Digital Infrastructure to Support Social Prescribing in Viet Nam: What Should Technology Enable, and Where Should Human Navigation Remain Central?

Digital Infrastructure to Support Social Prescribing in Viet Nam: What Should Technology Enable, and Where Should Human Navigation Remain Central?

Category: Implementation & Practice

Digital technology can make a referral faster. It cannot, by itself, make the referral appropriate.

A community directory can be searchable and still be out of date. A referral can be transmitted instantly and still reach a service with no capacity. A dashboard can show that a referral was “sent” while the person remains at home, unsure whom to contact or whether anyone will contact them. An algorithm can suggest a nearby activity without understanding that the person cannot climb stairs, does not feel comfortable attending alone or simply does not want that kind of support.

This is the central design challenge for digital infrastructure in social prescribing. Technology can help make community resources visible, move information between organizations, track whether connections progress and generate data for improvement. But social prescribing is also relational. It involves understanding what matters to a person, helping them make sense of options, navigating uncertainty and responding when the first connection does not work.

The question for Viet Nam is therefore not simply “How can we digitize social prescribing?” A more useful question is: “Which parts of a community-connected pathway should technology make easier, safer and more reliable, and which parts still depend on human relationships, judgment and navigation?”

This distinction is increasingly important. A 2026 concept analysis published in Frontiers in Public Health searched eight databases through May 2026 and included 30 publications on digital social prescribing. It found substantial variation in how the concept was defined and implemented. Some literature described relatively narrow digital referral tools, while other work described broader cross-sector models in which technology supports needs identification, personalization, community-resource connection and coordination. The authors ultimately framed digital social prescribing as a socio-technical system, rather than a technology product operating in isolation.

That is a useful starting point for Viet Nam.

Digital infrastructure is not social prescribing

A digital platform does not become social prescribing simply because it contains a list of community services.

The 2026 concept analysis identified recurring features including the use of technology, connection to non-clinical services, personalized plans based on people's needs and preferences, community-based resources and involvement of organizations from different sectors. Technology may support resource identification, referral management, communication, monitoring and evaluation.

But this was a concept analysis, not an effectiveness review. It does not establish that adding technology improves social-prescribing outcomes, nor does it show that one digital model is superior to another. The evidence base remains heterogeneous, and much of the literature is conceptual or descriptive.

For SPVN, that distinction matters. Viet Nam does not need to create a separate model called “digital social prescribing” before locally workable social-prescribing pathways have been developed. A more practical approach is to think about digital infrastructure that supports social prescribing and community-connected care.

Technology should serve the pathway. The pathway should not be redesigned simply to serve the technology.

A Western Pacific scoping review published in 2026 further illustrates how uneven the regional landscape remains. It included 42 peer-reviewed studies and 13 grey-literature sources, but only one peer-reviewed study came from Viet Nam. Its searches were completed on 1 November 2024, so it should not be interpreted as a current national assessment of Viet Nam's digital readiness. What it does show is substantial variation in social-prescribing structures, community capacity, referral mechanisms and use of technology across the region.

For Viet Nam, this argues for building digital infrastructure around clearly defined local problems rather than assuming that an established platform from another health system can simply be imported.

What should technology make easier?

The first task is making community resources visible and usable. A navigator needs reliable information about what resources provide, who they may support, eligibility, accessibility, cost, location, participation requirements, contact routes and — where feasible — current capacity. This is why digital infrastructure should connect closely with community asset mapping and a living resource directory.

A searchable directory is useful only when the information behind the search is trustworthy. Technology can make updates easier, show when information was last verified, support filtering and allow community partners to report changes. It cannot solve the underlying governance questions of who owns the information, who verifies it and who is responsible when it becomes outdated.

Technology can also make referral and connection management more reliable. Once a person has agreed on an appropriate option, a digital system can reduce unnecessary manual steps, transmit the information genuinely needed by the receiving organization, record when a connection was initiated and provide a way to identify unresolved cases.

This becomes particularly valuable after a referral has been sent. A pathway may need to know whether the community organization received it, whether contact was attempted, whether the resource had capacity, whether the person actually connected and whether another option is needed. This is often described as a closed-loop referral.

Closed-loop referral does not mean that healthcare needs detailed information about everything a person later does in the community. Different connections require different levels of follow-up. A low-risk activity chosen independently by the person may require little or no formal feedback to healthcare. A pathway involving more complex needs, greater vulnerability or professional coordination may require clearer confirmation and escalation.

