An interactive web application to aid diagnostic test accuracy meta-analysis

Session: 

Oral session: Understanding and using evidence (3)

Date: 

Tuesday 18 September 2018 - 11:00 to 11:10

Location: 

Carrick 3

All authors in correct order:

Freeman S1, Kerby C2, Cooper N3, Sutton A3
1 University of Leicester; NIHR Complex Reviews Support Unit, United Kingdom
2 University of Leicester, United Kingdom
3 University of Leicester, NIHR Complex Reviews Support Unit, United Kingdom
Presenting author and contact person

Presenting author:

Suzanne Freeman

Contact person:

Suzanne Freeman
Abstract text
Background:
A meta-analysis of diagnostic test accuracy (DTA) studies synthesises multiple studies to evaluate the performance of a diagnostic test. Often there is variation between studies surrounding the threshold value used to determine whether a patient is healthy or diseased. DTA meta-analysis can be performed using either the bivariate or hierarchical summary receiver operating curve (ROC) models, and the results presented either around a mean point or as a summary ROC curve. Conventionally, summary ROC curves are published as static graphs. An alternative is to consider interactive graphs, which can allow multiple perspectives to be displayed and information to be tailored to the user's preference. Interactive graphs can also provide a useful tool to aid sensitivity analyses.

Objectives:
1) To develop a freely available, web-based 'point and click' interactive tool, which allows users to input their DTA study data and conduct meta-analyses for DTA reviews, including sensitivity analyses.
2) To illustrate the benefits of the interactive application using an existing DTA meta-analysis for detecting unhealthy alcohol use.

Methods:
To create our online, freely available, interactive application we used the existing R packages mada and Shiny to analyse the data and create an interactive user interface.

Results:
We created an interactive online application for conducting meta-analysis of DTA studies. The user interface was designed to be easy to navigate having different tabs for different functions. Benefits include the ability for users to enter their own data, to explore a range of different threshold values and what this means in terms of true and false positive rates, and the ability to conduct sensitivity analyses.

Conclusions:
We built a freely available, interactive, online application that meta-analyses DTA studies, plots the summary ROC curve and allows for sensitivity analyses to be conducted in a timely manner. This application will allow a wide range of users to carry out specialised analyses without needing software beyond an internet browser.

Patient or healthcare involvement:
This interactive application will allow patients and healthcare professionals to visualise the results of DTA meta-analyses and to explore assumptions and sensitivity analyses in an interactive manner.

Relevance to patients and consumers: 

Diagnostic tests aim to distinguish between patients with the disease and healthy individuals and form an essential part of today’s current medical practices. Diagnostic tests are often a pre-requisite to identifying treatment options, enabling access to services and predicting patient outcomes. One diagnostic test may be considered across multiple studies considering a range of thresholds. The interactive application we developed is freely available online and allows both healthcare professionals and patients to explore the accuracy of a diagnostic test across the whole range of threshold values and to explore sensitivity analyses.