Inseok Ko1, Hyejung Chang2. 1. Division of Biomedical Informatics, Seoul National University College of Medicine, Seoul, South Korea. 2. Department of Management, Kyung Hee University School of Management, Seoul, South Korea. Electronic address: hjchang@khu.ac.kr.
Abstract
BACKGROUND: Interactive visualization is an important approach to help to understand and to explain large amounts of data, particularly in light of decision support. Although data visualization have been introduced in healthcare and clinical fields, analytics has often been performed by data experts, focused on specific subjects, or insufficient statistical evidence. Therefore, this study suggests the procedures of effective and efficient visualization of big data for general healthcare researchers. Specifically, the procedure includes conventional regression analyses followed by interactive data visualization for prescription patterns of antihypertensive drugs. METHODS: As a large-scale nationally representative prescription data, the Korean National Health Insurance claims data were collected. Conventional descriptive and regression analyses were conducted for therapy decision and prescription patterns using the software R. Then, based on the statistically significant findings, dashboards were developed to visualize interactively the patterns of prescriptions using the software Tableau. RESULTS: Major characteristics (genders, age groups, healthcare institutions, and comorbidities) explained the differences in therapy and the average number of drugs prescribed as well as differences among most commonly prescribed drug classes. Two interactive dashboards were created for visualizing prescription patterns with incorporation of horizontal bar charts, packed bubble charts, treemaps, filled maps, radar charts, box and whisker plots, and filters. CONCLUSION: In the current big data era, interactive data visualization offers substantial opportunities to have comprehensive view, extract insights and evidence from the flood of vast amounts of data. This study's interactive visualizations can provide healthcare professionals insight into prescription patterns and demonstrate the value of creating interactive dashboards to support informed and timely decision-making. Exploring big data using interactive visualization is expected to deliver many future benefits in healthcare fields.
BACKGROUND: Interactive visualization is an important approach to help to understand and to explain large amounts of data, particularly in light of decision support. Although data visualization have been introduced in healthcare and clinical fields, analytics has often been performed by data experts, focused on specific subjects, or insufficient statistical evidence. Therefore, this study suggests the procedures of effective and efficient visualization of big data for general healthcare researchers. Specifically, the procedure includes conventional regression analyses followed by interactive data visualization for prescription patterns of antihypertensive drugs. METHODS: As a large-scale nationally representative prescription data, the Korean National Health Insurance claims data were collected. Conventional descriptive and regression analyses were conducted for therapy decision and prescription patterns using the software R. Then, based on the statistically significant findings, dashboards were developed to visualize interactively the patterns of prescriptions using the software Tableau. RESULTS: Major characteristics (genders, age groups, healthcare institutions, and comorbidities) explained the differences in therapy and the average number of drugs prescribed as well as differences among most commonly prescribed drug classes. Two interactive dashboards were created for visualizing prescription patterns with incorporation of horizontal bar charts, packed bubble charts, treemaps, filled maps, radar charts, box and whisker plots, and filters. CONCLUSION: In the current big data era, interactive data visualization offers substantial opportunities to have comprehensive view, extract insights and evidence from the flood of vast amounts of data. This study's interactive visualizations can provide healthcare professionals insight into prescription patterns and demonstrate the value of creating interactive dashboards to support informed and timely decision-making. Exploring big data using interactive visualization is expected to deliver many future benefits in healthcare fields.
Keywords:
Antihypertensive agents; Comorbidity; Data mining; Decision making; Drug prescriptions; Hypertension; Interactive visualization; National Health Insurance claims database; Prescriptions; Software
Authors: Andrew Stirling; Tracy Tubb; Emily S Reiff; Chad A Grotegut; Jennifer Gagnon; Weiyi Li; Gail Bradley; Eric G Poon; Benjamin A Goldstein Journal: J Am Med Inform Assoc Date: 2020-05-01 Impact factor: 4.497
Authors: Jawad Chishtie; Iwona Anna Bielska; Aldo Barrera; Jean-Sebastien Marchand; Muhammad Imran; Syed Farhan Ali Tirmizi; Luke A Turcotte; Sarah Munce; John Shepherd; Arrani Senthinathan; Monica Cepoiu-Martin; Michael Irvine; Jessica Babineau; Sally Abudiab; Marko Bjelica; Christopher Collins; B Catharine Craven; Sara Guilcher; Tara Jeji; Parisa Naraei; Susan Jaglal Journal: J Med Internet Res Date: 2022-02-18 Impact factor: 7.076