Lessons from the

Lessons field
from the field

Applying artificial intelligence and digital health technologies,

Viet Nam
Ho Quang Chanh,a Damien K Ming,b Quang Huy Nguyen,a Tran Minh Duc,a Luu Phuoc An,a Huynh Trung Trieu,c
Stefan Karolcik,d Bernard Hernandez, Perez,d Jennifer Van Nuil,a Ngan Nguyen Lyle,a Evelyne Kestelyn,a
Louise Thwaites,a Pantelis Georgiou,d Chris Paton,e Alison Holmes,b Nguyen Van Vinh Chauf & Sophie Yacouba

Problem Direct application of digital health technologies from high-income settings to low- and middle-income countries may be
inappropriate due to challenges around data availability, implementation and regulation. Hence different approaches are needed.
Approach Within the Viet Nam ICU Translational Applications Laboratory project, since 2018 we have been developing a wearable device
for individual patient monitoring and a clinical assessment tool to improve dengue disease management. Working closely with local staff at
the Hospital for Tropical Diseases, Ho Chi Minh City, we developed and tested a prototype of the wearable device. We obtained perspectives
on design and use of the sensor from patients. To develop the assessment tool, we used existing research data sets, mapped workflows and
clinical priorities, interviewed stakeholders and held workshops with hospital staff.
Local setting In Viet Nam, a lower middle-income country, the health-care system is in the nascent stage of implementing digital health
technologies.
Relevant changes Based on patient feedback, we are altering the design of the wearable sensor to increase comfort. We built the user
interface of the assessment tool based on the core functionalities selected by workshop attendees. The interface was subsequently tested
for usability in an iterative manner by the clinical staff members.
Lessons learnt The development and implementation of digital health technologies need an interoperable and appropriate plan for data
management including collection, sharing and integration. Engagements and implementation studies should be conceptualized and
conducted alongside the digital health technology development. The priorities of end-users, and understanding context and regulatory
landscape are crucial for success.

Introduction Local setting

Artificial intelligence and digital health technologies are In Viet Nam, a lower middle-income country, opportunities
improving clinical services, and are facilitating the decen- exist to apply artificial intelligence and digital health tech-
tralization of health care. Within low- and middle-income nologies to improve health outcomes, especially for many
countries, opportunities for applying these innovations are infectious diseases, which exert significant burdens on the
rapidly being realized for diagnostics, patient monitoring health-care system. The Vietnamese health-care system is in
and aiding clinical decision-making.1 However, significant the nascent stage of implementing digital health technologies.5
barriers to implementation remain, including data sharing For example, electronic medical records are being introduced
and individual rights to privacy, algorithmic bias, end- for routine clinical use. In contrast with high-income settings
user interpretability and accountability.2 To date, research where electronic medical records data is routinely generated,
addressing these barriers has mainly been conducted in availability of such data in our setting is limited. While infor-
high-income settings,3 and direct application of these find- mation on heterogeneity between health-care settings exists,
ings to low-and middle-income health-care settings is often care pathways, strategic health system priorities, infrastructure
inappropriate, given significant differences in context and and capacity for individual settings remain poorly character-
resource availability.2,4 ized, hampering the estimation of resources required and the
Here, we share our experience in conducting multidis- development of an implementation plan of new digital health
ciplinary research for the development and implementation technologies.2,4
of digital health technologies in Viet Nam. We outline chal-
lenges encountered and describe specific findings around
real-world implementation. Our approaches and methods
Approach
are applicable to similar settings and provide insights on The Viet Nam ICU Translational Applications Laboratory
design considerations from a low- and middle-income (VITAL) project, funded by the Wellcome Trust, started in
country perspective. 2018. The project aims to reduce morbidity and mortality from

a
Oxford University Clinical Research Unit, 764 Vo Van Kiet Street, District 5, Ho Chi Minh City, 72708, Viet Nam.
b
Centre for Antimicrobial Optimisation, Imperial College London, England.
c
Hospital for Tropical Diseases, Ho Chi Minh City, Viet Nam.
d
Centre for Bio Inspired Technology, Imperial College London, England.
e
Department of Information Science, University of Otago, New Zealand.
f
Department of Health of Ho Chi Minh City, Viet Nam.
Correspondence to Sophie Yacoub (email: syacoub@​oucru​.org).
(Submitted: 11 November 2022 – Revised version received: 16 March 2023 – Accepted: 22 March 2023 – Published online: 26 May 2023 )

