Introduction

Recent developments in health technology have positively affected multiple and essential sectors of the economy, especially the healthcare sector, by providing solutions that guarantee the exchange of medical knowledge and information and establish long-lasting health outcomes1,2. Digital health technologies, such as wearables devices, computerized decision support systems, and telemedicine improve the technical performance and satisfaction of healthcare employees, demonstrate potential to decrease direct and indirect costs of medical services, and enhance the quality of delivered care3. Worldwide, using digital solutions in practice seems inevitable, with modality-specific prevalence (e.g., 50.8% for telemedicine, 89.9% for electronic health records, and 91.9% for social media platforms)4,5,6. However, the prevalence of use might be even higher, as no previous study has collated and assessed the overall prevalence of using digital health technologies by healthcare providers. Likewise, several studies have suggested that ethnicity, race, geographic location, age, and medical specialty directly interfere in the adoption of technology use, evidencing the importance of understanding variables accounting for the digital divide and disparity of access7,8,9.

Several barriers to healthcare’s overall quality, transparency, and efficiency naturally arise during or following the creation, implementation, and maintenance of digital health technologies. Therefore, during the design of any health-related project, it is essential to identify and quanti-qualitatively analyze its risks and facilitators, enhancing the likelihood of obtaining favorable outcomes and optimizing the chances of success. The efficient implementation of digital technologies, characterized by proper implementation of a systematic management approach, including strategic planning, resource allocation, and control and evaluation processes, is fundamental to refining healthcare services, equipment, and technologies10,11,12. In reaction to these aforementioned elements, multiple efforts have strengthened healthcare systems through employing DHTs for healthcare professionals and stakeholders from low-, middle-, and high-income countries. For instance, the World Health Organization (WHO) endorsed in the 73rd World Health Assembly the institution of the Global Strategy on Digital Health 2020–2025, in which four guiding principles rely on the acknowledgment that the institutionalization of digital health in a national system requires a decision and commitment by countries, recognition that successful digital technologies require an integrated strategy, promotion of the appropriate use of digital interventions for health, and recognition of the urgent need to address the major impediments faced by least-developed countries implementing digital health technologies13. Furthermore, the Regional Digital Health Action Plan for the WHO European Region 2023–2030 has a critical regional focus area on strengthening digital literacy skills and capacity-building in the general population, with particular attention to the health workforce, for the use of digital health services and disease prevention and management14. Due to these global actions, numerous studies have focused on assessing barriers to and facilitators for many technologies15,16,17.

To date, hundreds of clinical trials based on specific technologies applied to the healthcare professionals’ environments have assessed the implementation of digital interventions in the healthcare system, while several systematic reviews have combined these publications, evidencing their effectiveness, safety, and feasibility. However, a summary of enablers and restraints to healthcare professionals’ coordinated and integrated use of digital health technologies has not been published yet, making the current evidence dispersed, misused, and overlooked. Therefore, in this overview of systematic reviews and semantic-based occurrence meta-analysis, we report all published evidence from existing systematic reviews covering and mentioning barriers and facilitators to the solid use of digital health technologies by healthcare providers.

Results

Study selection and characteristics

Our database and PROSPERO search are shown in Fig. 1. Our January 21, 2022 search retrieved 9,912 records, of which 139 underwent full-text review (Fig. 1, section A). Based on the inclusion and exclusion criteria, 47 studies and seven ongoing studies were included. On March 1, 2023, 2,717 new publications were identified through an additional database search (Fig. 1, section B). Of those, 142 studies were shortlisted for full-text assessment, and 60 reviews were added to our umbrella review. Two additional ongoing studies or protocols were identified. In total, this overview of systematic reviews included 108 primary systematic reviews and nine ongoing studies (Fig. 1, section C).18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,87,88,89,90,91,92,93,94,95,96,97,98,99,100,101,102,103,104,105,106,107,108,109,110,111,112,113,114,115,116,117,118,119,120,121,122,123,124,125 One study was identified from alternative resources.64 Justification for the exclusion of 165 studies is presented in Supplementary Information 1 (pp 2–7). Included study characteristics are characterized in Table 1 and Table 2. One study is pending classification as it required translation. No additional data needed to be requested from the corresponding authors.

