Ralph D’Agostino

Ralph D’Agostino stepped down from his role as editor of Statistics in Medicine (SIM) December 31, 2019. The major areas in which he excelled include being the lead biostatistician for the Framingham Heart Study, a biostatistical consultant to The New England Journal of Medicine, a member and consultant on federal drug advisory (FDA) committees, and a 52-year faculty member in mathematics and statistics at Boston University.

Editorial Service, Statistics in Medicine
D’Agostino has been a key player at Statistics in Medicine (SIM) since its first volume in 1982, with an article in the first edition, “The Logistic Function as an Aid in the Detection of Acute Coronary Disease in Emergency Patients (a Case Study)”. Of major importance, he became the lead editor of the popular Tutorials in Biostatistics portion of the journal in 1995, a post he held until December 2019.

In the decade from 2010–2019, there have been 115 such tutorials, many created for both students and instructors. One informative example, which was not even labeled a tutorial, comes from a SIM paper by Timothy Heeren and D’Agostino titled “Robustness of the Two Independent Samples t-Test When Applied to Ordinal Scaled Data.” It had a huge impact on how service courses in both statistics and biostatistics were taught. The conventional wisdom in 1987, when there were small sample comparisons from two independent groups with respect to a discrete ordinal outcome, was to employ the Wilcoxon test, and students would be seriously docked on exams if they employed the t-test. Heeren and D’Agostino proved teachers wrong. The null properties held up exceedingly well when they compared the p-values obtained from the t-test to the maximum exact p-value obtained over all null distributions in their 2 by K table of outcomes. As a special case, they noted the t-test is more robust than Fisher’s exact conditional test for comparing two independent small sample binomials.

D’Agostino saw a huge expansion of SIM. The first volume in 1982 had four issues with 386 printed pages; the volume before he joined the team (1997) had 24 issues with 2,936 pages; and his last volume as an editor (2019) had 30 issues with 5,672 pages. As an author’s editor, D’Agostino personally read an incredible number of submissions to SIM and constructively weighed in on the review process. His presence was major, especially when the reviewers were in conflict. Most importantly, he left no stone unturned when it came to ensuring good science won out, even if it flew in the face of conventional wisdom.

Statistical Consultant to the Editor, The New England Journal of Medicine
Since 2007, D’Agostino has served as statistical consultant to the editor of The New England Journal of Medicine. With an impact factor of 70.7 (number of citations in 2018 and 2019 divided by number of citable publications in those years), this factor is well above all other medical journals. Many of the most important clinical trials and diagnostic discoveries that frame clinical practice are published in this journal, and it is extremely important to properly vet the research design and analysis methodology of these papers. This appointment is a marker of the trust the medical community has in D’Agostino.

Senior Statistician and Co-Principal Investigator, Framingham Cohort
In 1983, D’Agostino assumed the role of senior statistician to the Framingham Heart Study (and co-principal investigator from 1985–2015), arguably the single most important and oldest continuously funded cohort studies of heart disease and stroke. He was funded in this role through 2015 but remains active in data mining for the cohort, where any survivors are now well over 100 years of age.

This study of adult residents of the city of Framingham, Massachusetts—a Boston suburb with current population of 68,000—began in 1948 and prospectively followed a cohort of 2,843 women and 2,236 men. An offspring study of 2,656 children of 1,644 husband-wife original cohort members began in 1971 and continues to this day.

During his tenure, the National Institutes of Health have awarded hundreds of millions of dollars for the research activities involving these cohorts. D’Agostino co-authored 305 peer-reviewed papers involving the Framingham cohort between 1984–2019. Of these, 48 were published from 2010–2019. Two landmark accomplishments of his work on this cohort are (1) leading the development of the Framingham Risk Scores and (2) playing a key role in the development of guidelines for cholesterol.

One explicit Framingham risk function (score) is given for men and women and encompasses age, blood pressure, total cholesterol, high density cholesterol, smoking, and diabetes. A Cox proportional hazards model was used with dependent variable 1 (myocardial infarction and/or coronary death) or 0 otherwise. The cholesterol guidelines involved risk scores for the estimation of the risk of coronary disease or stroke.

Service on Federal Drug and Device Advisory Committees
Drug and device advisory committees of the US Food and Drug Administration (FDA) are empaneled to review safety and effectiveness data, as well as review proposals by industry to test new products. The statistical components of these activities are of paramount importance, and these committees rely heavily on their biostatistical members. D’Agostino has been appointed to and/or is currently on no fewer than 13 of these committees. Of special note, he has served on the advisory committee for over-the-counter drugs continuously since 1976. He was twice awarded a special citation from the FDA commissioner (1981 and 1995), as well as the FDA Advisory Committee Service Award in 2008.

Faculty, Department of Mathematics and Statistics, Boston University
Immediately after earning his PhD in statistics in 1968 from Harvard University under the joint directorship of William Cochran and Fredrick Mosteller, D’Agostino joined the mathematics faculty (now mathematics and statistics) at Boston University. He is still active there, now in his 52nd year (60th if you count his four years as an instructor from 1964–1968 and his undergraduate training before that). He has been an exemplary member of the department, having served two six-year stints as department chair. He currently co-directs the Boston University Statistics and Consulting Unit (director from 1986–2015) and co-directs the biostatistics department’s MA/PhD program (since 1988). He also served as associate dean of the graduate school from 1976–1978.

D’Agostino has published 726 peer-reviewed papers (240 from 2010–2019), including 56 in SIM and 29 in The New England Journal of Medicine. His papers have been cited in Google Scholar more than 225,000 times and his H Index stands at 220, among the best numbers in any field.

D’Agostino will turn 80 on August 12.