Key Take Home
Digital Health interventions for PA promotion in primary healthcare are acceptable, but their effectiveness varies. Qualitative studies show what stakeholders desire from digital health interventions: easy to use, complements pre-existing primary care provision, and is patient-centred.
Introduction
Regular physical activity (PA) results in wide ranging physical and mental health benefits (Hardman and Stensel, Reference Hardman and Stensel2009). Despite this, one third of adults in the European Union fail to meet the PA guidelines as defined by the World Health Organisation (WHO) (Bull et al., Reference Bull, Al-Ansari, Biddle, Borodulin, Buman, Cardon, Carty, Chaput, Chastin and Chou2020), with almost half (45%) reporting they never exercise or play sport (OECD and WHO, 2023). This is important, as physical inactivity has a large detrimental impact on healthcare services, which are already stretched by the increasing burden of non-communicable diseases (Bull et al., Reference Bull, Cho, Friedman, Santos and Willumsen2022).
Primary healthcare professionals have wider exposure to the whole population than any other health professional – regularly seeing those in need of PA advice and viewed by the public as a trusted source of information (Lion et al., Reference Lion, Vuillemin, Thornton, Theisen, Stranges and Ward2019; McNally, Reference McNally2015). Given this level of exposure, it is not surprising that PA promotion delivered via primary healthcare has been shown to be effective at increasing PA in patients (Kettle et al., Reference Kettle, Madigan, Coombe, Graham, Thomas, Chalkley and Daley2022) and is cost-effective (Campbell et al., Reference Campbell, Holmes, Everson-Hock, Davis, Buckley Woods, Anokye, Tappenden and Kaltenthaler2015). PA promotion in healthcare settings can take a number of different formats, but in general refers to PA counselling, PA on prescription or exercise referral (Orrow et al., Reference Orrow, Kinmonth, Sanderson and Sutton2012). Acknowledging the cost-effectiveness of PA promotion in primary healthcare, the WHO for Europe highlighted PA counselling in primary healthcare as one of its ‘best buys’ in an economic analysis of cost per disability-adjusted life years averted (WHO, 2017). Despite primary healthcare being a key point of influence for PA behaviours, evidence shows poor implementation of PA promotion by general practitioners (GPs) (Barnes and Schoenborn, Reference Barnes and Schoenborn2012; Chatterjee et al., Reference Chatterjee, Chapman, Brannan and Varney2017).
The development of technology has the potential to address many of the barriers to promotion of PA in primary healthcare (Kennedy and Hales, Reference Kennedy and Hales2018). A scoping review by Wattanapisit and colleagues (Wattanapisit et al., Reference Wattanapisit, Tuangratananon and Wattanapisit2020) explored the usability and utility of technology for delivering PA promotion in primary healthcare. It found mixed findings on usability and utility of technology in assisting with PA promotion, with the major barriers to use included complexity and technical issues. Since the publication of this scoping review, the world has experienced a global pandemic (COVID-19), necessitating accelerated technological advancement in health sectors. Consequently, this study aimed to update the scoping review performed in 2020, providing an overview of digital health interventions tailored for PA counselling in primary healthcare and investigated their acceptability and effectiveness. The secondary aim was to categorise the method of action of the digital health interventions according to a recognised behaviour change model.
Methods
This scoping review was registered in the Open Science Framework (DOI 10.17605/OSF.IO/R26QB). The scoping review was conducted according to the recommendations of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) extension for Scoping Reviews (Tricco et al., Reference Tricco, Lillie, Zarin, O’Brien, Colquhoun, Levac, Moher, Peters, Horsley, Weeks, Hempel, Akl, Chang, McGowan, Stewart, Hartling, Aldcroft, Wilson, Garritty and Straus2018) and was modelled on a previous review (Wattanapisit et al., Reference Wattanapisit, Tuangratananon and Wattanapisit2020).
Search strategies
A systematic search strategy was performed across five databases: Cochrane Library, Embase, MEDLINE, PubMed and Web of Science (from 01/01/2020 until 05/12/2023). The search terms consisted of three components: PA counselling, technology and primary healthcare. This is outlined in Table 1.