Fig. 1: PRISMA flow chart diagram.
figure 1

Reason 1—wrong intervention or platform was unclear. Reason 2—the study did not provide any relevant outcome influencing healthcare providers. Reason 3—targeted population was not healthcare providers. Reason 4—study design used did not match our inclusion criteria.

Full size image
Table 1 Main characteristics of included studies evaluating the impact of digital health solutions on health workers (n = 108).
Full size table
Table 2 Population being evaluated, studies’ methodologies and technologies being evaluated.
Full size table

Few studies (n = 20; 18.5%) initially targeted evaluating the creation, implementation, long-lasting use, and self-reported barriers and facilitators to using digital health technologies by healthcare professionals25,27,29,43,45,51,66,68,70,72,73,74,82,86,93,96,98,101,107,120. Thus, the remaining reviews were cautiously evaluated in order to identify a report of any barrier or facilitator to using digital health technologies by healthcare workers. Included reviews were heterogeneous in terms of the digital health technologies being assessed (e.g., alert systems, clinical reminders applications, computerized clinical decision support systems, electronic documentation systems, mobile health applications, social media platforms, and telemedicine tools) and enrolling different healthcare professionals (e.g., general practitioners and specialists, nurses, pharmacists, community healthcare workers) at several levels of care (primary, secondary, and tertiary health facilities).

Most reviews (n = 63; 58.3%) were executed in North America, Europe (n = 61; 56.4%), and Asia (n = 50; 46.2%). Thirty-three reviews suggested barriers and facilitators in the African territory (30.5%), while 28 reported data from Latin American and Caribbean regions (25.9%). Our study involved reviews from low- (e.g., Kenya, Rwanda, Uganda, and Ghana), middle- (e.g., Brazil, China, Russia, South Africa, and India), and high-income countries (e.g., Japan, the Czech Republic, United States of America, and Australia).

According to our bibliometric analysis, our data were classified into five clusters based on identifier clustering assessment, and recorded keywords by co-occurrence frequency are shown in Table 3 and Fig. 2. The ten most common identifiers were “healthcare professionals,” “technology,” “review,” “barrier,” “care,” “systematic review,” “factor,” “patient,” and “implementation”.

Table 3 Top author-provided identifiers among included reviews.
Full size table
Fig. 2: Overview of the network map of the most frequently identified terms among included studies.
figure 2

Please note that in the network visualization, items are represented by their label and by default also by a rectangles. The size of the label and the circle of an item is determined by the weight of the item. The higher the weight of an item, the larger the label and the circle of the item. The color of an item is determined by the cluster to which the item belongs.

Full size image

Taking into account the 37 (34.2%) records providing data regarding the number of healthcare professionals considered in primary studies, sample sizes ranged from 22 to 106,876 (totaling approximately 345,000 healthcare workers), with a mean of 3,197 (SD 12,364), and a median of 1,545 (IQR 258 to 9,016). Most studies did not precisely consider one medical specialty, disease, or condition. However, some reviews focused on diseases of the respiratory system (e.g., tuberculosis, asthma, and chronic pulmonary obstructive disease)19,22,31,32,46,93,101,123, pregnancy, childbirth, or puerperium (e.g., maternal health, postpartum hemorrhage, and reproductive health)19,22,23,26,31,35,46,56,61,77,94, certain infectious or parasitic diseases (e.g., malaria, human immunodeficiency virus infection, and influenza)19,22,23,28,31,46,50,61, endocrine, nutritional, or metabolic diseases (e.g., diabetes mellitus)57,64,76,93,123, mental and behavioral disorders (e.g., post-traumatic disorder syndrome, stress, depression, and burnout)23,41,44,64,70,76,94,125, neoplasms50,67,85,123, diseases of the circulatory system (e.g., hypertension)19,25,48,50,57,67,93,123, diseases of the blood or blood-forming organs (e.g., anemia)22, and diseases or disorders of orofacial complex (e.g., oral lesions)28,42. Identified reviews mostly included quantitative (randomized and non-randomized trials, surveys, economic analysis, structured questionnaires, and experimental studies), qualitative (e.g., non-structured interviews, literature reviews, focus groups, observation, and cultural reports), and mixed-method reviews (sequential exploratory and concurrent transformative studies). An additional description of included reviews is shown in Table 2.