Bull World Health Organ 2023;101:487–492 | doi: http://dx.doi.org/10.2471/BLT.22.289423 487

 Lessons from the field
Applying artificial intelligence and digital health technologies, Viet Nam Ho Quang Chanh et al.

infectious diseases by using innovative pathways. First, we mapped workflow facilitate health data exchange across
technology and clinical approaches to and clinical priorities in our setting to different systems.
improve the management of critical characterize differences in management Furthermore, we acknowledged
care. This multidisciplinary project is a between patient cohorts which could the use of large secondary data sets for
collaboration between the Hospital for affect deployment.7 We then organized model development could introduce
Tropical Diseases; Oxford University 10 in-person interviews and two work- hidden biases that are difficult to detect
Clinical Research Unit in Ho Chi Minh shops with hospital staff who directly and account for in our models.9 While
City; Imperial College London; Uni- care for dengue patients, to explore patient diversity and large cohort size
versity of Oxford and other partners. their perspectives towards dengue and from pooled data were beneficial in
The implementation site is the 550-bed their needs and desired functionalities providing a representative sample over
public Hospital for Tropical Diseases, of tools for improving disease manage- time, the narrow scope and homogenous
which is a referral centre for infectious ment. We used the information obtained nature of research data can adversely
diseases for the south of Viet Nam. to guide the system frontend design. affect the downstream generalizability
As a part of the project, we have Finally, we conducted 50 separate obser- and real-world performance of artificial
been conducting research and imple- vations of 15 clinicians to evaluate the intelligence models.9 We are therefore
mentation studies to investigate the usability of the assessment tool. continuing to refine our models using
role and utility of novel digital health We engaged with hospital stake- data from our prospective ongoing
technologies for improving the manage- holders during the project’s conceptu- studies to validate the clinical utility and
ment of dengue, a major public health alization phase as well as throughout generalizability, and to reduce biases.
issue in Viet Nam. Our work includes the project operation. We also actively Furthermore, population groups un-
the development of a novel wearable, attended digital health technology derrepresented or excluded in research
and an artificial intelligence-driven workshops to engage with government data represent significant challenges
electronic clinical decision support sys- authorities and policy-makers including to address health inequity.10 Therefore,
tem, namely D-SCAPE (dengue severity the Vietnamese health ministry. continuous real-world evaluation is
classification and prediction wearable) needed to ensure clinical utility and
and D-CAT (dengue clinical assessment improve the performance through con-
tool), respectively.
Relevant changes tinuous learning.
In collaboration with engineers at We are working on refining the designs
Implementation
Imperial College London, we developed of the digital health technologies in
a wearable prototype that could optically our project. Based on patient feedback, New digital health technologies need to
obtain plethysmograms. While this first we are altering the wearable sensor to be sustainable, pragmatic, interoperable
