CVD Epidemiology Leads the Way in Data Handling
Senior workers in CVD epidemiology recall the chatter of mechanical, then electric calculators as hand-written data were entered from field and laboratory notebooks. Before long, thoseworking in the new field were able to put aside their calculators and slide rules and start sorting punched cards. Then, in the 1960s, not long after workers became comfortable with cross-classifications and simple statistical tests advanced by card-sorting, digital electronic computers burst upon the scene. The result was a powerful facilitation of data storage and management, computation of correlations and regressions, and the facile handling analytically of multiple variables simultaneously with programs for the multiple logistic and hazard ratios (Truett, Cornfield, and Kannel 1967; Cox 1972). These advances were technically and emotionally akin to another 1960’s phenomenon then under way: man’s dramatic blasting off into space with flights to the moon!
While parallel development of computer technology had a powerful impact on design and analysis in epidemiological studies, the discipline of CVD epidemiology itself made major and early contributions to biomedical data handling. It was the pressure of dealing with massive field data collected in the many CVD surveys and cohort studies of the 1950s and 60s that generated an intense motivation to devise ways of managing and analyzing data on a new and immense and more complex scale for the biosciences.
Anecdotal examples of the evolution of data entry and storage, and of computing problems, and the techniques developed to cope with them, provide a flashback to those times. The effort here is not to treat this important technological history in exhaustive detail, which is well considered elsewhere, but to tell a few stories central to our topic of early prevention research with population data.