Barriers and facilitators identified in included reviews and potential recommendations

The final domains created based on the thematic analysis can be accessed in Figs. 3, 4, and the summary of findings of the top seven barriers and facilitators can be accessed in Table 4. Our linguistic and semantic-based analysis stratified the data into 21 barriers and 19 recommendations. Predominant barriers were associated with infrastructure and technical (RFO of 6.4% [95% CI 2.9–14.1]), personal and psychological barriers (RFO of 5.3% [95% CI 2.2–12.7]), time and workload-related (RFO of 3.9% [95% CI 1.5–10.1]), training and educational (RFO of 3.4% [95% CI 1.3–8.9]), and legal- and ethical-related factors (RFO of 3.6% [95% CI 1.3–9.6]). Most predominant enablers related to the offer of training and educational activities (RFO of 3.8% [95% CI 1.6–9.0]), healthcare provider perception of digital health technologies usefulness and willingness to use (RFO of 3.8 % [95% CI 1.8–7.9]), the existence of government and multisector incentives (RFO of 3.0% [95% CI 1.4–6.6]), adherence promotion campaigns (RFO of 2.2% [95% CI 1.1–4.3]), involvement of healthcare providers in the process of digital health technologies development and implementation (RFO of 2.0% [95% CI 0.8–4.9]), and intuitive navigation in healthcare technology systems (RFO of 1.9% [95% CI 0.7–5.2]).

Fig. 3: Relative frequency meta-analysis of most reported barriers for the use of digital health technologies by healthcare professionals.
figure 3

Frequencies (expressed as % and their confidence interval) are distributed among each categorized barriers as well as by healthcare technology modality.

Full size image
Fig. 4: Relative frequency meta-analysis of most reported facilitators for the use of digital health technologies by healthcare professionals.
figure 4

Frequencies (expressed as % and their confidence interval) are distributed among each categorized facilitators as well as by healthcare technology modality.

Full size image
Table 4 Summary of qualitative findings.
Full size table

As represented in Figs. 3, 4, several semantic clusters were described throughout included reviews. Herein, we outline and exemplify the five most common barriers and facilitators to the design, implementation, longitudinal maintenance, and evaluation of digital health technologies by healthcare professionals. The remaining barriers and facilitators are explained in detail in Supplementary Information 2 (pp 8). Infrastructure and technical barriers were the most frequently described barriers among included reviews, relating to issues with a limited or insufficient network, lack of existing technologies, lack of devices, compatibility with daily workflow, connectivity speed, healthcare capacity of technology integration, interconnectedness, absence of standardized/harmonized systems at different facilities, limited access to electricity, and requirement of a functional database system or large disk space. Notably, technical issues seem to be the worst in rural and countryside regions. Firstly, counteracting connectivity-related barriers involves ensuring availability (especially in rural areas) and affordability, guaranteeing high-speed fiber connectivity, and increasing the number of reliable local networks. In addition, we found reviews suggesting that to overcome infrastructure and technical barriers, the involvement of healthcare professionals in developing and implementing any health technology tools is fundamental, enhancing their capacity to manage such applications and increase their independence from co-workers and support centers. Remarkably, all reviews stated that user engagement and collaboration with system developers or associated stakeholders is crucial in all design and development stages, deployment, and continued utilization, as created applications are fit for purpose, based on understanding and addressing healthcare providers’ needs and expectations.