protoype did not display any data col- increase comfort. After the workshop and aligned with existing health-care
lected, our aim is that future protypes participants had selected the core processes,11 as well as solving real-life
will show pulse, oxygen saturation and functionalities of the assessment tool, problems that matter to the population.
trends in haematocrit levels. We tested we built the user interface, which sub- To ensure that the project fulfilled these
the prototype on 50 dengue patients ad- sequently was tested for usability in an points, we discussed with stakehold-
mitted to the implementation hospital, iterative manner by the clinical staff ers before the implementation, and
and we explored patients’ perspectives members. conducted several activities through-
on design and use of the sensor. We also out the development of the digital
collected physiological data including Lessons learnt health technologies. Interviews with
pulse, pulse wave and oxygen saturation, individual health-care workers around
Data
from 250 hospitalized dengue patients decision-making processes during the
with different severity grades, using a Despite the potential of using artificial development revealed insights about the
commercial wearable named SmartCare intelligence models developed from usability of the clinical decision support
(SmartCare Analytics, Oxford, England) pooled data, we found seasonality and systems, and the need to balance clinical
to develop artificial intelligence models disease prevalence affected our model usefulness and digital complexity of the
for dengue physiological monitoring. performance.8 These findings highlight tool. Clinicians highlighted that digital
The development of the assessment the need for continuous data collection health technologies should not signifi-
tool consisted of training and testing methods to sustain model performance cantly alter existing pathways in terms of
artificial intelligence models alongside for dengue tools. As the data avail- data entry and patient flow.7 As process
implementation work, which included ability and quality issues hindered the mapping revealed different patient flow
landscape exploration, user interface application of automated pipelines for between adult and paediatric patients,
design and stakeholder usability test- data collection, pre-processing and the implementation of the assessment
ing. To develop the artificial intelligence sharing in our project, we emphasize tool should thus have flexibility ac-
predictive algorithms, we used existing the importance of strengthening the cording to the age of the patient.7 Data
research data sets, including pooled capacity for electronic medical records entry to two electronic systems (hospital
data from 8000 Vietnamese patients transition and interoperability. We are electronic medical records and the as-
collected over 19 years.6 Concurrently, therefore facilitating the upgrade of cur- sessment tool) would discourage uptake
we conducted a multistage qualitative rent hospital electronic medical records of the new tool in clinical practice. We
study to better understand existing to comply with the HL7 Fast Healthcare therefore recommend developers of
health-care pathways and the possible Interoperability Resources, an inter- digital health technologies to consider
impact of the interventions on these national industry standard, which will strategies for system integration into

