Why and how should we conduct a thorough search for existing mobile health applications before deciding to develop one from scratch

In step with the gradual digital transformation of the healthcare domain, more and more mobile health (mHealth) applications (apps) are being developed and researched to enhance, augment, and extend cardiovascular care. Examples are apps designed to support the primary and secondary prevention of cardiovascular disease, long-term management of hypertension or heart failure, medication adherence, cardiac rehabilitation, exercise therapy in intermittent claudication, or physical activity behaviour in people with cardiovascular disease.¹

The investment required to develop a new working prototype for an app is relatively low. There is, therefore, a risk of unnecessary proliferation and duplication of apps, as new app prototypes are continuously being developed, be it in formally funded projects, by students in educational projects, or by innovative individuals who have coding skills and who perceive the need for a new app.

Healthcare professionals and clinical researchers may find it appealing to develop a new app tailored to their perceived clinical need and local use case. But we argue that it is paramount to first conduct a thorough and systematic search for apps that are already available to minimize the risk of unnecessarily wasting resources (creative energy, digital infrastructure, time, personnel, and financial resources) in creating yet another app. Conducting a search for apps that are ‘out there’ and might be suitable for our purpose mirrors the imperative to conduct a search of the scientific literature prior to embarking on a research project to ensure that the research question has not already been answered by others.

Healthcare professionals and researchers are generally familiar with systematically searching electronic databases for health-related scientific publications, for example to conduct a literature review. But information about already available mHealth apps is not necessarily found in scientific electronic databases, and a search for apps might need to include additional relevant criteria pertaining to the form and function of the app. We therefore need to apply an extended search strategy to give ourselves the best chance of identifying suitable existing apps.

We suggest conducting a search in four steps. First, the desired form and function of the app are articulated. This would be the equivalent to the PICO,² SPIDER,³ or similar approaches for constructing a search strategy for scientific literature, enabling us to define the relevant concepts (target user(s), functions of the app, accessibility features, and technical aspects) and to specify the appropriate search terms. We recommend using everyday language by writing down phrases such as the following:

• ‘The app should be able to…’ (e.g. support cardiac patients in completing their recommended exercise training).

• ‘With the app, patients with hypertension should be able to…’ (e.g. record and self-monitor their blood pressure measurements and their medication intake).

• ‘Healthcare professionals should be able to use the app to…’ (e.g. conduct secure remote video consultations with patients).

• ‘The app should have…’ (e.g. accessibility features such as large text size and screen reader).

• ‘The app should…’ (e.g. work on both Android and Apple operating systems and comply with the European Union General Data Protection Regulation).

These phrases then serve to derive specific search terms and to define inclusion and exclusion criteria for the screening of search results.

Second, the search is conducted. In addition to scientific databases from the healthcare domain and clinical trial registries, we recommend searching an extended set of information sources: scientific databases from the technical domain (e.g. computer science and human computer interaction); national registries for vetted and approved mHealth apps (e.g. the German DiGA-Verzeichnis,⁴ mHealth Belgium,⁵ or Healthdirect Australia⁶); international digital health repositories (e.g. the World Health Organization’s Digital Health Atlas⁷); app stores (Google Play Store, Apple App Store, Microsoft Store, and Amazon App Store); and the World Wide Web (searched using a web browser such as Google Chrome, Mozilla Firefox, and Microsoft Edge). More recently, artificial intelligence chat systems offer another pathway to search for information on the World Wide Web (e.g. Microsoft Copilot, Google Gemini, and Perplexity AI). Additionally, mHealth apps are often developed by—or in collaboration with—organizations and networks such as scientific societies/associations, patient organizations and networks, health professional associations, higher education organizations, and research institutions; and enquiries directed at the appropriate points of contact within such organizations and networks will often provide access to valuable internal knowledge. Enquiries directed at individual experts in the field (researchers, clinicians, and app developers) will often yield information that is not easily found or not accessible in the public domain.

Third, the search is documented. This is important for efficiency, as it can be easy to lose oversight of which information sources have already been searched and which results have been obtained from which information source. For those seeking funding for a project, a well-documented search can provide crucial evidence and reassurance to the funding agency or other decision-makers that the investment is justified by demonstrating either that an mHealth app that meets the specific needs is (most likely) not available yet or that the best available candidate app has been identified in a thorough and systematic search. We recommend the use of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) checklist extension for searching (PRISMA-S), which supports comprehensive reporting of literature searches.⁸ Checklist items 4 (searching online resources and browsing), 6 (contacting authors, experts, manufacturers, or others), and 7 (any additional information or sources used) can be used to document information sources other than electronic databases of scientific literature. While it is relatively straightforward to document complex and reproducible search strings for electronic databases, search strategies for other information sources might need to be recorded differently. For example, searches conducted on the World Wide Web yield millions of results and are best documented by recording the search terms and the number of consecutive search results or pages that were screened. Screenshots or printouts of websites can be helpful to document search results that cannot be conveniently exported in bulk.

Fourth, the search results are screened against inclusion and exclusion criteria. Similar to the process of screening scientific literature (first titles and abstracts, then full texts), we suggest to first screen based on the information available online, which could include short descriptions of an app but also detailed reports, user manuals, advertisements, and user reviews. Potentially suitable candidate apps should then be downloaded, installed, and trialed to confirm that they meet the inclusion criteria. Moreover, it is important to consider technical aspects, regulatory implications, the scientific evidence base, and transferability to the intended geographic, cultural, clinical, or healthcare policy context. Technical and regulatory considerations require an assessment against relevant standards and requirements. The scientific evidence base requires critical appraisal of research reports. Transferability can be explored during a testing period with healthcare professionals and patients who are representative of the intended user groups.

In conclusion, a thorough search and appraisal of existing mHealth apps is good practice and contributes to reducing waste in research and in evidence-based clinical practice. We hope that our suggestions provide healthcare professionals and clinical researchers with helpful guidance on how to conduct a comprehensive and systematic search for an app.

Data availability

There are no new data associated with this article.

References

Author notes