Personal and psychological barriers involved complex thematic components, including the healthcare professionals’ resistance to change, difficulty understanding the technology, perception of less human interaction, technophobia, ages, education levels, professional experience, low literacy, poor writing skills, linguistic features, adherence behavior, and fear of using particular health technology. Moreover, unwillingness, low expectations, skepticism from healthcare providers, and low motivation for compliance were also associated with personal barriers. For counterbalancing these barriers, healthcare professionals’ perception of usefulness and willingness was a highly cited facilitator, characterized by the degree to which the employees believe that using specific digital health technologies would enhance their performance and the proportion of participants intending to utilize that technology. Furthermore, personal and psychological barriers could be addressed by using and adopting training programs and educational activities appropriately tailored to healthcare professionals’ needs and coverage of deficient abilities. High-quality, real-time technical support and coaching also appeared as a component that increased healthcare providers’ efficiency, decreased implementation fear, and potentially could reduce internal conflicts during system adoption. Importantly, training programs may be developed with the ongoing involvement of the intended community to understand their needs and knowledge gaps. Moreover, evidence shows that user-friendly design, intuitive system navigation, and easy-to-use interfaces are critical to improving overall product performance and facilitating data collection and input, data processing, and further analysis.

Some reviews suggested that the limiting factors for the broad use of digital health technologies are associated with healthcare workers’ concerns about increased workload and altered workflow, which could hinder the sustainability of the digital health technologies. Additionally, these newly implemented technologies would require additional purchase time and increased set-up, implementation, training, access, adaptation, and establishment stages. In addition, healthcare professionals commonly stressed that digital health technologies would impact the quality of delivered care, as recently trained professionals would need a longer time to convert acquired data into the implemented system. However, although time might be required to acquire the right skills and operating competencies, with adequate training, continuous technical support, and peer-to-peer collaboration, threats associated with increased time to complete a specific task are significantly reduced. Useful written guidelines, instructions, and handouts appear to be important facilitators that could be easily implemented73. Likewise, incentives from government agencies and multisectoral organizations were shown to significantly improve digital health technologies’ effectiveness and chances of success in large-scale healthcare systems. Therefore, this conceptual perspective should be shown to healthcare providers, as increased effectiveness is directly related to the appropriate use of time and less wasteful processes.

Fourth, legal- and ethical-related barriers were shown to be a relevant factor for healthcare providers, as privacy and security concerns, national legislation, jurisdiction, and the existence of unclear legal liability regarding response protocols would directly affect healthcare professionals. Possible interventions for these barriers are associated with the development of safer data storage systems, the establishment of requirements on safety and security in cooperation with healthcare professionals and patients, or the creation of an international legal framework and legislative norm, which would clarify security regulation policies that could help ensure patients’ privacy and confidentiality, as well as define healthcare professionals’ liabilities.

Lastly, deficient or inexistent training and educational activities were evidenced to significantly impact the success and efficiency of digital health technologies in the healthcare environment . Some reviews highlighted that without training, healthcare providers tend to feel low self-efficacy when utilizing any digital health technologies, resulting in negative attitudes toward these technologies. In addition, as evidenced by healthcare workers, prior technology introduction, vendor training, in-depth seminars, workshops, or correlated training activities are unusual, and regular quality process assessment following implementation to ensure efficiency are also rare. Interestingly, reviews not only highlighted that training was fundamental to the success of using digital health technologies but also suggested that training per se would also be delivered through certain digital health technologies, such as mobile technologies and computers. Thus, the training offer positively affects healthcare professionals’ experience with digital health technologies, especially when monetary incentives are added to this variable, given the time invested in obtaining the proper abilities to operate any digital health technologies.

Using the AMSTAR 2 methodological quality assessment tool, most reviews had a very critically low overall methodological quality, as shown in Table 5. Nine-nine reviews were classified as very low quality, six as low quality, and only three were rated to have a high methodological quality. Two top-ranked reporting inadequacies related to the lack of evaluating the presence and likely impact of publication bias (95.2%), and the disregard of the risk of bias when interpreting the results of the review (95.2%). Where judgment was lost, this generally associated with the lack of prior protocol (50.9%), absence of justification for excluding individual studies (88.8%), lack of risk of bias assessment from individual studies being included in the review (63.8%).