488 Bull World Health Organ 2023;101:487–492| doi: http://dx.doi.org/10.2471/BLT.22.289423

 Lessons from the field
Ho Quang Chanh et al. Applying artificial intelligence and digital health technologies, Viet Nam

existing electronic medical records to

Box 1. Summary of main lessons learnt
maximize uptake and sustainability.
Furthermore, to increase long-term • An interoperable and appropriate plan for data management including collection, sharing
uptake, we asked about patients’ per- and integration is important.
ceptions towards the wearable sensor, • Engagements and implementation studies should be conceptualized and conducted
and we learnt that the patients and alongside the digital health technology development.
their relatives expressed concerns that • The priorities of end-users and understanding context and regulatory landscape are crucial
its use would reduce interactions with for success.
health-care workers. Although digital
literacy and the role of technology in
many low- and middle-income country Table 1. Challenges and potential solutions in applying digital health technologies,
settings are increasing at a significant lessons from Viet Nam
pace, better understanding of attitudes
towards digital health technologies is Challenges, by stage Potential solutions
needed. An implementation research ap- Data and development
proach is therefore crucial in any digital • Limited data availability for model • Strengthening capacity for data collection and
health technology project, and should development ensuring system interoperability
ideally run alongside development of the • Ensuring representative data is • Multidisciplinary collaboration (clinical, data science,
available, and model fairness information technology) for development work;
technology to better guide the iterative
support local workforce development and retention;
development and delivery. and specific focus on addressing model bias
During the project we had difficul- Implementation
ties in recruiting local artificial intelli-
• Difficulties to integrate digital • Co-design of interventions with focus on integration
gence researchers, hence local research health technologies in existing and minimizing disruption to workflow.
capacity needs to be boosted through health-care systems • Better understanding of health-care context through
collaboration with local academic • Enhancing uptake and behaviour methods including process mapping and human-
institutions, as well as investment in change centred design
higher education. Another important • Addressing concerns of end- • Interventions targeting digital literacy and
and complex challenge to the viability of users and bodies regarding engagement with relevant communities. Involve
familiarity of using digital health local institutions and policy-makers during public
many digital health technology projects technologies in local settings engagement
in low-resource settings is the migration
Legal and regulatory framework
of digital practitioners to high-income
• Lack of artificial intelligence • Initial adoption of existing guidelines while
settings. governance capability and local collaborating with stakeholders to continuously re-
Regulation expertise evaluate their relevance to local setting
• Limited experience with • Adoption of flexible frameworks and early
Recent released policies focusing on intellectual property engagement in intellectual property to ensure clear
enhancing electronic medical records • Varying digital ecosystem understanding of rights between stakeholders
interoperability, picture archiving, and maturity • Sharing experiences with regulatory bodies and using
systems to support communications research findings to inform and co-develop policies
relevant to the setting
between health-care bodies,12 provide
an opportunity to develop digital health
technologies within our setting. These
policies facilitate the digital transfor- limited. Adapting existing guidance on sensor based on feedback from patients;
mation, by increasing data availability artificial intelligence developed in high- validate the artificial intelligence models
and interoperability between systems. income settings can provide important embedded in the assessment tool; and
We engaged with government authori- insights,13,14 although all stakeholders further explore potential pathways for
ties and policy-makers, including the involved need to be cognizant of the regulatory submission.
health ministry, early on in our project relevance and direct applicability of the The lessons learnt (Box 1) from our
to better understand the priorities and guidance. Given the fast pace of digital project, as well as the challenges and
infrastructure investment. We found health technology development in Viet potential solutions in applying digital
that policy-makers were receptive to Nam, it is important to have a formal, health technologies (Table 1) in Viet
our experience and supported plans on iterative process led at the national Nam could be useful for other settings in
future roll-out. level, to capture local needs and develop low- and middle-income countries. We
Artificial intelligence-enhanced necessary regulations. 12 At the same suggest that challenges specific to low-
applications in health care can only time, heterogeneity between individual and middle-income countries need to be
create impact providing they exist health-care settings requires that suffi- explicitly considered, and often require
within a legal and regulatory framework. cient support and expertise are available a different approach than replicating
Understanding these environments is for local bodies (such as research ethics or extrapolating the solutions from the
important in facilitating collaboration committees) for evaluation of novel high-income countries. Application of
and safeguarding users.4 In Viet Nam, digital health technologies. artificial intelligence and digital health
we have found that the availability of lo- The project is still ongoing, and the technologies requires development
cal guidelines supporting development next steps in our dengue work package of digital infrastructure, placing local
of novel digital health technologies is are to refine the design of the wearable health-care workers and stakeholders

Bull World Health Organ 2023;101:487–492| doi: http://dx.doi.org/10.2471/BLT.22.289423 489

 Lessons from the field
Applying artificial intelligence and digital health technologies, Viet Nam Ho Quang Chanh et al.

at the centre of the implementation pro- Acknowledgements Tropical Medicine and Global Health,
cess, and co-development of regulatory HTT is also affiliated with the Oxford Nuffield Department of Medicine, Uni-
frameworks across different relevant University Clinical Research Unit, Ho versity of Oxford, England.
bodies.  Chi Minh City, Viet Nam. LT, CP and
SY are also affiliated with the Centre for Competing interests: None declared.