Table 1. Search terms
Study eligibility
Studies were included if they were: (1) original published peer-review quantitative or qualitative articles conducted in primary healthcare settings, (2) technical or governmental documents (3) published in English and (4) published between 1st January 2020 and 5th December 2023. Studies were excluded if: (1) they consisted of review articles, opinion excerpts, protocols, conference abstracts, (2) full-text was not accessible after all avenues exhausted, (3) participants included paediatric populations or (4) participants had specific diseases requiring specialist care (e.g. cancers, chronic obstructive pulmonary disease).
For the purpose of this review, digital health was defined in accordance with the WHO Global Strategy on Digital Health (WHO, 2021) and includes mobile-phone applications, websites and wearable technology.
Selection of evidence sources
After duplicate removal, the titles, abstract and full texts were screened by a single reviewer, with 10% selected for a second review, to check for discrepancies in agreement. Any discrepancies were resolved on discussion between authors. The Rayyan software was used to manage the studies found, with the final included studies exported to Endnote citation manager version 20. One author performed data extraction using an extraction template file.
Quality appraisal
Study quality was assessed using the Mixed Methods Appraisal Tool (MMAT) Version 2018, accounting for the wide range of study types included (Hong et al., Reference Hong, Fàbregues, Bartlett, Boardman, Cargo, Dagenais, Gagnon, Griffiths, Nicolau and O’Cathain2018). The MMAT tool is designed for mixed method research, but is also suitable for use within qualitative research, quantitative descriptive studies and both randomised and non-randomised studies. The MMAT uses five quality criteria that differ according to the type of study, with the possible outcome from whether each criterion being met ‘yes’, ‘no’ and ‘can’t tell’. The quality of included studies was assessed using the appropriate category for the study type. Each article was appraised by a single author and scored (using the 0–5 scoring system). A score out of 5 was provided for each study (see Table 2, summary of included studies) as a guide to study quality.
Table 2. summary of the included studies
Analysis
Descriptive statistics were used to present quantitative data. Each intervention was categorised according to the stage of action using the 5A’s model and highlighted in the WHO BRIEF project (integrated brief interventions for noncommunicable disease risk factors in primary care: the manual) (Anderson et al., Reference Anderson, Ferrieria-Borges, Wickramansinghe, Malykh, Hetz and Breda2022). The model provides an evidence-based structure for healthcare professionals to address health promotion. The stages of the 5A’s model are: (1) ask and measure exposure to risk factors (2) advise patients to change exposure to risk factors, (3) assess readiness to change, (4) assist patients in acquiring the motivation, self-help skills or support needed and (5) arrange follow-up support (Anderson et al., Reference Anderson, Jorenby, Scott and Fiore2002; Glasgow et al., Reference Glasgow, Emont and Miller2006).
The effectiveness of a digital health intervention was determined based on whether the study reported statistically significant positive outcomes in comparison to a control, normal care or baseline condition. Acceptability was considered only when explicitly assessed and reported by the study authors, such as through user satisfaction, engagement metrics or qualitative feedback. All assessments of effectiveness and acceptability were based solely on self-reported data as presented in the original studies, with no additional secondary analysis conducted.
Qualitative data were analysed using a conventional content analysis, as described by Hsieh and colleagues (Hsieh and Shannon, Reference Hsieh and Shannon2005), and this was used to identify patterns within qualitative data to allow for systematic coding and categorisation. The results were obtained via the following steps: (a) familiarisation with the papers by one author; (b) identification of previously identified themes from qualitative papers were re-coded into sub-categories; (c) these sub-categories were then categorised into themes according to similarities and differences. All themes were initially developed by one author and finalised in discussion with other authors.
Results
Summary of search results and study selection
Our systematic search retrieved 2,717 studies, including 170 duplicates, leaving a total of 2,547 studies for review. Following title and abstract screening, 94 studies were included for full text review. We identified 29 studies for inclusion, as shown in the PRISMA flow diagram (Figure 1).
Figure 1. PRISMA flow diagram.