Digital infrastructure should also fit the work people are already doing. If staff need to document the same referral in an electronic medical record, a navigation platform, email and a spreadsheet, digitalization has increased administrative burden rather than reduced it.

Appropriate interoperability with existing healthcare systems may therefore become valuable. But interoperability should mean exchanging defined information for a defined purpose, not requiring every community organization to operate inside a hospital information system.

Finally, technology can support learning. A pilot can use structured data to understand which resources are used, where connections fail, how long people wait, which needs are difficult to address and whether particular population groups are systematically missed.

Those data can improve implementation. They should not become surveillance of people's social lives.

A referral is not a connection

One of the most useful things digital infrastructure can do is make visible the difference between sending a referral and achieving a meaningful connection.

Without follow-up, a digital system can create false certainty. A referral has been recorded, so the pathway appears to have worked. But outside the screen, the telephone number may have been wrong, nobody may have answered, the activity may have been full, the person may have become anxious about attending alone or the resource may not have matched what the person thought they had agreed to.

Automation can identify some of these problems. It can flag an unacknowledged referral, remind a navigator to follow up or show that no outcome has been recorded.

It cannot resolve every reason why a connection failed.

A randomized controlled trial published in 2026 provides useful evidence here. Dahrouge and colleagues randomized 326 primary-care patients with health or social needs in Ontario, Canada, to either the Access to Resources in the Community model or 211-Ontario. The ARC model provided comprehensive, longitudinal navigation with informational, instrumental and emotional support, while the comparison group was signposted to inbound navigation services providing mainly informational assistance.

At three months, 50.3% of participants randomized to comprehensive navigation reported accessing at least one needed health or social resource, compared with 35.8% in the signposting group, an absolute difference of 14.5 percentage points.

This trial does not establish that human navigation is superior to digital technology. That was not the comparison. It shows that, in this setting, more comprehensive and longitudinal navigation achieved better access to needed resources than a lower-intensity signposting model.

The implication for digital design is more modest but important: information alone may not be enough for everyone.

A good digital pathway should therefore support different levels of assistance rather than force everyone into self-service. Some people may be comfortable finding and contacting community resources themselves. Others may need a navigator to interpret options, make the first connection, address practical barriers or follow up when something does not work.

Where human navigation should remain central

The conversation about what matters is one of the clearest places where human navigation retains value.

A digital questionnaire can ask about social isolation, financial difficulty, physical activity, food, transportation or caregiving. It can organize answers and suggest categories of support.

But a need is not automatically a priority.

Someone may screen positive for loneliness but have no interest in joining a social group. A person experiencing financial difficulty may consider caregiver exhaustion more urgent. An older adult may technically be eligible for an activity but feel afraid to attend a new place alone. A situation initially described as a social problem may actually require clinical assessment, professional social work or another form of intervention.

These situations require context and conversation.

The 2026 concept analysis similarly argues that digital infrastructure should complement rather than replace frontline human support. Technology can strengthen coordination and access, but professional judgment, relational continuity and individualized support remain important where people's circumstances are complex or uncertain.

Human navigation is particularly valuable when needs are multiple, community options are limited, circumstances change quickly, a person is unsure whether they want support or the first connection has already failed.

It also matters when risk is present. Technology can prompt questions or flag predefined warning signs, but determining whether someone requires clinical care, professional social work, safeguarding or community support still involves appropriate professional judgment and accountability.

Sometimes the most useful question after a failed referral is not “What other service should we send this person to?” It is “What made the first connection fail?”

That is relational work.

Design for people, navigators, healthcare teams and community partners

Digital referral systems have several different users, and what helps one group can create work for another.

A person using the pathway may want simple information, meaningful choices and an easy way to ask for help. A navigator needs current resource information, clear referral status and a manageable view of follow-up. A healthcare professional may need to initiate a connection without repeatedly leaving the clinical workflow. A community organization needs enough information to respond without being overwhelmed by administrative requirements.

A human-centered design study published in Applied Clinical Informatics illustrates this challenge. Haynes and colleagues conducted 48 user tests across three iterative rounds while developing a social-care referral platform involving patients, clinical teams and community-based organizations. The study showed that patient-facing interfaces can support self-screening and self-referral, that community organizations have diverse workflow needs and that integration with electronic health records may make referral processes easier for healthcare teams.