‫ملخص‬
‫ فييت نام‬،‫تطبيق تقنيات الذكاء االصطناعي والصحة الرقمية‬
،‫املواقع املحلية فييت نام هي دولة ذات دخل منخفض إىل متوسط‬ ‫املشكلة إن التطبيق املبارش لتقنيات الصحة الرقمية من األوضاع‬
‫ويعد نظام الرعاية الصحية فيها املرحلة الوليدة لتنفيذ تقنيات‬ ‫بالدول ذات الدخل املرتفع إىل الدول ذات الدخل املنخفض إىل‬
.‫الصحة الرقمية‬ ‫ قد يكون غري مناسب بسبب التحديات املتعلقة بتوافر‬،‫املتوسط‬
‫ نقوم بتغيري‬،‫التغرّيات ذات الصلة بنا ًء عىل مالحظات املريض‬ ّ ‫ ومن ثم فإن هناك حاجة إىل‬.‫ وتنظيمها‬،‫ وتنفيذها‬،‫البيانات‬
‫ قمنا بوضع واجهة‬.‫تصميم املستشعر القابل لالرتداء لزيادة الراحة‬ .‫أساليب خمتلفة‬
‫املستخدم اخلاصة بأداة التقييم بنا ًء عىل الوظائف األساسية املختارة‬ ‫األسلوب نحن نقوم ضمن مرشوع خمترب التطبيقات االنتقالية‬
‫ تم اختبار الواجهة الح ًقا للتحقق‬.‫للحارضين يف ورشة العمل‬ ‫ بتطوير جهاز‬2018 ‫ منذ عام‬،‫لوحدة العناية املركزة يف فييت نام‬
‫من قابليتها لالستخدام بطريقة تكرارية بواسطة أعضاء الفريق‬ ‫ وأداة للتقييم اإلكلينيكي‬،‫يمكن ارتداؤه ملراقبة املرىض من األفراد‬
.‫اإلكلينيكي‬ ‫ من خالل العمل عن كثب مع‬.‫لتحسني إدارة مرض محى الضنك‬
‫الدروس املستفادة إن تطوير تقنية الصحة الرقمية وتنفيذها حيتاج‬ ‫ هو تيش من‬،‫املوظفني املحليني يف مستشفى األمراض االستوائية‬
‫إىل خطة مناسبة وقابلة للتشغيل املتبادل إلدارة البيانات بام يف ذلك‬ .‫ قمنا بتطوير نموذج أويل للجهاز القابل لالرتداء واختباره‬،‫سيتي‬
‫ جيب وضع تصور لعمليات املشاركة‬.‫التجمع واملشاركة والتكامل‬ .‫حصلنا عىل منظور لتصميم واستخدام املستشعر من املرىض‬
‫ وإجرائها جن ًبا إىل جنب مع تطوير تقنية الصحة‬،‫ودراسات التنفيذ‬ ،‫ استخدمنا جمموعات بيانات البحث احلالية‬،‫لتطوير أداة التقييم‬
‫ وفهم السياق‬،‫ تعترب أولويات املستخدمني النهائيني‬.‫الرقمية‬ ‫ وأجرينا‬،‫وخططنا عمليات سري العمل واألولويات اإلكلينيكية‬
.‫أمرا بالغ األمهية لتحقيق النجاح‬ً ،‫واملشهد التنظيمي‬ ‫ وعقدنا ورش عمل مع موظفي‬،‫مقابالت مع أصحاب املصلحة‬
.‫املستشفى‬

摘要
越南应用人工智能和数字健康技术
问题 由于数据可用性、实施和监管方面存在的挑战， 当地状况 在中低收入国家越南，卫生保健系统正处于
将数字健康技术从高收入环境直接应用到中低收入国 实施数字健康技术的初级阶段。
家可能并不合适。因此，需要采用不同的方法。 相关变化 根据患者反馈，我们正在改进可穿戴传感器
方法 在越南 ICU 转化应用实验室项目中，我们自 的设计以提升舒适度。我们根据研讨会与会者选择的
2018 年以来一直在开发用于个体患者监测的可穿戴设 核心功能构建了评估工具的用户界面。随后，临床工
备和临床评估工具，以改善登革热疾病管理。我们与 作人员通过迭代方法测试了该界面的可用性。
胡志明市热带病医院的当地工作人员开展密切合作， 经验教训 数字健康技术的开发和实施需要一个可互操
开发并测试了可穿戴设备原型。我们获得了患者对传 作且适当的数据管理计划，包括数据的收集、共享和
感器设计和使用的看法。为了开发评估工具，我们采 集成。应将参与和实施研究概念化，并与数字健康技
用了现有的研究数据集，制定了工作流程和临床优先 术的开发一起进行。终端用户的优先事项以及对背景
事项，采访了利益相关者，并与医院工作人员一起举 和监管环境的了解对成功至关重要。
行了研讨会。

