Increasing the Efficiency of Presenting Study Population Descriptive Characteristics: The “Table 1” Macro
descriptive statistics, macros
Background/Aims: Basic characteristics of a data set or cohort are often described using a “Table 1.” Numerous iterations of this table can be required before the final product is satisfactory. Reproducing this table multiple times can become onerous and increases the chance for data entry errors. Some programmers create their Table 1 through the output delivery system (ODS) in SAS and some output data into Excel. A recently funded Center for Effectiveness and Safety Research project supported us in developing macros to create a “Table 1” macro that can easily accommodate format and content changes and output data in either ODS or Excel format.
Methods: The Kaiser Permanente Colorado analytic team developed two different macros for creating “Table 1,” one for frequencies, means and p-value output through ODS into a pdf or rtf file and one for frequencies and means output into an Excel file. We describe the features, advantages and disadvantages of each macro.
Results: By converting the Table 1 development process into a macro, we have increased the efficiency of producing and reviewing the results of descriptive analysis while at the same time facilitating different examinations of the data. Both macros have the flexibility to present typical descriptive statistics, such as frequencies, means and comparative p-values. Output from ODS produces a table that is ready to distribute without additional manipulation by the programmer. However, it can be challenging to set up the macro variables correctly. An advantage for the Excel output macro is ease of use, whereas some disadvantages include the time required to format the spreadsheet after data are exported into Excel and moving data into the correct fields. These macros are durable in that they can be applied across multiple projects and variables can be customized for any Table 1 analysis.
Discussion: These macros increase the efficiency of producing and reviewing descriptive analyses and reduce errors that can be introduced from data entry and creating tables de novo. Future work includes combining the macros so that the user can choose the preferred output destination.
Carroll NM, Tavel HM, Clarke CL, Powers J, Feigelson HS, Raebel MA. Increasing the Efficiency of Presenting Study Population Descriptive Characteristics: The “Table 1” Macro. J Patient Cent Res Rev 2015;2:138-139. http://dx.doi.org/10.17294/2330-0698.1190