This was a design and usability study, not evidence that the platform improved health outcomes. Its practical lesson is nevertheless important: there is no single user of referral technology.

For Viet Nam, an early system should therefore not assume that every community organization has a sophisticated information system, an API or dedicated IT staff. A technically elegant platform can weaken the pathway if it creates excessive administrative work for small community partners.

Interoperability should be proportionate. A large formal partner may eventually support structured system-to-system exchange. A smaller community group may need a secure form, simple notification process, telephone confirmation or another low-burden mechanism.

The objective is not to make every organization digitally identical. It is to create enough shared information for the pathway to work reliably.

Technology can also improve access for some people while creating new barriers for others. Emerging digital-social-prescribing literature highlights risks related to digital literacy, internet access, socioeconomic disadvantage, accessibility, fairness and bias. These concerns are especially relevant to community-connected care because some of the people who need navigation most may also find digital systems hardest to use.

A digital pathway should therefore preserve meaningful non-digital routes. Someone should not lose access to community support because they do not own a smartphone, cannot complete an online form, do not use email or feel uncomfortable sharing personal information through an app.

Digital first should not become digital only.

An assisted-digital model may be more appropriate. Someone who prefers self-service may search for resources and initiate a connection independently. Another person may use the same underlying infrastructure through a navigator, healthcare professional, social worker or caregiver.

Accessibility also means more than connectivity. Interfaces need understandable language, readable design and support for disability and different levels of digital confidence.

Digital infrastructure also cannot create community capacity that does not exist. A sophisticated matching system does not create a support group where none exists. It cannot make an unaffordable service affordable. It cannot remove a waiting list simply by displaying it.

Sometimes the most useful output from a digital platform will be:

There is currently no appropriate resource available.

That is not necessarily a technical failure. It may be important implementation information about a gap in community capacity.

Viet Nam's digital and legal infrastructure creates opportunities and boundaries

The timing of this discussion is particularly relevant because Viet Nam's digital-health environment is developing rapidly.

Circular No. 13/2025/TT-BYT, effective from 21 July 2025, provides the current framework for implementing electronic medical records. It requires EMR implementation to operate within the broader legal framework governing healthcare, electronic transactions, information systems, cybersecurity, personal-data protection and related requirements.

In April 2026, the Ministry of Health issued Decision No. 965/QĐ-BYT, approving the nationwide electronic medical record implementation plan for 2026–2030. The plan sets the goal that by 2030, 100% of medical examination and treatment establishments will implement electronic medical records without paper medical records, together with synchronized connectivity and life-course health-data management.

This creates opportunities for future social-prescribing pathways to connect appropriately with healthcare information systems.

It does not mean that every interaction between a person and a community organization should become part of the medical record.

As a design principle, the medical record should contain information necessary for healthcare continuity, safety and accountability without automatically reproducing every interaction that occurs in the community. Healthcare teams need enough information to understand relevant referrals and follow-up. Community organizations need enough information to provide the agreed support. Navigators need enough information to coordinate the pathway.

Nobody needs every piece of information merely because it can technically be shared.

This becomes especially important when community partners sit outside the healthcare sector. Digital infrastructure should connect sectors without silently turning community organizations into extensions of the medical record.

Personal referral data may also contain different kinds of information. Some relate directly to health; others may concern finances, housing, caregiving, family circumstances, disability, transportation or contact details. These should not all be casually collapsed into a single category of “health data.”

Viet Nam's Law No. 91/2025/QH15 on Personal Data Protection and Decree No. 356/2025/ND-CP have both been in force since 1 January 2026. Health information receives specific protection under Article 26 of the Law. Where health information is collected or processed, consent requirements and other applicable provisions need to be considered, subject to statutory exceptions. Other personal data remain subject to the broader requirements of the Law and its implementing regulations.

Digital infrastructure therefore needs to define in advance what information is actually necessary, which organization is responsible for which data, who can access them, how access is managed, how errors are corrected, how long information is retained and what happens when personal data move between healthcare and community organizations.

A useful architecture should keep community-resource information separate from personal referral information wherever possible. A living resource directory does not need to contain the health histories of the people using those resources.

Other parts of Viet Nam's digital legal framework are also relevant. Law No. 20/2023/QH15 on Electronic Transactions has been effective since 1 July 2024; Law No. 60/2024/QH15 on Data since 1 July 2025; and Law No. 116/2025/QH15 on Cybersecurity since 1 July 2026.