Table 5 Quality assessment rating of systematic reviews included in the digital health solutions applied to healthcare workers environment overview.
Full size table

We mapped the aforementioned data and complementary results, as shown in Fig. 5 (also available for virtual access through the GitMind platform).126 As evidenced in supplementary information 3 (pp 9), we found several terms with similar semantic structures. Thus, we coded each barrier or facilitator and identified recommendations, suggesting the possibility of a complex and broad linguistic connection and relationship amongst codes. These thematic relationships are not limited in our analysis and can be explored and exhausted.

Fig. 5
figure 5

Conceptual map of reported barriers and potential facilitators and recommendations to overcome these barriers.

Full size image

Discussion

To our knowledge, this is the first overview of systematic reviews to collate, cluster, and synthesize the quantitative, qualitative, and mixed methods body of literature associated with barriers and facilitators to and use of several digital health technologies by healthcare professionals at all levels of care. The decision for carrying out this valuable, but complex study, relies on the noticeable detachment of research data and investigation groups in the field of Medical Informatics, who usually inadvertently duplicate technical and financial resources given the existing gaps in the literature. Here we report 21 overarching barriers and 19 facilitators, mostly interconnected, containing a complex sequence of thematic describers and identifiers. Understanding and overcoming identified barriers to the fully integrated and coordinated use of DHTs by any class of healthcare providers and evaluating its facilitators could positively impact successful creation, implementation, adoption, training, and long-term services or product utilization.

The evidence suggests that healthcare providers and managers predominantly face infrastructure, technical-, training-, legal-, ethics-, time-, and workload-related barriers to using digital health technologies, regardless of the level of care or digital technology. In the second level of semantic occurrence, several restraining factors to the wide use of digital health technologies were combined and reported, including psychological and personal barriers, lack of supervisory support, ownership issues, and healthcare system-cultural-, social-, and financial-related limiting features. Nevertheless, we are aware that some of the classified items are interconnected, meaning that the prevalence of occurrence ranking should not be used as a priority guide for policymakers and health organizations when addressing these barriers. For instance, the highlighted barrier “81B” (regarding the simplicity of contents usually transferred in mobile applications or clinical alert systems) might be directly related (or potentially caused due to) to the technical limitations per se (considering devices screen’s reduced size (“2B”), the complexity of the systems themselves and the information they carry (“5B”), or even because the lack of standardization and customizability of such systems and technologies (“7B”). Therefore, the creation of artificial intelligence-based mind mapping representing these interconnections is of utmost relevance126.

Creating and applying digital health technologies to healthcare environments must be driven by a regime of comprehensive assumptions instead of empirical models and processes. Our results corroborate with published systematic reviews that have already evidenced patient-reported barriers and facilitators to utilizing digital health solutions for self-care127,128,129. For instance, self-management of low-back pain using mobile health applications was mainly challenging due to information technology, usability-accessibility, quality-quantity of content, tailoring-personalization, and motivation-support barriers127. In contrast, flexibly structured and intuitive navigation, trustworthy content and sources, content accounting for individual needs and priorities, and the opportunity to influence the application design appeared as relevant facilitators affecting the uptake and utilization of digital health interventions for self-management of lower back pain127. Likewise, Powell and colleagues suggested that a lack of awareness, self-motivation, training, privacy, and security concerns are the most common patient-derived barriers to using electronic portals128. Emphasized facilitators correlated with use engagement by a leader (i.e., physician), free access and control over health information, and an adequate communication profile. Therefore, as the relationships between our identified barriers and facilitators and existing patient-related evidence highlight, the development of digital healthcare solutions should consider multiple factors, which can facilitate or deteriorate broad goals of high-quality use of information technology in the healthcare environment.

During protocol modeling, our research group discussed the possibility of including reviews that summarize evidence on barriers and facilitators involving students in health fields. The decision was not to include these reviews because these students are not yet legally considered professionals or critically necessary workforce, and they are not considered essential in healthcare settings130,131. However, one aspect found in these excluded reviews was revealed in our overview with significant frequent and relevant findings: the use of digital health technologies for training and educational purposes. Although distance education dates from 1728132,133, e-learning or virtual learning started during the early 1980s at the University of Toronto134 and has been developing ever since, particularly during the COVID-19 pandemic135,136. Currently, several high-income countries, such as New Zealand and the United States of America, have already integrated and implemented the Information and Communication Technology constructivist learning model in their national or statewide policies, ensuring that students have the chance to become digitally competent citizens137,138. These actions effectively decrease multiple barriers observed related to limited or no computer skills, restricted knowledge and technology literacy, and lack of reliability in technological tools. However, it has been suggested that numerous low- and middle-income countries still struggle with device acquisition, connectivity issues, tutors’ level of expertise and lack of motivation, absence of basic infrastructure, and the unwillingness of the government to implement such solutions129.