Of the included studies, 13 were randomised control trial, 2 mixed-method studies, 6 prospective cohort studies, 2 retrospective cohort studies and 6 qualitative phenological studies. Study quality was mixed, with nine deemed high, seventeen moderate and three low quality as assessed by the MMAT scoring system.
Lifestyle counselling domains
Of the twenty-nine studies included in this scoping review, nineteen used interventions focused solely on PA. Five studies focused on interventions that integrated the four other domains of lifestyle medicine: PA, nutrition, alcohol and smoking. The final five studies focused on interventions combining PA and nutrition (Figure 2).
Figure 2. Digital Health interventions separated by counselling domains they deliver (n = 29).
Method of digital health intervention delivery
Twenty-three out the twenty-nine studies (79%) evaluated or assessed a specific digital health intervention. Of these studies, eighteen (78%) investigated a mobile application, with twelve of these combined with a fitness tracker. A further four (17%) digital health interventions were web-based, with one intervention using social media messaging to deliver education and motivation (Figure 3).
Figure 3. Method of delivery of digital health intervention (n = 23).
Method of implementation
Of the 23 studies that evaluated a specific digital health intervention, each was categorised based on how it aligned with the 5A’s model for behavioural counselling (Ask, Advise, Assess, Assist, Arrange), as summarised in Table 3. Most interventions addressed multiple components of the model, though with varying emphasis.
Table 3. Method of implementation of PA promotion
The “Ask” component—assessing PA levels—was addressed in 16 of the 23 studies (70%), typically through self-reported questionnaires. The “Advise” step, involving personalised recommendations to increase PA, was identified in 14 studies (61%). Only one study explicitly incorporated the “Assess” step, which involves evaluating an individual’s readiness or confidence to change behaviour.
The “Assist” category was the most commonly addressed, with 22 interventions (96%) providing tools or strategies to support behaviour change. This included features such as PA tracking (n = 12, 52%) and educational content (n = 15, 65%) aimed at enhancing motivation and self-efficacy. Finally, the “Arrange” step—typically referring to planning follow-up or referrals—was present in only one study (4%).
Effectiveness and acceptability
Nineteen studies reported on the effectiveness of the digital health intervention on increasing PA. Of these, eight did not report any significant improvements in PA due to the digital health intervention when compared to control groups. However, eleven studies (57%) did report significant improvements in PA levels of digital health intervention users. No studies identified that the digital health interventions decreased PA levels of patients when compared to standard care or no intervention.
Of the studies that reported on the acceptability of the digital health intervention, ten presented positive outcomes (see Figure 4). Four of these studies reported a high completion rate in the intervention group, implying a translation of acceptability into practice (Dhinagaran et al., Reference Dhinagaran, Sathish, Soong, Theng, Best and Car2021; King et al., Reference King, Campero, Sheats, Sweet, Hauser, Garcia, Chazaro, Blanco, Banda and Ahn2020; Lugones-Sanchez et al., Reference Lugones-Sanchez, Recio-Rodriguez, Agudo-Conde, Repiso-Gento, Adalia, Ramirez-Manent, Sanchez-Calavera, Rodriguez-Sanchez, Gomez-Marcos and Garcia-Ortiz2022; Young et al., Reference Young, Miyamoto, Dharmar and Tang-Feldman2020).
Figure 4. Acceptability and effectiveness of digital health interventions, presented by number.
Three studies showed poor uptake on the digital health intervention (Hawkes et al., Reference Hawkes, Miles, Ainsworth, Ross, Meacock and French2023; Mattila et al., Reference Mattila, Horgan, Palmeira, O’Driscoll, Stubbs, Heitmann and Marques2022; Parker et al., Reference Parker, Barr, Stocks, Denney-Wilson, Zwar, Karnon, Kabir, Nutbeam, Roseleur and Liaw2022). All of these were education or advisory based without an exercise-tracking capacity, with one(Hawkes et al., Reference Hawkes, Miles, Ainsworth, Ross, Meacock and French2023) finding engagement was higher when support was combined with a health-coach.
