University of Twente Student Theses


eHealth for risk screening and early diagnosis : a scoping review on the accuracy and availability of online diagnostic tools

Jovicic, Nikola (2020) eHealth for risk screening and early diagnosis : a scoping review on the accuracy and availability of online diagnostic tools.

This is the latest version of this item.

[img] PDF
Abstract:Introduction: Health applications under the form of symptom checkers or other diagnostic tools have promising implications to not only reduce some of the burdens of the modern health care market, but also in empowering its users in becoming increasingly and more positively involved in their own health care. However, not much is known about the availability of such tools, their diagnostic capabilities and accuracy, and their impact on the behaviors and attitudes of its users. Objective: To conduct a scoping review in order to explore these gaps and to report what is actually known about diagnostic tools across the literature by mapping all known tools and studies in that subject. Methods: A search strategy was devised, and three databases were searched for potential studies: PubMed, Scopus, and PsycInfo. 330 studies were identified and were subject to full-text reviews. We included studies of any design as long as they appraised the accuracy of diagnostic tools available to the general public and provided information on the behavioral impact upon the use of such. Thus, 31 studies were selected for final review. Data was extracted in tables, where the characteristics of the tools and studies were summarized and presented. Results: Three different types of diagnostic tools have been identified: Diagnostic Symptom Checkers, Symptom Checkers with Triage Functions, and Risk Calculators. 80 unique diagnostic tools have been identified, of which 46 had been covered in 3 studies alone. Most diagnostic tools operate on question-based algorithms; however, most studies did not report on either which algorithm it runs on, their validation status, or how many diagnoses they can actually produce. Overall, diagnostic and triage accuracy tends to be poor, with only a few studies showing good rates. However, their accuracy is also extremely variable, which is best demonstrated in studies assessing multiple diagnostic tools where they received the same input before attempting their diagnosis. We have identified a handful of studies on risk calculators, yet only one had its predictive capabilities tested on actual disease incidences. Lastly, the behavioral aspect on the use of diagnostic tools is critically underexplored, with very few studies reporting on changes in behavioral attitudes and actions, such as only half of users actually complying to the advice given by such tools. Conclusion: Almost all diagnostic tools are outperformed by health professionals and their rates of accuracy remain sub-optimal. Considering not only the scarcity of relevant studies in the field and also the evidence, studies need to increase their efforts in clearly outlining the characteristics of their diagnostic tools, assess a broader range of medical condition as well as examine already known tools to more medical conditions, and measure what people actually do with the health information given.
Item Type:Essay (Master)
Faculty:BMS: Behavioural, Management and Social Sciences
Subject:77 psychology
Programme:Psychology MSc (66604)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page