This review was conducted in accordance with Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines with a protocol accessible on PROSPERO (registration number: CRD42023403205). A systematic search of three electronic databases (Medline, Embase and Cumulative Index to Nursing and Allied Health Literature) was carried out between 21 February and 6 March 2023, to identify peer-reviewed articles published from inception until 6 March 2023. A grey-literature (Google, Google Scholar) and manual search of the reference lists of included articles was conducted to find additional studies that met the inclusion criteria. Search strategies for all databases are available in online supplemental appendix 1.

Studies were included if they met the following criteria: (1) implemented a qualitative or mixed-methods study design, using interviews, focus groups and open-ended questionnaires; (2) reported clinicians’ experiences with DHIs prior to, during or following surgery/in the perioperative context; and (3) evaluated a digital health intervention intended for use by clinicians or adult patients, as described by the WHO.

Studies were excluded if they: (1) reported only patient or other non-clinician data, (2) presented a digital health intervention used by paediatric clinicians or patients and (3) were not available in English. Editorial comments, reviews, protocols, abstracts and conference proceedings were also excluded.

The search results were uploaded to Covidence (Veritas Health Innovation, Melbourne, Australia; www.covidence.org) for screening and data extraction. Following deduplication, the primary reviewer (AA) filtered articles by title and abstract and screened full-text articles against the eligibility criteria. Concordance checking was undertaken by a second reviewer (CWH) on a sample of 10% of full-text articles. All included articles were also reviewed by CWH. A third reviewer (EVC) was available to resolve disagreements regarding eligibility, where consensus could not be reached. Cohen’s Kappa was calculated to establish inter-rater reliability.

AA extracted data on author, year of publication, country, aims, methods, sample size, clinician participants, clinical specialty and intervention characteristics, using Microsoft Excel (see table 1).

The included articles were imported into NVivo 12 (2017), a qualitative data management programme. To ensure a comprehensive interpretation and analysis of the data, a framework synthesis approach was taken.22 The author used an inductive-deductive, line-by-line coding technique to analyse the data. Following review by CWH, codes were compared across studies and organised into themes, based on their conceptual similarities. Themes were explored in detail to evaluate their alignment with the TFA, before they were systematically mapped to the seven TFA constructs.27 Group discussion between AA, CW and EVC facilitated further examination of the relationships between codes, themes and TFA constructs.

Evaluation of the quality of included studies was performed using the Critical Appraisal Skills Programme (CASP) tool. This was selected as it is the most commonly used checklist for quality appraisal in healthcare-related quality evidence synthesis.32 33 It is endorsed by Cochrane and the WHO for this purpose.33 34 It uses the following criteria: (1) Was there a clear statement of the aims of the research? (2) Is a qualitative methodology appropriate? (3) Was the research design appropriate to address the aims of the research? (4) Was the recruitment strategy appropriate to the aims of the research? (5) Was the data collected in a way that addressed the research issue? (6) Has the relationship between researcher and participants been adequately considered? (7) Have ethical issues been taken into consideration? (8) Was the data analysis sufficiently rigorous? (9) Is there a clear statement of findings?

Quality assessment was undertaken by one reviewer (AA) and independently verified by a second reviewer (CWH) (figure 1). To allow for a comprehensive exploration of the available qualitative data, publications were not excluded based on quality.

No ethical approval was sought for this study as it involves the use of qualitative data from published studies which are freely available in the public domain.

Patients or the public were not involved in the design, conduct, reporting or dissemination plans of our research.

Included studies were conducted between 2012 and 2022, with the majority (n=13) published between 2019 and 2022. Studies were undertaken across six countries: the UK (n=6), the USA (n=5), Canada (n=2), the Netherlands (n=2), Taiwan (n=1) and Uganda (n=1). Four categories of DHIs were evaluated: telehealth interventions (n=7), mobile health applications (n=6), website-based programmes (n=4) and electronic health record (EHR)-integrated software (n=1). A summary of study characteristics is presented in table 1. Clinician participants included surgeons, physicians, nurses, physiotherapists, dieticians and psychologists. The studies represented DHIs used in a range of specialities (online supplemental appendix 1, table 1.

Our analysis identified that the seven TFA constructs (perceived effectiveness, affective attitudes, opportunity costs, ethicality, burden, intervention coherence and self-efficacy) effectively described clinicians’ perceptions of DHIs. On further analysis, 20 more detailed themes emerged. These are summarised in figure 3, with representative quotes provided in table 2.