Content analysis
Six of the studies included in this scoping review (Bondaronek et al., Reference Bondaronek, Dicken, Singh Jennings, Mallion and Stefanidou2022; Breeman et al., Reference Breeman, Keesman, Atsma, Chavannes, Janssen, van Gemert-Pijnen, Kemps, Kraaij, Rauwers and Reijnders2021; Buss et al., Reference Buss, Varnfield, Harris and Barr2022; Wattanapisit, et al., Reference Wattanapisit, Amaek, Wattanapisit, Tuangratananon, Wongsiri and Pengkaew2021; Wattanapisit, et al., Reference Wattanapisit, Wattanapisit, Tuangratananon, Amaek, Wongsiri and Petchuay2021; Woldamanuel et al., Reference Woldamanuel, Rossen, Andermo, Bergman, Åberg, Hagströmer and Johansson2023) were qualitative phenomenological studies, which explored stakeholder (patients, primary healthcare practitioners and tech businesses) views of digital health interventions. From these papers three main themes emerged:
-
(1) Ease of use. This was important for both practitioners and patients, and simplicity was identified as an essential component (Bondaronek et al., Reference Bondaronek, Dicken, Singh Jennings, Mallion and Stefanidou2022; Breeman et al., Reference Breeman, Keesman, Atsma, Chavannes, Janssen, van Gemert-Pijnen, Kemps, Kraaij, Rauwers and Reijnders2021; Wattanapisit, et al., Reference Wattanapisit, Wattanapisit, Tuangratananon, Amaek, Wongsiri and Petchuay2021; Wattanapisit, et al., Reference Wattanapisit, Wattanapisit, Tuangratananon, Amaek, Wongsiri and Petchuay2021; Woldamanuel et al., Reference Woldamanuel, Rossen, Andermo, Bergman, Åberg, Hagströmer and Johansson2023).
-
(2) The need for any digital health intervention to act as an adjunct to primary healthcare. Healthcare professionals are time pressured, and therefore interventions need to be easy to implement with minimal time investment and a strong evidence-base regarding effectiveness. As an extension of this, and in-keeping with the need for simplicity, a desire was expressed for any digital health intervention to require minimal resourcing. Finally, to deliver a holistic and combined service, a need for integration with pre-existing systems and electronic health records is required (Bondaronek et al., Reference Bondaronek, Dicken, Singh Jennings, Mallion and Stefanidou2022; Breeman et al., Reference Breeman, Keesman, Atsma, Chavannes, Janssen, van Gemert-Pijnen, Kemps, Kraaij, Rauwers and Reijnders2021; Wattanapisit, et al., Reference Wattanapisit, Wattanapisit, Tuangratananon, Amaek, Wongsiri and Petchuay2021).
-
(3) Digital health interventions should be patient centred. All stakeholder groups expressed a wish for personalised interventions that support the individual and their autonomy, whilst also offering the possibility of interpersonal communication if required (Bondaronek et al., Reference Bondaronek, Dicken, Singh Jennings, Mallion and Stefanidou2022; Breeman et al., Reference Breeman, Keesman, Atsma, Chavannes, Janssen, van Gemert-Pijnen, Kemps, Kraaij, Rauwers and Reijnders2021; Wattanapisit, et al., Reference Wattanapisit, Wattanapisit, Tuangratananon, Amaek, Wongsiri and Petchuay2021; Wattanapisit, et al., Reference Wattanapisit, Wattanapisit, Tuangratananon, Amaek, Wongsiri and Petchuay2021; Woldamanuel et al., Reference Woldamanuel, Rossen, Andermo, Bergman, Åberg, Hagströmer and Johansson2023).
Discussion
Comparison with existing literature
This scoping review identifies the wide variety (related to both method and means) of digital health interventions. Although the majority of included studies (65.5%) focussed solely on PA, other lifestyle domains were included in some digital health interventions, including nutrition, alcohol and smoking cessation advice. The integration of different lifestyle domains is likely to impact outcome, as outlined in a recent narrative review (Leese et al., Reference Leese, Al-Zubaidi and Smith2024). An uplift might be particularly pronounced when PA and nutrition counselling are co-delivered (Johns et al., Reference Johns, Hartmann-Boyce, Jebb, Aveyard and Group2014). However, outcomes appear to be impaired when smoking cessation advice is co-delivered with other lifestyle interventions(Meader et al., Reference Meader, King, Wright, Graham, Petticrew, Power, White and Sowden2017; Schulz et al., Reference Schulz, Kremers, Vandelanotte, van Adrichem, Schneider, Candel and de Vries2014).
