Rising Trends, Gaps in Research Observed in Review of Digital Health Intervention Studies

Published on: 

This systematic review highlighted the rise in studies on artificial intelligence and teledermatology for skin cancer, as well as other gaps in related research.

There is a rising trend of digital health intervention (DHI) studies on store-and-forward (S&F) teledermatology and artificial intelligence applications for skin cancer, according to new findings, though gaps exist in research on DHI for chronic skin issues such as psoriasis and atopic dermatitis.1

These new findings and more were the results of a recent systematic review of existing literature regarding research on DHIs in the field of dermatology, with the investigators expressly looking for trends and gaps in research.

This latest research was authored by Patrick Reinders, from the Institute for Health Services Research in Dermatology and Nursing (IVDP) at the University Medical Center Hamburg-Eppendorf in Germany.

These investigators noted some strong evidence on artificial intelligence applications designed to help classify skin cancer current research, adding that broader conclusions had not yet been fully established.2

“However, a general and comprehensive overview of studies on DHIs in dermatology does not exist,” Reinders and colleagues wrote. “Consequently, this systematic mapping of the literature characterizes and clusters studies on DHIs in dermatology regarding the type of DHI, indication, aim, study country, outcome parameter, target group and study design.”

Background and Findings

The investigators selected studies which followed several specific inclusion as well as exclusion criteria. Publications that had been made prior to 2010 were excluded by the investigators due to the disparity between older technologies and the existing technological standards.

The inclusion criteria used by the research team encompassed research that utilized quantitative methods in order to examine DHIs in the field of dermatology. Studies which centered around artificial intelligence (AI) were kept by the team if they had compared the AI algorithm to standard care practices, assessed AI prospectively in real-world types of scenarios, or used market-available applications, excluding those which mainly focused on AI's mathematical development.

The inclusion criteria the investigators used also allowed for research featuring DHIs aligned with the World Health Organization's classification and relating to the topic of the dermatology field. In addition, the studies had to have been original, and they had to be peer-reviewed and published in English or German in the time between August of 2010 and August of 2022.

Furthermore, the research team’s criteria for exclusion would be if research turned out to be case reports or series, if the studies mainly aimed at developmental aspects and had minimal quantitative evaluation, and if the studies on AI strongly emphasized the process of mathematical development.

Two examiners were used by the team to assess all abstracts and titles, and in specific situations of disparity or uncertainty, their records were deliberated with a third evaluator until a unanimous decision was able to be made. Subsequently, the residual complete texts underwent evaluation for their suitability by a minimum of 2 reviewers, with disagreements leading to discussions to determine the studies’ eligibility.

The investigators ended up establishing the following outcome parameters for the included studies:

  • Effectiveness or efficacy was noted when the study assessed the potential advantages of a DHI either in ideal circumstances or real-world scenarios, compared to the standard care. This parameter encompassed factors like indices for educational intervention test scores, quality of life, morbidity, and habitual indices including adherence to treatment.
  • Diagnostic performance encompassed sensitivity, specificity, accuracy, and associated scores—an example being F1 score—along with the diagnostic agreement between the standard care and the DHI.
  • The DHI’s efficiency was taken into account by the team for metrics gauging the effects of an intervention specifically with regard to the resources invested. This element of analysis was only usable with studies which were comparative.
  • Usability of DHIs was examined in alignment with the International Organization for Standardization (ISO) (2013) definition: 'the degree to which a product can be employed by designated users to attain objectives effectively, efficiently, and with satisfaction.'
  • Regarding acceptance, a comprehensive interpretation was adopted by the investigators, which used aspects including attitudes surrounding new DHIs, willingness to take part in an intervention, actual use of DHI, and satisfaction following engagement with such technology.

From the initial 12,009 records the research team found in MEDLINE searches, their final analysis ended up with 403 studies in total, primarily made in Western nations such as the US with 133 studies, Germany with 32 studies, and Spain with 23 studies. There were a diverse set of DHIs found, and among them 261 were shown to be focused on healthcare providers, 66 on clients such as caregivers, patients, and healthy individuals, and 67 focused on both client and provider relations.

Among the investigators’ interventions they were able to find, 254 were shown to be aimed at diagnosis of several different health issues, while S&F teledermatology—with 187 studies—and artificial intelligence applications for image analysis—with 65 studies—stood out as major study categories. In the area of client-oriented DHIs, 31 of the identified studies were focused on supporting changes in health behavior, and 77 focused on monitoring health statuses of clients.

Interestingly, the most common targets of these studies were shown to be DHIs with a focus on skin cancer with 148 studies, wounds with 29, and psoriasis with 29. The team’s identified research mainly examined diagnostic performance with 166 studies, DHI acceptance with 92, and effectiveness with 98, whereas usability and efficiency were shown to have gotten comparatively less attention.

Overall, while DHIs in the field of dermatology have mostly emphasized AI and teledermatology use, particularly relating to skin cancer, the analysis found a wide range of interventions which aimed at different groups of users, different research objectives, and different medical indications.

“Nevertheless, to fully comprehend the capabilities of DHIs, additional research is necessary in promising areas such as the management and monitoring of chronic skin diseases and triage of patients,” they wrote. “Additionally, the assessment of DHIs for clients is essential, as these technologies are widely available, yet they are not supported by a large evidence base.”


  1. Reinders, P, Augustin, M, Kirsten, N, Fleyder, A, Otten, M. Digital health interventions in dermatology—Mapping technology and study parameters of systematically identified publications. J Eur Acad Dermatol Venereol. 2023; 00: 1–10.
  2. Haggenmüller S, Maron RC, Hekler A, Utikal JS, Barata C, Barnhill RL, et al. Skin cancer classification via convolutional neural networks: systematic review of studies involving human experts. Eur J Cancer. 2021; 156: 202–216.