dr2 Consulting brings diagnostic meta-analysis into the 21st century with dr-ROC. Until now, you could get dedicated summary ROC programs that would analyze your data for you, but their interfaces were idiosyncratic at best, and you'd have to open a spreadsheet or graphics program and retype your results to get graphs or tables that were suitable for publication.
Now, dr-ROC is here, and you can do everything in one program. Just think of the time and money you'll save! With one data entry and a few clicks, you can get a greater depth of analysis than any other diagnostic meta-analysis package, plus a host of customizable publication-quality graphs.
And there's more. dr-ROC does heterogeneity analysis, statistical analysis of user-supplied study quality scores, and pre-test/post-test probabilities, among other analyses. Subgroup analyses are facilitated by a group category in your input, so you can switch studies in and out of the analysis without having to cut and paste. dr-ROC also includes data checking and alert routines to try and catch data entry errors before they make it to your manuscript.
dr-ROC is designed for flexibility--to analyze data the way you want it analyzed, not just the way a programmer thinks it should be analyzed. You can choose to weight the analysis by inverse variance, by study size, by quality, or a user-specified parameter, or you can do an unweighted analysis. And if you're not sure about the effects of a methodologic choice like that, you can quickly analyze both ways and see what kind of difference it makes.
The same flexibility applies to the output side too. Each graph includes a set of check boxes and menus that toggle display components on or off. The options are clearly marked--no more alphabet codes to memorize. Check the boxes you want, click OK, and your graph is ready. dr-ROC also offers more graph types than any other package of its kind. Besides the summary ROC graph, you'll get plots of the original data with confidence limits on sensitivity and specificity, a logit plot to illustrate the regression and explain how it works, a post-test probability graph, and a forest plot of study-by-study diagnostic effect sizes.
dr-ROC is based on software used for years in production meta-analysis, then completely rewritten from the ground up to combine an easy-to-use interface with maximum flexibility and power. It has been validated by comparison of results to other programs, and the basic summary ROC algorithm (1) has been validated by meta-analysis of numerous simulated data sets with known characteristics (2). The new interface benefits from years of experience in meta-analysis of diagnostics for technology assessment production, not just ivory tower exploration.
Thoroughly and lovingly documented, using dr-ROC is an education in itself. The open structure of the program allows you to collect intermediate results and/or do further calculations, while new graphs like the logit plot help you see the relationships among studies.
dr-ROC runs within Microsoft Excel, so it works on both Macintosh and Windows computers. The basic program can analyze up to 25 studies. dr-ROC 100, which can analyze up to 100 studies, is available at no charge to licensed users of dr-ROC
System requirements: dr-ROC requires a licensed copy of Microsoft Excel.