Résumé
Utilisation de l'intelligence artificielle et des technologies numériques en matière de santé au Viet Nam
Problème Transposer directement des technologies numériques en prototype de ce dispositif portable. Nous avons ainsi pu recueillir l'avis
matière de santé des pays à revenu élevé vers des pays à revenu faible et des patients quant à la conception et à l'utilisation du capteur. Pour
intermédiaire peut se révéler inapproprié en raison des défis que posent élaborer l'outil d'évaluation, nous avons employé des ensembles de
la disponibilité, la mise en œuvre et la réglementation des données. Par données de recherche existantes, cartographié les processus et les
conséquent, différentes approches sont nécessaires. priorités cliniques, interrogé les parties prenantes et organisé des ateliers
Approche Dans le cadre du projet Viet Nam ICU Translational avec le personnel de l'hôpital.
Applications Laboratory, nous développons depuis 2018 un dispositif Environnement local Le Viet Nam est un pays à revenu faible, où le
portable pour le suivi individuel des patients, ainsi qu'un outil déploiement des technologies numériques n'en est encore qu'à ses
d'évaluation clinique visant à améliorer la gestion de la dengue. En débuts dans le secteur des soins de santé.
étroite collaboration avec les équipes locales de l'Hôpital des maladies Changements significatifs En nous fondant sur les avis des patients,
tropicales, à Ho Chi Minh Ville, nous avons mis au point et testé un nous avons modifié le design du capteur portable en vue de le rendre

490 Bull World Health Organ 2023;101:487–492| doi: http://dx.doi.org/10.2471/BLT.22.289423

 Lessons from the field
Ho Quang Chanh et al. Applying artificial intelligence and digital health technologies, Viet Nam

plus confortable. Nous avons également imaginé l'interface utilisateur des données adéquat et interopérable, notamment en termes de
de l'outil d'évaluation en nous inspirant des principales fonctionnalités collecte, de partage et d'intégration. Des études sur l'engagement et
sélectionnées par les participants aux ateliers. Cette interface a ensuite le déploiement doivent être conceptualisées et menées parallèlement
été testée de manière itérative par les membres du personnel clinique au développement de telles technologies. Tenir compte des priorités
afin d'en mesurer la facilité d'utilisation. des utilisateurs finaux, mais aussi comprendre le contexte et le cadre
Leçons tirées L'élaboration et la mise en œuvre des technologies réglementaire est essentiel à la réussite de ce projet.
numériques en matière de santé requièrent un plan de gestion

Резюме
Применение искусственного интеллекта и технологий цифрового здравоохранения, Вьетнам
Проблема Прямое применение технологий цифрового Местные условия Во Вьетнаме, который относится к странам
здравоохранения, заимствованных из опыта стран с высоким со средненизким уровнем дохода, система здравоохранения
уровнем дохода, может быть нецелесообразным в странах с находится на начальной стадии внедрения технологий
низким и средним уровнем дохода из-за сложностей, связанных цифрового здравоохранения.
с доступностью данных, внедрением и регулированием. Поэтому Осуществленные перемены На основании отзывов пациентов
необходимы различные подходы. с целью повышения удобства использования была изменена
Подход В рамках проекта лаборатории трансляционных конструкция носимого датчика. На основании выбранных
приложений для ОИТ во Вьетнаме с 2018 года разрабатывается участниками семинара основных функциональных возможностей
носимое устройство для индивидуального мониторинга был разработан пользовательский интерфейс инструмента
пациентов и инструмент клинической оценки для более оценки. Впоследствии интерфейс был протестирован на удобство
эффективного ведения болезни денге. В тесном сотрудничестве использования итеративным способом сотрудниками клиники.
с местным персоналом больницы тропических болезней Выводы Для разработки и внедрения технологий цифрового
в г. Хошимине был разработан и протестирован прототип здравоохранения необходим функционально совместимый
носимого устройства. От пациентов были получены сведения и соответствующий план управления данными, включая
о конструкции и использовании датчика. Для разработки сбор, обмен и интеграцию. Исследования по вовлечению и
инструмента оценки были использованы доступные наборы внедрению должны разрабатываться и проводиться параллельно
данных исследований, сопоставлены рабочие процессы и с разработкой технологий цифрового здравоохранения.
клинические приоритеты, опрошены заинтересованные стороны Приоритеты конечных пользователей и понимание контекста
и проведены семинары с участием персонала больницы. и нормативно-правовой базы имеют решающее значение для
достижения успеха.

