predictor of bacteremia than white blood count alone, and could be used to define subgroups with risks ranging from less than 1% to greater than 10%(Fig. 3).

Figure 3

Risk of bacteremia in children 3-24 months of age based on height of fever and absolute neutrophil count (ANC). (Reproduced with permission from Kuppermann et al. Pediatric Research 39:135A, 1996⁽⁶⁾.)

The third question generated by my patient was whether earlier antibiotic therapy would have prevented her fatal outcome. As a first attempt to probe this point, Jaffe et al.⁽⁷⁾ conducted a randomized, double blinded trial of amoxicillin versus placebo among children at risk for occult bacteremia, showing that the rate of complications, 10.5% versus 12.5%, was only minimally better in the treated group. Subsequently, a number of investigators from around the country participated with me in a comparison of oral amoxicillin and intramuscular ceftriaxone, using a higher dose of amoxicillin than in the previous study by Jaffe et al. Among approximately 7000 young children with a temperature of over 39°C, who were randomized to receive one of these two antibiotics, we observed significantly fewer predefined definite infectious complications with ceftriaxone (0/101 versus 5/91.p = 0.02)⁽⁸⁾ in the 192 patients with bacteremia.

Once these clinical studies had been completed, it was time to consider more general guidelines for managing well appearing febrile children. The sample size required for a study to demonstrate a significant reduction in meningitis after treatment for occult bacteremia, 40000 patients, posed an obstacle to performing a randomized, prospective clinical study, but created an opportunity for me to interact with investigators versed in another culture, decision analysis. Using existing data, Lieu et al.⁽⁹⁾ developed a computerized simulation of a hypothetical cohort of 100000 highly febrile infants 3-36 months of age and calculated the number of infections that would occur among bacteremic children with six different strategies, assuming that the incidence of bacteremia was 3% and that amoxicillin was used as the initial treatment with an effectiveness of 20%. The first pass, using these baseline assumptions, showed that appropriately selected interventions could prevent 48% of the expected 600 cases of sepsis, meningitis, and other focal infections. A sensitivity analysis for the different strategies, which varied the efficacy of antibiotic therapy from 0 to 100%, demonstrated that, as more effective antibiotics were administered, three approaches (Fig. 4), employing early presumptive treatment, could eliminate nearly 100% of sequelae.

Figure 4

Sensitivity analysis of strategies for evaluation and treatment of febrile children. (Reproduced with permission from Lieu et al., Journal of Pediatrics 118:21-29. 1991⁽⁹⁾.)

The final question from my patient was whether alternative or additional therapeutic measures could have been used to treat her once she became septic. The study of this piece of the puzzle lent itself, at least in part, to the culture of the animal laboratory. Collaborating with George Siber, my colleagues and I produced models of endotoxemia and sepsis in rabbits, which we monitored both biochemically and hemodynamically. Animals challenged intravenously with endotoxin by Baldwin et al.⁽¹⁰⁾ or intraperitoneally with E. coli by Saladino et al.⁽¹¹⁾, became hypotensive and acidotic over a period of 6 hours, and subsequently died, in most cases from shock.

At this point, I learned that Stan Watson and a group of scientists at Associates of Cape Cod were purifying a protein from the horseshoe crab,Limulus polyphemus, that was capable of avidly binding to endotoxin. Endotoxin induces the activation and degranulation of horseshoe crab hemocytes, which results in hemolymph coagulation. The protein being studied by Watson and colleagues-eventually named endotoxin-neutralizing protein, or ENP, once it had been sequenced and cloned-served as an inhibitor of endotoxin-mediated activation of the gelation cascade. Theorizing that ENP might have a role in endotoxin blockade in human disease, we began a collaborative series of experiments.

Initially our group demonstrated that ENP prevented the onset of endotoxic shock in rabbits when given mixed with or just before a dose of endotoxin. Rabbits receiving endotoxin developed tachycardia and hypotension, which were prevented by the administration of ENP, as a result perhaps of the ability of ENP to lower circulating endotoxin levels and down-regulate the production of tumor necrosis factor, as well as other cytokines⁽¹²⁾. Rabbits treated with ENP had a significantly improved survival rate compared with controls (Fig. 5)⁽¹²⁾.