In this scoping review, 78% of all digital health interventions were delivered by a mobile application. A plethora of PA mobile applications exist, with over 150,000 existing in 2017 (Kennedy and Hales, Reference Kennedy and Hales2018). There is a lack of standardisation between mobile applications, with research highlighting they are frequently limited in their scope, function or compliance with the WHO PA guidelines (Foster, Reference Foster2019; Schoeppe et al., Reference Schoeppe, Alley, Rebar, Hayman, Bray, Van Lippevelde, Gnam, Bachert, Direito and Vandelanotte2017). This lack of standardisation, makes assessment of effectiveness and validity challenging (Baker et al., Reference Baker, Gustafson, Shaw, Hawkins, Pingree, Roberts and Strecher2010).
The categorisation of digital health interventions into a recognised behaviour change model (5A’s) has, to the best of the author’s knowledge, not previously been done. By providing a framework for analysis it allows the identification of what features contribute to digital health intervention effectiveness and acceptability.
The results in our study regarding the acceptability of digital health interventions for PA promotion in primary healthcare combined with previous research (Gonçalves et al., Reference Gonçalves, Moraes and Silva2022) support the stance of the WHO in their Global Action Plan on PA (WHO, 2019), which highlights technology interventions as a viable and strategic means of engaging patients in PA and supporting health-related behaviour change. The results from the content analysis provides clear guidance as to what patients, practitioners and digital health-developers desire: a patient-centred application which provides autonomy, is simple to use and works as an adjunct to ongoing primary healthcare services. To act as an adjunct for primary healthcare, it must be quick and simple to use and integrate with pre-existing systems.
Despite the endorsement by the WHO regarding digital health interventions being acceptable, this review found no conclusive evidence as to their effectiveness. There are several possible reasons for this, including: (1) different abilities and needs of distinct patient populations, (2) changing intergenerational needs and desires, (3) technological literacy and availability existing along chronological and geographical disparities and (4) healthcare-system differences. Alongside this variability in studies, the absence of clear evidence is exacerbated due to a poor quality of existing studies (Eland-de Kok et al., Reference Eland-de Kok, van Os-Medendorp, Vergouwe-Meijer, Bruijnzeel-Koomen and Ros2011; Zangger et al., Reference Zangger, Bricca, Liaghat, Juhl, Mortensen, Andersen, Damsted, Hamborg, Ried-Larsen and Tang2023).
A key consideration for the broader implementation of digital health interventions is how they complement and enhance the work of primary healthcare providers. While many of the reviewed studies focused on patient-facing outcomes, no studies included in this review explored whether digital health interventions increased the rate of PA promotion by primary care professionals or addressed the communication pathways between digital tools and healthcare professionals. Although Mendes and colleagues (Mendes et al., Reference Mendes, Nunes Silva, Santos Silva, Marques, Godinho, Tomás, Agostinho, Madeira, Rebelo-Marques and Martins2020) explored the utilisation of digital health interventions in Portuguese primary healthcare, it is not clear whether this represents any change to pre-intervention levels of PA promotion. For a digital health intervention to be effective within a primary care, it is likely that primary healthcare professionals need to be engaged with the intervention and informed of patient progress and outcomes. This could occur through integration with electronic health records, automated alerts or summary reports that support clinical decision-making.
Strengths and limitations
This scoping review followed the previously published protocol and the PRISMA guidelines outlined for reporting scoping reviews (Tricco et al., Reference Tricco, Lillie, Zarin, O’Brien, Colquhoun, Levac, Moher, Peters, Horsley, Weeks, Hempel, Akl, Chang, McGowan, Stewart, Hartling, Aldcroft, Wilson, Garritty and Straus2018). In a rapidly changing context, particularly in light of the recent COVID-19 pandemic, this provides an up-to-date overview of digital health interventions and to the best of our knowledge this is also the first study to have defined the method of digital health intervention by a well-known behaviour change model (5 A’s model). Only including papers from January 2020 until December 2023 (as per protocol) was decided in the context of previous work (Wattanapisit et al., Reference Wattanapisit, Tuangratananon and Wattanapisit2020).