Resumen
Aplicación de la inteligencia artificial y las tecnologías sanitarias digitales en Vietnam
Situación La aplicación directa de las tecnologías sanitarias digitales Marco regional En Vietnam, un país de ingresos medios bajos, el
desde los entornos de ingresos altos a los países de ingresos bajos y sistema sanitario se encuentra en una fase incipiente de implantación
medios puede resultar inadecuada debido a los problemas relacionados de tecnologías sanitarias digitales.
con la disponibilidad de datos, la implementación y la regulación. Por Cambios importantes A partir de los comentarios de los pacientes, se
consiguiente, se necesitan enfoques diferentes. ha modificado el diseño del sensor portátil para aumentar su comodidad.
Enfoque Dentro del proyecto Viet Nam ICU Translational Applications Se creó la interfaz de usuario de la herramienta de evaluación según
Laboratory, se está desarrollando desde 2018 un dispositivo ponible las principales funciones que seleccionaron los asistentes al taller.
para la monitorización individual del paciente y una herramienta Posteriormente, los miembros del personal clínico comprobaron la
de evaluación clínica para mejorar el manejo de la enfermedad del usabilidad de la interfaz de manera iterativa.
dengue. En estrecha colaboración con el personal local del Hospital Lecciones aprendidas El desarrollo y la implementación de la
para las Enfermedades Tropicales de Ciudad Ho Chi Minh, se desarrolló tecnología sanitaria digital necesitan un plan interoperable y adecuado
y probó un prototipo del dispositivo ponible. Se obtuvo la opinión para la gestión de datos, que incluya la recopilación, el intercambio y la
de los pacientes sobre el diseño y el uso del sensor. Para desarrollar integración. Los compromisos y los estudios de implementación deben
la herramienta de evaluación, se utilizaron conjuntos de datos de conceptualizarse y llevarse a cabo junto con el desarrollo de la tecnología
investigación existentes, se trazaron flujos de trabajo y prioridades sanitaria digital. Las prioridades de los usuarios finales y la comprensión
clínicas, se entrevistó a las partes interesadas y se organizaron talleres del contexto y el panorama reglamentario son fundamentales para
con el personal del hospital. lograr los objetivos.

References
1. Wahl B, Cossy-Gantner A, Germann S, Schwalbe NR. Artificial intelligence 2. Ciecierski-Holmes T, Singh R, Axt M, Brenner S, Barteit S. Artificial
(AI) and global health: how can AI contribute to health in resource-poor intelligence for strengthening healthcare systems in low- and middle-
settings? BMJ Glob Health. 2018 Aug 29;3(4):e000798. doi: http://​dx​.doi​ income countries: a systematic scoping review. NPJ Digit Med. 2022 Oct
.org/​10​.1136/​bmjgh​-2018​-000798 PMID: 30233828 28;5(1):162. doi: http://​dx​.doi​.org/​10​.1038/​s41746​-022​-00700​-y PMID:
36307479

Bull World Health Organ 2023;101:487–492| doi: http://dx.doi.org/10.2471/BLT.22.289423 491

 Lessons from the field
Applying artificial intelligence and digital health technologies, Viet Nam Ho Quang Chanh et al.