Perceived effectiveness was explored in all studies (n=18).35–52 Clinicians’ views of DHI effectiveness were shaped by their potential utility (n=18),35–52 observed success (n=15)35–39 41–43 45–48 50–52 and accessibility to patients (n=12).35–37 39–42 44 46 48 49 51

Clinicians exhibited a range of affective attitudes towards DHI. Positive affective attitudes were observed in the majority of studies (n=15).35 36 38–45 47 49–52 This included optimism (n=15)35 36 38–45 47 49–52 and open-mindedness (n=4).40–42 44 Negative affective attitudes such as scepticism (n=8)35–37 40 43 44 46 47 and apprehension (n=10)35 36 40–46 49 appeared in numerous studies (n=12).35–37 39–47 49 52 Indifference was expressed in one study (n=1).47

Most studies raised concerns about the opportunity costs of using DHIs (n=15).35–38 40–47 49 51 52 Clinicians believed that the adoption of various interventions had the potential to impact patient safety (n=9),35 36 40–42 45–47 52 data privacy and security (n=5),37 40 44 46 52 and efficiency (n=15).35–38 40–47 49 51 52

Multiple studies (n=13)35–42 44 46 49 51 52 emphasised the implications of DHI implementation on a clinician’s professional obligation to promote patient autonomy (n=6),38 39 42 44 49 51 beneficence (n=9),35 38–42 44 46 51 non-maleficence (n=7)36 40–42 46 49 52 and justice (n=6).37 39 41 44 46 49

Perceived burdens of DHIs were identified in several studies (n=12).35 37 40–44 46–49 51 These included cost (n=7),40–44 47 48 increased workload (n=6)35 40 42 46 47 49 and the need for user training (n=6).37 40–42 44 51

Across studies, intervention coherence was linked to the clinicians’ perspective of the simplicity of the interventions (n=12).36 37 40 42–47 50–52

Clinician perceptions of self-efficacy were related to their views on the prerequisites for use of interventions (n=10).39–41 43–45 47 48 51 52

This systematic review aimed to assess clinicians’ perceptions of DHIs in perioperative care. Our results show that, across 18 studies, perceived effectiveness was the most commonly identified TFA construct, followed by affective attitudes, opportunity costs, ethicality, burden, intervention coherence and self-efficacy. This information is crucial, given clinicians’ role as key stakeholders in the implementation of DHIs. Indeed, clinicians’ perspectives carry substantial implications for the long-term adoption and efficacy of these technologies as they are the ones to allocate resources efficiently and identify patients most suitable for treatment.27 53 These findings support previous studies which indicate that clinician beliefs regarding the utility and success of DHIs positively influence their acceptance.54

Despite the importance of clinician involvement in intervention development, a recent review noted that their collaboration with the developers of DHIs was insufficient.55 Therefore, DHIs remain in the early stages of implementation and lack evaluation during practice.56 This may undermine clinician confidence in DHIs, contributing to the recurring focus on their effectiveness. This is evident in the diverse affective attitudes exhibited in this study. Clinicians’ optimism and open-mindedness regarding the value of DHIs to patients and providers align with prior studies on digital interventions.57 However, their scepticism and ambivalence regarding the security and utility of DHIs in surgical and clinical settings have also been reported previously.58–60 These attitudes significantly impact acceptability, consistent with a systematic review by Sekhon et al.27

Clinicians have expressed legitimate apprehensions about patient safety, data security, privacy and efficiency.61–63 With the integration of digital health into perioperative pathways, the malfunction or failure of DHIs could have far-reaching ethical consequences and opportunity costs.60 64 65 Prior research also relays the negative impact of these factors on care provision,66 exemplified by a systematic review revealing that 67% of smartphone calculator apps placed diabetes patients at serious risk of insulin overdose.63 Furthermore, as DHIs to diagnose melanoma were reported to be inaccurate in 30% of cases, physicians continue to discourage their use.67 This lack of confidence in DHI efficacy may arise from clinicians’ continued safety concerns.55 68 Moreover, without a clear sense of the benefits of new interventions, they may be more wary of risks.