Foremost, we chose only six health solutions as systematic and feasible choices for comprehensive data processing. Nevertheless, we observed additional modalities of health solutions being implemented worldwide (e.g., laboratory and radiology automatic reporting systems, picture archiving and communication systems, cloud-based systems, and advanced and business analytics), and our synthesis may miss emerging or recent technologies52,74,114. For instance, studies have suggested that electronic laboratory reporting systems not only improve surveillance for notifiable conditions but can also be helpful in real-time laboratory testing in emergency departments and significantly improve organizational framework and efficiency139,140. Correspondingly, cloud-based computing systems have been increasingly applied in the healthcare system to ensure secure storage, handling, and processing of medical information141. Regardless of the digital health solution being implemented and utilized, healthcare workers and patients benefit from it. By improving real-time patient access to their results and providing better patient involvement with care, the incidence of unwanted tests or extra prescriptions decreases, and the overall quality of care is subsequently enhanced142,143.

We observed a limited number of reviews assessing the potential challenges and enablers for artificial intelligence models, machine learning algorithms, and platforms utilizing features such as augmented reality40,54,63,70,78,85,94,99. However, although the restricted number of studies assessing these subgroups in the field of digital technologies, core barriers and facilitators remained like other subgroups. Nevertheless, we highlight the need for further research with these technologies, as alternative barriers and facilitators would arise.

Due to the wide variety of digital health technologies currently being used in several medical specialties and levels of care, we had to restrict our report in different ways, limiting our certainty of evidence. Similarly, our series of analyses did not consider the existence of subgroup singularities by type of healthcare professional. As suggested in our map based on bibliometric data, only physicians, community health workers, and nurses appeared as recurrent keywords among all studies within the 42 systematic reviews eligible for inclusion. Therefore, studies analyzing impeding and enabling factors to the general use of digital health technologies in other healthcare providers (e.g., pharmacists, physiotherapists, physical educators, speech therapists, healthcare governmental agents, biologists, social services agents, healthcare managers, dentists, and psychologists) cause a “professional class bias” event that should be addressed in future studies. Likewise, factors like age, racial group, gender, country income index, or geographic location could affect a different subgroup (e.g., potential higher reporting of barriers of professionals practicing in low- or middle-income countries would focus more on technical and infrastructure features). Moreover, we neglected that digital health technologies utilized in the healthcare environment are usually concomitant and integrated. Thus, we may have considered the reported health solution independently instead of using a translational and adapted assignment methodology. Therefore, the provided RFO represented only the tendency of domain observance and reporting and not the identical picture of healthcare professionals’ reality. To conclude, we are aware that some highlighted barriers and facilitators could be assigned to a broader subtheme (e.g., lack of supervisory support in training and educational skills). However, during the overall execution, we observed that some terminologies and coding were commonly reported separately, so we decided to maintain them as individual elements to ensure the representativeness of the findings. Interestingly, the use of the AMSTAR 2 tool for evaluating the methodological quality of all included reviews should also be stated as a limitation, as the approach was primarily intended to systematic reviews of randomized controlled trials. Nevertheless, as most AMSTAR domains are on the elements that any review is structured (e.g., search strategy, protocol, extraction, combing studies, and publication bias), we believe that applying this methodology to our include reviews do not hinder the observed results. Likewise, although we Apart from these minor methodological limitations, the major strength of our study is the strict adhesion to international guidelines for reporting of systematic reviews (e.g., Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement and the Cochrane Handbook of Systematic Reviews and Meta-Analyses) and the execution of the entire study with international and blinded collaboration. We acknowledge that more than one methodology for evaluating the certainty of the evidence in qualitative research exists. We applied the GRADE CERQual method to check the overall quality of evidence for the seven most-reported barriers and facilitators. Generally, the evidence quality is high, with all considered domains without major concerns but with methodological limitations. We judged this domain as a moderate concern based on the phenomena of interest, adequate data collection and extraction, and quality in reporting observed data. In addition, expert groups have been discussing.