3. Möllmann NRJ, Mirbabaie M, Stieglitz S. Is it alright to use 9. Näher AF, Vorisek CN, Klopfenstein SAI, Lehne M, Thun S, Alsalamah S, et al.
artificial intelligence in digital health? A systematic literature Secondary data for global health digitalisation. Lancet Digit Health. 2023
review on ethical considerations. Health Informatics J. 2021 Oct- Feb;5(2):e93–101. doi: http://​dx​.doi​.org/​10​.1016/​S2589​-7500(22)00195​-9
Dec;27(4):14604582211052391. doi: http://​dx​.doi​.org/​10​.1177/​ PMID: 36707190
14604582211052391 PMID: 34935557 10. Målqvist M, Hoa DTP, Liem NT, Thorson A, Thomsen S. Ethnic minority health
4. Global strategy on digital health 2020-2025. Geneva: World Health in Vietnam: a review exposing horizontal inequity. Glob Health Action.
Organization; 2021.Available from: https://​apps​.who​.int/​iris/​handle/​10665/​ 2013 Mar 4;6(1):1–19. doi: http://​dx​.doi​.org/​10​.3402/​gha​.v6i0​.19803 PMID:
344249 [cited 2023 Mar 12]. 23462107
5. Future of digital health in Vietnam [internet]. Hanoi: KPMG; 2020. Available 11. Mandel JC, Kreda DA, Mandl KD, Kohane IS, Ramoni RB. SMART on FHIR: a
from: https://​assets​.kpmg​.com/​content/​dam/​kpmg/​vn/​pdf/​publication/​ standards-based, interoperable apps platform for electronic health records.
2021/​digital​-health​-vietnam​-2020​-twopage​.pdf [cited 2023 Mar 15]. J Am Med Inform Assoc. 2016 Sep;23(5):899–908. doi: http://​dx​.doi​.org/​10​
6. Ming DK, Hernandez B, Sangkaew S, Vuong NL, Lam PK, Nguyet NM, et al. .1093/​jamia/​ocv189 PMID: 26911829
Vietnam ICU Translational Applications Laboratory (VITAL) investigators. 12. Tran DM, Thwaites CL, Van Nuil JI, McKnight J, Luu AP, Paton C. Vietnam
Applied machine learning for the risk-stratification and clinical decision ICU Translational Applications Laboratory (VITAL). Digital health policy
support of hospitalised patients with dengue in Vietnam. PLOS Digit Health. and programs for hospital care in Vietnam: scoping review. J Med Internet
2022 Jan 18;1(1):e0000005. doi: http://​dx​.doi​.org/​10​.1371/​journal​.pdig​ Res. 2022 Feb 9;24(2):e32392. doi: http://​dx​.doi​.org/​10​.2196/​32392 PMID:
.0000005 PMID: 36812518 35138264
7. Nguyen QH, Ming DK, Luu AP, Chanh HQ, Tam DTH, Truong NT, et al. 13. Ethics and governance of artificial intelligence for health: WHO guidance.
Vietnam ICU Translational Applications Laboratory (VITAL) investigators. Geneva: World Health Organization; 2021. Available from: https://​www​.who​
Mapping patient pathways and understanding clinical decision-making in .int/​publications/​i/​item/​9789240029200 [cited 2022 Oct 20].
dengue management to inform the development of digital health tools. 14. Evidence standards framework for digital health technologies. London:
BMC Med Inform Decis Mak. 2023 Feb 2;23(1):24. doi: http://​dx​.doi​.org/​10​ National Institute for Health and Care Excellence; 2022. Available from:
.1186/​s12911​-023​-02116​-4 PMID: 36732718 https://​www​.nice​.org​.uk/​corporate/​ecd7 [cited 2022 Oct 20].
8. Ming DK, Tuan NM, Hernandez B, Sangkaew S, Vuong NL, Chanh HQ, et al.
The diagnosis of dengue in patients presenting with acute febrile illness
using supervised machine learning and impact of seasonality. Front Digit
Health. 2022 Mar 14;4:849641. doi: http://​dx​.doi​.org/​10​.3389/​fdgth​.2022​
.849641 PMID: 35360365

492 Bull World Health Organ 2023;101:487–492| doi: http://dx.doi.org/10.2471/BLT.22.289423