794 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS ]•.•,'l'r ansm i ttal Sheets Key Punched Cards Cards sorted to a specific report (s) Report Generation Computer d.etc' Unit Record Processing Sortin Routine /•Card to tape utility Input •etic • Tape 19 Generated Reports Resort cards for another specific report (s) Figure 6. Computer system flowchart of the dandruff pilot clinical program

DATA PROCESSING OF ANTI-DANDRUFF RESEARCH 795 Listed below are the generated reports by identification letter and title. •em A B C H I J D E F G K O P L M N R S Title Condition of scalp at start dry Condition of scalp at start--oily Condition of scalp at start--normal Correlation of condition of scalp at start to quantity of loose dandruff Correlation of condition of scalp at start to quantity of adherent dandruff Correlation of condition of scalp at start to quantity of loose and adherent combined Quantity of loose dandruff Quantity of adherent dandruff Quantity of loose and adherent dandruff combiued Degree of itching Overall improvement rating Degree of dermatitis Reduction of dandruff vs age groups Reduction of dandruff vs years with dandruff Hair loss during combing and shampooing (Start vs eighth week) Test product rating Comparison of dandruff results Recorded likes about test product Recorded dislikes about test product COMPARISON' TIME AND COST The estimates described below are based on a dandruff clinical study involving 717 subjects. It should be noted that the cost of manually compiling and analyzing data remains constant for each clinical study of the same size and nature, regardless of the number of studies conducted. The net cost of each electronic compilation would be the actual running cost plus a pro-rata share of the initial programming cost. This share of programming cost diminishes as the number of studies increases assum- ing no changes iu programming are required. The time requirements and dollar costs per study (assuming five studies are to be run) are shown below: Manual method Electronic method Cost per study $2500 $800 Time requirement per study 6 weeks 1.8 weeks CONCLUSION What has been described here is properly referred to as a "pilot pro- gram." It was instituted to determine feasibility, costs, and time re- quirements. The computer program did not include tests of statistical significance or plotting of curves. Such a program can be modified to