The practical message for an early pilot is not that every project needs a complex legal architecture before it can start. It is that privacy, security and information governance should be designed into the pathway from the beginning rather than added after a platform is already operating.

AI should be optional, explainable and proportionate to risk

Artificial intelligence may eventually be used to rank resources, suggest options, identify patterns or help navigators search complex directories.

But an early social-prescribing pathway does not need AI simply because AI is available.

Viet Nam's Law No. 134/2025/QH15 on Artificial Intelligence has been effective since 1 March 2026. Decree No. 142/2026/ND-CP, effective from 1 May 2026, provides detailed rules including risk classification and obligations relating to AI systems. Decision No. 33/2026/QĐ-TTg, effective from 15 August 2026, promulgates the national list of high-risk AI systems.

If AI is introduced for matching, ranking or recommendation, the system should therefore be assessed according to its actual function, intended use and applicable risk classification.

An AI-enabled social-prescribing recommender should not automatically be assumed to be a high-risk AI system simply because it is used somewhere in healthcare. Classification depends on what the system actually does, how its outputs are used and whether it falls within the applicable legal criteria.

From an implementation perspective, a simpler approach may often be preferable at first. Transparent filters based on location, accessibility, eligibility, cost and personal preference, followed by a human conversation, may be more useful and easier to govern than an opaque recommendation engine.

Where algorithmic recommendations materially shape the options shown to a person or navigator, transparency and appropriate human oversight should be built into the process.

A recommendation should support a choice, not become the choice.

Digital infrastructure also does not redefine professional boundaries. Circular No. 51/2024/TT-BYT, effective from 1 March 2025, regulates the implementation of social work in medical examination and treatment establishments licensed to operate as hospitals. Circular No. 29/2026/TT-BYT, effective from 25 August 2026, regulates social-work practice and the updating of social-work knowledge.

These regulations do not establish social prescribing. They do matter, however, when a digitally enabled pathway intersects with functions already performed by hospital social workers or when someone's circumstances require professional social-work intervention rather than community navigation.

A digital interface should not make clinical care, professional social work and low-intensity community connection appear interchangeable simply because all three can be selected from the same screen.

The technology should reflect the reality of professional responsibility.

Start small and evaluate the pathway, not the software

For an early Vietnamese pilot, the safest digital strategy may be deliberately modest.

The first version does not need artificial intelligence, a national resource database or integration with every hospital information system. It needs to solve a small number of real implementation problems well.

A useful minimum infrastructure might include a maintained resource directory, a simple way to initiate a connection, visibility of referral status where appropriate, a mechanism for identifying unsuccessful connections, basic communication between navigator and community partner, and enough structured information to evaluate whether the pathway is functioning.

Even these functions require ownership. Who maintains the directory? Who follows up an unanswered referral? Who determines when a connection is closed? Who corrects inaccurate information? Who decides what personal information a community partner can see? Who responds to a privacy or cybersecurity incident?

If those questions do not have clear answers, adding more features will not make the pathway more mature.

It may only make the uncertainty digital.

As implementation matures, a successful pilot might move toward greater interoperability with electronic health records, more structured community-partner interfaces, person-facing self-navigation options and more sophisticated analytics.

Digital maturity should follow implementation maturity.

Evaluation should also focus on whether technology improves the pathway, not simply whether people use the software. Login counts, clicks, referrals generated and resources viewed are easy to measure, but they do not show whether community-connected care became better.

More meaningful questions include whether technology reduces the time needed to find an appropriate resource, makes failed connections visible sooner, increases completed connections, reduces duplicated administrative work and helps community organizations communicate changes in capacity.

Evaluation should also ask whether people understand their options, retain meaningful choice and receive human support when they need it. Digital equity should be visible in the data. Are older people less likely to complete digitally initiated connections? Do people with low digital confidence leave the pathway earlier? Are rural users offered fewer realistic options? Do recommendations repeatedly point toward services some groups cannot afford or physically reach?

Community-partner burden matters too. A system that saves time for a hospital while creating substantial new data-entry work for a small community organization has not necessarily improved the pathway.

Privacy, security and safety are also implementation outcomes. Inappropriate information sharing, unauthorized access, mistaken referrals and failures to escalate serious concerns should be monitored alongside traditional performance measures.