Although digital health technologies and their numerous types of technologies positively affect the healthcare environment, barriers impacting the successful creation, adoption, implementation, and sustainability of digital interventions are commonly reported by healthcare workers. Notwithstanding, the identification and deployment of different enabling factors allow the utilization of digital technologies in a holistic and integrated way. This overview of reviews emphasizes remarkable limiting features that should be considered by all stakeholders and provides advice to overcome these issues, with the expectation of increasing professional satisfaction and, perhaps, the quality of delivered care.

Methods

This overview of systematic and scoping review (herein referred to as “overview”) protocol was registered on PROSPERO (CRD42022304372, supplementary information 4, pp 10–20) and it was part of a broader study conducted by the Data and Digital Health Unit of the Division of Country Health Policies and Systems of the World Health Organization, Regional Office for Europe3. This initiative provides strategic direction, technical assistance, and tailored support to countries and policymakers to strengthen their capacity to generate timely, credible, reliable, and actionable health-related data. The scientific community is currently defining an explicit, systematic, and transparent methodology to create evidence- and agreement-based reporting guidelines for overviews of reviews144. Therefore, we used the Preferred Reporting Items for Systematic Reviews and Meta-Analysis reporting recommendations145, the Cochrane Handbook guidelines146, and reports published by Fusar-Poli et al.147 and Cornell et al.148 guiding the practice on how to effectively conduct an umbrella review. As our study relies upon secondary data, ethics approval was waived. It is worthwhile mentioning that although in our protocol we initially stated that a standard meta-analysis would not be carried out, we decided to mathematically evaluate the obtained results. The technique utilized for the word- and sentence-based assessment (particularly associated with discourse analysis) is a well-known summarizing strategy used in the field of Human Sciences and was systematically presented and implemented in our research team after the protocol preparation. Therefore, in consonance with the requirements of continuous scientific evolvement and improvement, we decided to apply this newly introduced technique. However, this deviation does not alter the core of this project.

Data sources and searches

We searched five databases (Cochrane Database of Systematic Reviews, Embase®, Epistemonikos, MEDLINE®, and Scopus) and the PROSPERO protocol registration platform from inception to Jan 23, 2022, for systematic and scoping reviews evaluating barriers and facilitators to using digital health technologies by healthcare professionals worldwide. We also performed a manual search of reference lists of reviews shortlisted for full-text review and planned to contact the authors of included review to retrieve additional data.

An experienced information specialist and the expert team tailored search strategies to each database using Medical Subject Headings (MeSH) and free-text identifiers associated with the research topic149,150,151,152. The search included three main categories of key terms. Digital health technologies search identifiers included terms such as “telemedicine,” “telehealth,” “mobile health,” “mHealth,” “artificial intelligence,” “machine learning,” “social media,” “natural language processing,” and “computer decision support systems,” healthcare professional-related terms included “healthcare worker,” “healthcare provider,” and “healthcare support worker,” and systematic review filters used were “systematic review,” “meta-analysis,” and “scoping review.” Our terms are defined in recently published studies in the World Health Organization guidelines on digital health technologies for strengthening health systems, the World Assembly Resolution on Digital Health, and The Lancet Digital Health. In supplementary information 5 (pp 21-28), we present the detailed search strategy for the databases.

Study selection

Eligibility was evaluated by two independent investigators who primarily screened titles and abstracts and subsequently reviewed the full texts using Covidence® (Veritas Health Innovation, Melbourne, Australia)153. Systematic and scoping reviews deemed eligible must have used at least two databases for their assessment, should have described the search methods, and evidenced the use of a transparent methodology for study selection and data extraction. Moreover, these reviews were only included if a qualitative analysis of barriers and facilitators to using digital health technologies by healthcare providers was clearly noted. We did not place limits on targeted healthcare professionals, medical specialty, level of care, language, and publication date. However, in order to avoid bias and results inflation, those studies strictly prioritizing the assessment of digital technologies for students and education in the field of health sciences were excluded.

