Article type
Year
Abstract
Background: The highlight of previous systematic reviews has been focusing on meta-analyses of randomized-controlled trials and non-randomized studies. In several clinical issues with a lack of meaningful sized comparative studies, although the issue is important, there are rare modalities to analyze or visualize the mapping of evidence.
Objectives: To establish a novel tool for ‘evidence mapping’ in clinical issues which have multiple treatment options but also have numerous, dispersed, and small-sized evidence.
Methods: We developed a web-based plotting program using Java-script and named it ‘plotting-e-map (PLOEM)’. In the example of recurrent pancreatic cancer, there are five kinds of treatment options but which therapy is the best is still controversial. Because of its clinical characteristics, meaningful-sized comparative study is rare. Instead, literature screening showed 75 studies including case reports. Using the PLOEM program, we assigned ID numbers for 75 individual studies and inserted the basic information (study type, publication year, author, sample size, etc.) of all the studies into the application.
Results: The visualized evidence mapping is shown in Fig 1. There are numerical dots (from 1 to 75): the shape of each dot represents the study type: 1) case reports in blue diamonds; 2) case series (sample size 2 to 10) in green ellipses; 3) single-armed cohort studies in yellow hexagons; 4) observational comparative studies in pink pentagons; and 5) prospective comparative studies in brown rectangles. The number inside each dot matches the study ID, and each dot is linked to its corresponding study by clicking on it. Users can easily understand the trend for 40 years in this condition. Figs 2 and 3 emphasize linkages between studies and gray alphabets indicate types of study combination. Users can try to conduct meta-analyses as a concept of subgroup analysis.
Conclusion: Our new modality (PLOEM) enables users to look over research trends at a glance and to perform subgroup meta-analyses. To distribute this program widely, we are developing an open-access website and patches for statistical analytic programs such as Stata or R.
Objectives: To establish a novel tool for ‘evidence mapping’ in clinical issues which have multiple treatment options but also have numerous, dispersed, and small-sized evidence.
Methods: We developed a web-based plotting program using Java-script and named it ‘plotting-e-map (PLOEM)’. In the example of recurrent pancreatic cancer, there are five kinds of treatment options but which therapy is the best is still controversial. Because of its clinical characteristics, meaningful-sized comparative study is rare. Instead, literature screening showed 75 studies including case reports. Using the PLOEM program, we assigned ID numbers for 75 individual studies and inserted the basic information (study type, publication year, author, sample size, etc.) of all the studies into the application.
Results: The visualized evidence mapping is shown in Fig 1. There are numerical dots (from 1 to 75): the shape of each dot represents the study type: 1) case reports in blue diamonds; 2) case series (sample size 2 to 10) in green ellipses; 3) single-armed cohort studies in yellow hexagons; 4) observational comparative studies in pink pentagons; and 5) prospective comparative studies in brown rectangles. The number inside each dot matches the study ID, and each dot is linked to its corresponding study by clicking on it. Users can easily understand the trend for 40 years in this condition. Figs 2 and 3 emphasize linkages between studies and gray alphabets indicate types of study combination. Users can try to conduct meta-analyses as a concept of subgroup analysis.
Conclusion: Our new modality (PLOEM) enables users to look over research trends at a glance and to perform subgroup meta-analyses. To distribute this program widely, we are developing an open-access website and patches for statistical analytic programs such as Stata or R.