The ultimate question is straightforward:

Does the technology make it easier for people to reach appropriate support without weakening choice, equity, relationships, privacy or professional accountability?

If it does not, digitalization is not improvement.

Technology should make community-connected care more human, not less

The strongest digital infrastructure may be the infrastructure people notice least.

It keeps community information current. It helps a navigator identify realistic options. It moves the minimum necessary information to the right place. It shows when a connection has stalled. It reduces duplicated work and generates useful information for improvement.

And then it gets out of the way of the conversation.

Current evidence does not justify assuming that automated screening, matching and referral can replace relational navigation where personalized guidance and ongoing support are needed. The 2026 concept analysis reinforces the importance of keeping digital social prescribing human-centered, while the randomized trial comparing comprehensive navigation with signposting provides stronger evidence that the intensity and continuity of navigation support can matter for people's ability to reach needed resources.

For Viet Nam, this suggests a clear direction. Build digital infrastructure that makes community resources more visible and information more reliable. Use it to support referral, follow-up and learning. Connect it appropriately with the country's developing digital-health infrastructure. Protect personal information from the beginning. Preserve non-digital and assisted-digital routes. And keep human support available wherever trust, complexity, judgment, uncertainty or failed connections require more than information alone.

The purpose is not to automate community connection.

It is to make human navigation more informed, community partnerships more workable, and connections between healthcare, social needs and community resources more reliable.

That is the role digital infrastructure should play in responsible community-connected care.

References

1. Zhang Y, Xie A, Wang R, Fan Y. Digital social prescribing: a concept analysis. Frontiers in Public Health. 2026;14:1857845. doi:10.3389/fpubh.2026.1857845.

2. Dahrouge S, Gauthier AP, Durand F, et al. A Randomized Controlled Trial of Social Prescribing: Comparing a Comprehensive Navigation Model With Signposting. Annals of Family Medicine. 2026;24(4):301–310. doi:10.1370/afm.250265.

3. Lwin KS, Wong MSJ, Tan MH, et al. Organization and implementation of social prescribing in the Western Pacific Region: a scoping review. The Lancet Regional Health – Western Pacific. 2026;67:101714. doi:10.1016/j.lanwpc.2025.101714.

4. Haynes D, Cheng P, Weaver M, Parsons H, Karaca-Mandic P. A Human-Centered Approach for Designing a Social Care Referral Platform. Applied Clinical Informatics. 2025;16(1):67–76. doi:10.1055/a-2425-8731.

5. World Health Organization Regional Office for the Western Pacific. A toolkit on how to implement social prescribing. Manila: WHO Regional Office for the Western Pacific; 2022.

6. Ministry of Health of Viet Nam. Circular No. 13/2025/TT-BYT guiding the implementation of electronic medical records. Effective 21 July 2025.

7. Ministry of Health of Viet Nam. Decision No. 965/QĐ-BYT approving the plan for nationwide implementation of electronic medical records for 2026–2030. 10 April 2026.

8. National Assembly of Viet Nam. Law No. 20/2023/QH15 on Electronic Transactions. Effective 1 July 2024.

9. National Assembly of Viet Nam. Law No. 60/2024/QH15 on Data. Effective 1 July 2025.

10. National Assembly of Viet Nam. Law No. 91/2025/QH15 on Personal Data Protection. Effective 1 January 2026.

11. Government of Viet Nam. Decree No. 356/2025/ND-CP detailing certain articles and measures for implementation of the Law on Personal Data Protection. Effective 1 January 2026.

12. National Assembly of Viet Nam. Law No. 116/2025/QH15 on Cybersecurity. Effective 1 July 2026.

13. National Assembly of Viet Nam. Law No. 134/2025/QH15 on Artificial Intelligence. Effective 1 March 2026.

14. Government of Viet Nam. Decree No. 142/2026/ND-CP detailing certain articles and measures for implementation of the Law on Artificial Intelligence. Effective 1 May 2026.

15. Prime Minister of Viet Nam. Decision No. 33/2026/QĐ-TTg promulgating the List of High-Risk Artificial Intelligence Systems. Effective 15 August 2026.

16. Ministry of Health of Viet Nam. Circular No. 51/2024/TT-BYT on the implementation of social work in medical examination and treatment establishments. Effective 1 March 2025.

17. Ministry of Health of Viet Nam. Circular No. 29/2026/TT-BYT on social-work practice and updating social-work knowledge. Effective 25 August 2026.