There are several limitations. Given defined inclusion criteria, the exclusion of any non-English studies, grey literature and abstracts may have resulted in a loss of some data. The scope of the search terms used was relatively narrow, a pragmatic and realistic response to time and resource constraints. While the review highlights a range of functionalities that digital health interventions can offer—including supporting, motivating, monitoring and promoting PA—these broader concepts were not fully reflected in the initial search strategy. As a result, it is possible that some relevant interventions were not captured. Future research should incorporate a wider range of search terms to better reflect the full spectrum of digital health interventions used to promote PA. The data-extraction and thematic analysis was conducted by a single author as part of a pragmatic decision-making process, however the effect of this was minimised by a 10% check for reliability by a second author. Given the aims of a scoping review to provide a holistic overview of the literature, although a study quality appraisal was performed, no studies were subsequently excluded. Furthermore, to allow for the inclusion of a wide variety of studies, a meta-analysis was not performed. A future meta-analysis would allow for quantification of the effectiveness of digital health interventions to assist the promotion of PA in primary healthcare.
Implications for practice
Of all the studies included, none looked at the impact of a digital health intervention on practitioners’ attitudes and behaviours. With the knowledge that PA promotion is effective at increasing patients levels of PA (Haskell, Reference Haskell2003) and is cost-saving (Campbell et al., Reference Campbell, Holmes, Everson-Hock, Davis, Buckley Woods, Anokye, Tappenden and Kaltenthaler2015), future research should explore the impact of technologies on practitioners attitudes and behaviours towards PA promotion.
Given the need to increase PA levels across all populations, covering the spectrum of age, health and geography, more robust large scale randomised-control trials with long -term follow-up across all these groups are required. The needs of different groups may not align (for example adolescent versus elderly) and so the inclusion of qualitative work to assess acceptability in these groups is also of value.
This review also sought to categorise the methods of action of the digital health interventions seeking to deliver PA promotion according to a recognised behaviour change model- the 5A’s framework. The results show the majority of interventions (96%) focused on “assisting” individuals to be more active. Interventions incorporating components that “assess” readiness to change and “arrange” on-going follow-up were far less common. This highlights a gap in the integration of tailored behavioural assessment and long-term structured follow-up in digital health interventions to support PA promotion, which could be vital for sustaining behaviour change over time. Future research should explore the role of including components that incorporate behavioural assessment and follow-up in digital health interventions.
The future of healthcare involves the integration of technology, and this includes PA promotion (WHO, 2019). The availability of technology is this field is already vast (Kennedy and Hales, Reference Kennedy and Hales2018), but evidence-based and trustworthy tools need to be created. The content analysis offers a useful oversight of stakeholder wishes and can guide future work to standardised the approach, for example by a delphi-study (Baker et al., Reference Baker, Gustafson, Shaw, Hawkins, Pingree, Roberts and Strecher2010). This has the potential to enhance the evidence base and lead to quicker implementation in national policy and on a global scale.
Conclusions
This scoping review found that digital health interventions for PA promotion in primary healthcare were acceptable, but findings regarding effectiveness were mixed. Findings from included qualitative studies provide clear guidance as to what stakeholders’ desire from digital health interventions in this area, with future output likely to benefit from clarity on application development and means of evaluation. Finally, further work is needed to evaluate the impact of digital health measures on practitioners’ PA attitudes and behaviours as well as patient PA levels.
Acknowledgements
None.
Author contributions
CL conceived the study and assistance for KA researched the literature. CL, HA and CK were involved in result interpretation and manuscript development. All authors reviewed and edited the manuscript and approved the final version.
Funding statement
This research did not receive any specific grant from funding agencies in the public, commercial or not-for-profit sectors.
Competing interests
Nothing to declare. This research did not receive any specific grant from funding agencies in the public, commercial or not-for-profit sectors.
Ethical standards
Not required.
Guarantor
CL.
Open access