Figure 5

Survival of septic rabbits treated with antibiotics and either endotoxin neutralizing protein (closed triangles) or saline(open squares). (Reproduced with permission from Saladino et al., Critical Care Medicine 24:1203-1207, 1996⁽¹²⁾.)

To gain a more fundamental understanding of ENP and its potential clinical applications required more work at a basic level. Thus, the groups at Associates of Cape Cod and the Dana Farber proceeded to study the ENP molecule. Shortly thereafter, George Siber, Bob Liddington, and colleagues succeeded in determining its crystal structure, which consists of an N-terminal α-helix followed by a four-stranded β-sheet and two C-terminal α-helices⁽¹³⁾. Analysis of the structure suggested an endotoxin-binding motif with similarities to epitopes in proteins such as bactericidal permeability increasing protein, but did not fully explain the superior therapeutic efficacy noted in our animal models of sepsis, when comparing ENP to structurally related compounds. Thus, Evelyn Kurt Jones began a series of investigations into whether ENP inhibited the effects of endotoxin by a mechanism other than simply binding directly to it⁽¹⁴⁾. In these experiments, endotoxin was added to a culture of mononulcear cells, and the binding of endotoxin to the cells was measured. Increasing doses of ENP (preincubated with the cells which were then washed) progressively inhibited binding of endotoxin to the cells. These results indicated that ENP interacted not just with endotoxin, but also directly with receptors on the mononuclear cells.

Now I would like to move away from my own specific areas of research and extend the concept of multiculturalism beyond being merely a source of fulfillment and lifelong friendships for me. Specifically, I would like to address the importance of multiculturalism in a larger context, starting with the relationship between the SPR and multicultural research in pediatrics and continuing with an exploration of its value to society as a whole. In terms of the SPR, I believe our organization has been able to foster the gathering of researchers that span these diverse cultures, while simultaneously benefiting from the presentation of superb science which has had a profound impact on the health of children. I think a few concrete examples will serve to clarify these concepts, and I have chosen three diseases for this purpose.

I consulted on a young girl with leukemia in 1978, while I was a fellow with Dr. Stanley Plotkin. Herpes zoster erupted on her trunk(Fig. 6), and a few days later, she died. Another young girl made two visits to the emergency department in the early 1980s with fever and was found to have nuchal rigidity (Fig. 7) the second time around. A lumbar puncture was done, revealing H. influenzae, and antibiotics were quickly administered. Nonetheless, she suffered neurologic sequelae. Finally, this infant, treated at Children's Hospital in Boston, had Pneumocystis pneumonia (Fig. 8). He is still alive, but his prognosis as a severely immunocompromised child infected with human immunodeficiency virus (HIV) is guarded.

Figure 6

Herpes zoster in a girl with leukemia.

Figure 7

Nuchal rigidity in a young girl with meningitis due toH. influenzae.

Figure 8

Pneumocystis carinii pneumonia in an infant infected with HIV.

Although these three patients did not fare well, had they been born in 1997, all might have avoided or survived their diseases, given the development of varicella zoster immune globulin, acyclovir, varicella vaccine, Hib vaccine, and zoduvidine. To put the prognosis of patients with these diseases in better perspective and to expand upon the relationship of multicultural research and the SPR, I conducted a study to test the hypotheses that:1) the SPR has provided a forum for multicultural research and2) the presentation of multicultural research at the annual meeting has benefited the SPR and potentially fostered scientific progress. To gather data, I reviewed SPR programs for abstracts related to varicella, H. influenzae, and HIV for a period of 20 years, commencing with 1977 and relying both on the subject index and a page-by-page search through the section on infectious diseases (ID). For each relevant abstract I recorded the title, authors, year published, and abstract number. Then, I assigned a culture category according to the following schema:

• 1 Basic science studies that were performed in the laboratory, usually at the cellular or subcellular level, or with genetically altered animals;

• 2 Animal studies that used models with nongenetically altered animals to determine the pathophysiology or treatment of various diseases;

• 3 Clinical studies that involved children as experimental subjects;

• 4 Epidemiologic studies that used large populations, existing data sets or computer models; and

• 5 Bicultural studies that correlated the results of assays in a basic science laboratory to the pathophysiology or outcome of clinical diseases.