Clinicians’ reluctance to embrace DHIs may also be attributed to their fear that these interventions could exacerbate existing inequalities, compromising the quality of treatment available to patients facing mental illness and socioeconomic challenges.69 70 These concerns are in line with the ‘inverse care law,’ which suggests that interventions may be least accessible to those who stand to benefit the most from them.71 For example, older populations are less able to use digital technology despite requiring health monitoring the most.72 The lack of benefit conferred by DHIs to older patients has been acknowledged by both clinicians and patients alike.73 This digital divide could limit care to patients marginalised by age, disability, low literacy or lack of digital access.69 Our study also reflects previously expressed concerns that the availability of DHIs on electronic platforms may undermine patient privacy and data security.72 74 75 These sentiments are justified as cyber thieves have recently targeted health insurance information, while millions of stolen phones put personal health records at risk.45 60 76 Such third-party access to data may also lead to discrimination and profiling by marketing agencies, causing psychological distress.68

Our results convey clinicians’ belief that DHIs may undermine efficiency, despite their potential to expedite care. This may be due to perceived difficulties in using technology, as previously highlighted by a systematic review.77 Clinicians also considered cost, increased workload and training requirements as burdens associated with DHIs. Concerns about limited grant funding for DHIs have been documented previously78–81 and clinicians’ perspectives on the financial implications of DHIs are frequently overlooked.55 82 This is relevant as most universal healthcare systems such as the NHS lack sufficient budgets for long-term DHI adoption.78 Perceived workload has been cited as another barrier to DHI adoption.54 74 75 79 80 83–86 The need for training in the use of DHIs contributes to perceptions of increased workload,74 87 as does low intervention coherence. This is an important consideration, as subjective clinician perceptions of workload are a greater predictor of burnout than actual workload.88 89 Addressing these challenges, the use of user-centred design principles has proven effective in promoting simplicity and ease of use of DHIs. This, in turn, facilitates their integration into existing workflows.25 90 Notably, perceived usability also plays a role in determining whether an intervention meets the needs of patients or providers, influencing its acceptance.91 These factors collectively shape perceptions about the prerequisites for DHI use, their attainability and in turn clinicians’ self-efficacy in engaging with DHIs.92 Despite these concerns, clinicians value the role of DHIs in promoting patient autonomy through increased access to information and communication channels.76 These views are echoed by patients, who view DHIs as predominantly beneficial.61 62

This is the first study to assess the acceptability of a wide range of DHIs in perioperative care, offering a comprehensive synthesis of a diversity of perspectives. Our focus on clinicians is an important strength, given their essential role in implementing DHIs. The qualitative inductive-deductive approach draws out important themes, which may not have been captured in traditional quantitative analyses. This contributes to a more nuanced understanding of the factors influencing the acceptability of multiple DHIs across specialties and perioperative phases. The utilisation of a validated framework (TFA) enabled a structured and systematic evaluation of the factors influencing DHI acceptability.27 This, alongside the rigorous methodology employed in screening, coding and synthesis maximised the objectivity and reliability of our findings. The inter-rater reliability of 0.81 suggests a high level of agreement among reviewers, indicating a consistent evaluation process. The ENTREQ (Enhancing transparency in reporting the synthesis of qualitative research) checklist and PRISMA checklist for this work can be found in the supplementary material.

However, due to the heterogeneity of the data, variations in perspective based on intervention type and specialty may have been overlooked. Furthermore, disparities in methodology and methodological rigour among constituent studies may have limited the reliability of inter-study comparisons. The lack of a standardised approach to weighting the evidence across studies is another potential limitation. Additionally, the predominantly single-reviewer approach to screening, coding and synthesis could also be a source of bias. Our database search was also restricted to studies published until 6 March 2023, which may limit the relevance of our findings to more recent developments in DHIs. Moreover, the absence of newer technology such as watch-based applications and wearable devices within our synthesis may have resulted in a narrower range of insights. The inclusion of only English-language studies also limits the relevance of the review to non-English cultural contexts. The over-representation of studies from high-income countries could also constrain the broader applicability of our findings.

The findings of this study provide useful information for the planning and development of DHIs as well as their incorporation into perioperative care pathways. Our narrative synthesis informs policymakers, service providers and DHI developers about the key factors influencing the acceptance of DHIs. These insights can serve as a foundation for enhancing the short- and long-term impact of DHIs. They may also guide the strategic involvement of clinicians in the design and deployment of DHIs. As such, this study highlights the importance of the clinician’s voice in DHI development. It emphasises the need for active clinician participation in co-creating solutions to address barriers underlying intervention implementation.

Future research should investigate DHI preferences to achieve a better understanding of which interventions are favoured by clinicians. Additionally, it would be valuable to explore the perspectives of allied health professionals involved in the management of patients in the perioperative setting. Future studies may also aim to include studies from a broader range of countries to enhance the applicability of the results to diverse socioeconomic contexts. Conducting subgroup analyses could allow for a deeper insight into perspectives by intervention type and specialty.

Not applicable.

Not applicable.