Data extraction and quality assessment

Two independent researchers appraised the methodological quality of included systematic reviews using the AMSTAR-2 tool154. Following the initial evaluation, a third researcher cross-checked rated domains. The methodological quality of reviews was classified as “critically low,” “low,” “moderate,” and “high.” Our research team is aware that the AMSTAR 2 tool is not intended to generate an overall score of the review’s quality. Thus, we emphasize that we considered the appraisal methodology holistically, mostly related to the provision of an extensive evaluation of quality, particularly weaknesses associated with poor conduct of the review or word counting limitation endorsed by a determined journal.

Relevant data (first author identification, publication year, published journal, number of included databases, review objectives, primary study design, type of healthcare professional, type of digital technologies being analyzed, number of included primary studies, and barriers, facilitators, and recommendations for using digital health technologies) was extracted from included reviews by two independent researchers using Microsoft Excel (Microsoft Corporation, Redmond, USA)155. In the second stage, four independent volunteer collaborators reassessed extracted data to resolve inconsistencies.

Data synthesis and analysis

We used VOSviewer to assess research hotspots associated with digital health technologies based on the principle of co-occurrence analysis156. The minimum number of co-occurrences was set as 3, normalization method as an association, random starts as 1, random seed as 0, resolution as 1, and we merged small clusters. We attempted to clean the network map as much as possible, as some keywords were not meaningful. Thus, we extracted data from the top 100 author-provided keywords and mapped them into a single keyword co-existing network. Representative and frequent terms are expressed as larger nodes, and the thickness of the link between two or more nodes represents the strength of the relationships between them.

Our findings were evaluated and collated using an adapted version of a thematic synthesis developed by Thomas and Harden157. The 21 domains prioritized in the Enhancing Transparency in Reporting the Synthesis of Qualitative Research (ENTREQ) statement were followed158. First, qualitative data of included reviews on the main barriers and facilitators identified were coded line-by-line using QSR’s NVivo software (QSR International, Burlington, USA)159. In addition, primary highlighted concepts were re-evaluated by four volunteer collaborators who double-checked selected data and evaluated extraction errors or missing information. If needed, they also created new in-text selections. Furthermore, we organized free selections into similar themes to combine the preliminary results into descriptive themes. Lastly, we developed analytical themes that summarized barriers and facilitators closely related to the original remarks reported in included reviews. The explanatory delineation of thematic barriers and facilitators was a dynamic, deductive, and intuitive process, as different review authors had their peculiarities in academic and text writing. The alignment of thematic barriers and facilitators was discussed by all authors, resulting in the development of recommendations. In the result section, we have identified only the five most frequent barriers and facilitators. Recommendations were also emphasized for these five features. However, a complete list of barriers, facilitators, and recommendations can be accessed in supplementary information 2 (2.1 and 2.2). Where homogenous barriers were recognized (e.g., lack of leadership and local champions), guidance to overcome these barriers were prepared by the group of specialists (e.g., identification of processes weaknesses, implementation of improved strategies, and adjustment of progress based on stakeholder feedback). Similarly, the recommendations also considered the identified facilitators. Systematic reviews with similar research questions were expected to be included in our umbrella review. Consequently, the likelihood of two or more reviews including the same primary study in their analysis was meaningful160. Therefore, we carefully extracted and evaluated all references mentioned in the results section of each included review to exclude overlapping studies.

After establishing analytical themes, the frequency of occurrence for each categorized barrier and facilitator was aggregated into a standard meta-analysis of proportions. Certainty of the evidence was based on the GRADE-Cer-Qual approach161. Nominally identified results are indicated as the relative frequency of occurrence (RFO) and 95% confidence interval (CI). Analysis was executed using R software (version 4.1.1), using the metaprop function package. This study is deemed exempt as it does not assess data or intervene in humans.