During the 20 years, a total of 552 abstracts appeared in the ID section, 14% on varicella, 41% on H. influenzae, and 45% on HIV(Fig. 9). The total number of abstracts on these infections increased sharply over the two decades. Abstracts related to varicella were least common and remained at a constant level(Fig. 10). H. influenzae declined in frequency in the 1990s, perhaps coincident with the decreasing incidence of invasive disease. As one might anticipate, abstracts on HIV first appeared in the early 1980s and then grew rapidly in number. I divided the abstracts into the five culture categories, as defined earlier, and found that approximately 55% were clinical, 15% bicultural, 15% basic science, and smaller proportions either animal models or epidemiologic studies, truly a multicultural distribution(Fig. 11).

Figure 9

Distribution of abstracts on selected topics, 1977-1996.

Figure 10

Changes in frequency of abstracts on selected topics over two decades.

Figure 11

Distribution of abstracts according to type of research.

Where has multicultural research led us? Turning first to varicella, the available data are old and incomplete, but suggest a slight downward trend in deaths (Fig. 12). More dramatic results are available for H. influenzae. As illustrated in a graph from the Communicable Disease Center (Fig. 13), the last few years have witnessed a marked decline, if not the virtual disappearance, of invasive disease due to H. influenzae. For HIV, the most remarkable change to date has followed the widespread introduction of antiviral treatment into perinatal practice. The graph from the Morbidity and Mortality Weekly Report (Fig. 14) shows that the corner has been turned on a national level.

Figure 12

Deaths from varicella. (Reproduced with permission from Morbidity and Mortality Weekly Report 45:2-3⁽¹⁵⁾.)

Figure 13

Decrease in invasive disease due to H. influenzae. (Reproduced with permission from Morbidity and Mortality Weekly Report 45:902⁽¹⁶⁾.)

Figure 14

Decline in disease due to HIV (Reproduced with permission from Morbidity and Mortality Weekly Report, 45:1007⁽¹⁷⁾.)

I hope you will allow me some latitude with my study and agree that the data about these three diseases support my hypotheses about multicultural research and the SPR. Clearly, large numbers of abstracts from diverse scientific cultures have been presented each year, and it seems a safe assumption that the opportunity to hear the latest updates about the conquest of major causes of pediatric mortality made for meaningful and exciting meetings. Although harder to prove, I would offer for your consideration the conclusion that the scientific exchange across cultures and disciplines that occurred as a result of these 552 abstracts contributed at least in some small way to the progress achieved over the last 20 years.

Moving even beyond the SPR, I believe that multiculturalism has a broader and more profound importance for science and for child health. In an article appearing several years ago in the New York Times, reporters described the results of a learning exercise at Stanford University in which the graduate students in business and engineering set out to design and sell new kinds of bicycle lights. Traditionally, the engineers, referred to as“propellerheads” by their counterparts in the business school, would first have taken the lights from the drawing board to the final product and the MBAs, known as “bean counters” on the engineering campus, would then have attempted to sell the items. In the new paradigm, teams consisting of students from the two schools worked together concurrently on the project. The result of this coordinated effort was that professionals who learned to speak both languages facilitated the development of a more effective and affordable bicycle light than had previous groups engaged in what we in pediatrics might describe as “parallel play.”

Reviews of scientific progress nationally have found that similar principles apply to biomedical research. In 1994, Philip Jobe and Colleagues⁽¹⁸⁾ published in the Physiologist a report written for the Institute of Medicine, entitled: “The Essential Role of Integrative Biomedical Sciences in Protecting and Contributing to the Health and Well-Being of our Nation.” In addition to finding an essential link between subcellular and cellular studies and medical practice, they spoke to four potential consequences of insufficient training of individuals prepared to work as a team across the spectrum of research, including:

• 1 An impaired national capacity to develop intact animal models of human function and disease;

• 2 A declining pool of scientific manpower with the skill to conceptualize biomedical hypotheses and experiments at the level of the intact organism;

• 3 Deterioration of key elements of preclinical and clinical drug development; and

• 4 A rapidly diminishing capacity to provide the scientific data essential to medical education and to rational judgments in clinical practice.”

In their recent commentary in Nature Medicine, entitled“The Slowing of Treatment Discovery, 1965-1995,” Wurtman and Bettiker⁽¹⁹⁾ pointed out that, although the national effort in biomedical research to acquire knowledge about how the body works had succeeded magnificently, discoveries of new treatments occurred rather infrequently during the 30 years under analysis. They stated, “Treatment discovery turns out to be very much a directed or mission-oriented enterprise, requiring the participation of investigators committed to that task. The dialogue between basic, clinical, and applied researchers leading to drug discovery is a disordered one that invariably occurs over extended distances and times. Arguably the drug discovery process could be accelerated by planning and coordination among the various professional groups involved.”

Although I have chosen to conclude the body of my remarks with comments about the importance of multicultural research, I want to continue to emphasize the personal benefits as well. From this perspective, if you engage in multicultural research, I cannot guarantee that you will make a pivotal discovery, but I can promise, based on my experience and my observations of work done by others, that you will contribute to pediatric science, develop rewarding relationships, and define for yourself not only a worthwhile, but also an enjoyable, career.

In closing, I would like to turn to a few relationships, from among many that I value, which have held particular importance for me.

In addition to basic scientific methodology, my first and foremost mentors, Werner and Gertrude Henle, taught me that one seldom has a completely original idea. Almost every time I rushed into their office with a new proposal in mind, Werner or Gertrude would pull down a well worn laboratory notebook, occasionally from before the war (and I don't mean Vietnam or even Korea), and dig out an experiment that they had performed to answer a similar question about a different organism. At the end of their medical training and at the inception of their scientific careers, the Henles fled Nazi Germany, gave up clinical practice, and immersed themselves in their laboratory. They worked together there for 50 years, gaining entry to the National Academy of Sciences, winning the Robert Koch Medal, and receiving the Bristol Myers Award for their achievements in basic science. Yet, they were the first to highlight for me the value of collaboration among investigators from different scientific cultures.

When I arrived in Boston in 1986, I was fortunate to connect very quickly with George Siber. We proceeded to work together for the next 10 years, and, through George, I established further collaborations. George offered me invaluable insight into the analysis of experimental data and a clearer understanding of the process of turning basic discoveries into clinical advances.

The last colleague I wish to thank is David Nathan. Although I never worked in David's laboratory or published a manuscript with him, he has helped me in many ways. Simply by being who he is, David has served as an example for me of someone who not only has made tremendous contributions as both a scientist and a physician but also has been able to multiply his contributions to pediatrics by appreciating and rewarding achievements by his colleagues, whether clinicians, clinical researchers, or basic scientists. More than anyone else, he has served as my adviser over the last decade.

The fact that I am up here on this podium, hopefully in the right place at the right time, is a tribute to Deb Anagnostelis and her staff, who so capably manage the SPR. Through dint of their hard work and organizational abilities, I have been kept focused on the issues this past year.

And finally, I could never have attained this office, or even membership in the SPR, were it not for the support of my family. I have even benefited from multiculturalism of a special sort at home, inasmuch as my wife, Jan Paradise, and I have been able over the years to co-author four publications, and, at various times, my children, Daniel, Madeline, and Carl, have helped me at the keyboard with data entry. Jan and the children have tolerated my long hours away and have nurtured me at home, and I am most grateful to them. Obviously, my work has truly been a team effort.