As the needs of our ageing population grow in intensity and diversity, there is a need to achieve precision in public health via data-driven profiling of population-level preventive care, while optimising medical and social services to address those needs. These initiatives will maximise population health and minimise health care costs. Nevertheless, population-level precision public health research is rare; its application to drive service planning and deployment at the population level is even rarer.1 Thus, with support from the Strategic Public Policy Research Funding Scheme managed by the Policy Innovation and Co-ordination Office of the Hong Kong SAR Government, we initiated a research programme to fill the gap in precision public health research and practice by triangulating data that represent population-level socioecology,2 such as personal-level clinical and functional data, relational-level data for individual households, community-level data regarding socio-demographic characteristics and physical living environments, data describing organisations that meet population-level needs, and data reflecting the impacts of governmental policy. We sought to identify individuals who can receive the greatest benefit from primary, secondary, and tertiary preventive care. The resulting profiles could inform population-level planning and allocation of the three tiers of preventive care programmes.

Nevertheless, our research objectives were confronted with challenges related to the following contextual factors: (1) the inherent biases and quality of real-world data extracted from medical services’ Electronic Health Records (EHRs) and social services’ record systems; (2) the fragmentation among services (and their respective databases) which are required to address needs arising from specific aspects of population-level socioecology, including the distinct medical and social needs that our siloed medical and social services seek to address; and (3) the coronavirus disease 2019 (COVID-19) pandemic and the associated social and public health measures which emerged shortly after project initiation and have persisted throughout its life cycle. To overcome these challenges, we adopted a multi-source analytical approach,3 whereby parallel and iterative analyses were performed across databases representing different socioecology aspects at the resident level. Specifically, an analytical profile developed in one database was applied to other databases with the goal of identifying research questions and facilitating the selection of corresponding features and analytics. The findings from multiple siloed databases could be triangulated to coherently address individual research objectives. In addition, where applicable, parameters extracted from siloed databases were integrated to model particular outcomes using our artificial intelligence (AI) algorithm, for which the input architecture was anthropomorphised4 according to spheres described in the socioecological prevention framework of the Centers for Disease Control and Prevention. This approach enabled structuring of the hierarchically interrelated input layers.

In the following text, we describe our multi-source analytical approach and emerging findings from our research programme. Although the academic outputs of our research programme are in various stages of peer review, this description of a data-driven process to formulate research questions and develop sampling frames for examination across siloed databases in the construction of a population-level coherent care profile may serve as an alternative approach for other researchers to consider when they face similar contextual challenges in population-level precision public health research.

For example, using the study populations’ EHRs (obtained via the Hospital Authority Data Collaboration Laboratory), we applied unsupervised and supervised machine learning algorithms in tandem to identify tertiary prevention needs and the service gaps that prevent those needs from being met in the study populations. Our analyses revealed that the highest rehospitalisation rates (>80%) and the shortest times between discharge and rehospitalisation occurred in sub-populations of patients who lacked specific ambulatory or postacute services. Nonetheless, these services were also available to patients who shared similar clinical and utilisation profiles but exhibited significantly lower rehospitalisation rates. Among the sub-populations with high rehospitalisation rates and low utilisation of rehospitalisation-mitigating post-discharge services, one had a typical profile (ie, population segment medoids) of patients aged 50 to 64 years with musculoskeletal pain–related disorders as primary diagnoses. These patients more frequently exhibited a history of multiple chronic illnesses and higher clinical complexity at index hospitalisation compared with other patients who had similar clinical and acute care utilisation profiles.

The profiling of sub-populations who fell through the service gaps and were rehospitalised at the highest rate enabled us to bring precision to tertiary prevention efforts and subsequently perform data-driven optimisation of population-level post-discharge service allocation, thereby minimising medical costs. Furthermore, the profile we constructed from EHRs could also be applied beyond medical settings to identify potential secondary prevention targets that may exacerbate the evolution of an underlying disease process, such that it interfered with quality of life among individuals who matched the EHR-based and machine-constructed profile, ultimately triggering health-seeking behaviour.

Thus, in a non-medical setting, we recruited residents of the study population aged 50 to 64 years who had musculoskeletal pain, according to community-based primary care clinicians. In addition to the residents’ socio-demographic characteristics, behavioural health, and co-morbid chronic illness statuses, clinicians also assessed anthropometric measures and biomarkers of metabolic dysfunction that are often direct or indirect precursors to the most common forms of chronic illnesses. These factors were included as predictive features in a random forest model for selection and risk-scoring of potential secondary prevention targets that could mitigate the exacerbation of pain symptoms. The model also included features representing various aspects of the residents’ living environments, which were separately parameterised and initially selected by our AI algorithm according to the following constraints: (1) they were sourced from multiple public domain datasets that belonged to governmental agencies such as the Census and Statistics Department, Housing Authority, Lands Department, Department of Health, and District Offices; (2) they were organised as layered input into a multi-headed hierarchical convolutional neural network, with an anthropomorphised architecture that captured the study population’s internal and external built environments and socio-demographic profiles; and (3) they were selected according to the statistical importances of their unique and combined contributions to residential building-level aggregates of general health based on census data and COVID-19 case counts from the Department of Health.

Finally, after parameterisation and selection in accordance with their degrees of importance to the population’s general health and COVID-19 susceptibility, features representing the built environments of the study district’s residential buildings were processed as follows: (1) they were entered into a random forest model together with the aforementioned individual-level measures to compare their respective importances in the onset of pain interference; and (2) they were scored according to their individual and combined adverse health effects, then assigned to individual residential buildings in the study district for optimised allocation of local primary prevention programmes.

Our analyses revealed that, although features representing residents’ socio-demographic characteristics and metabolic dysfunction had high importance with respect to the presence of pain interference in various residential quality of life domains, their feature importances were secondary to the importances of built-environment features, such as living area size, air quality, access to light, architecture conducive to social connectivity, and building age. In addition to scoring the risk of pain interference for individual residents, we scored the built environment of each building in public housing estates within the study district according to the likelihood that its residents would experience sufficient pain to interfere with their quality of life. This scoring approach can inform service planning in geospatially targeted secondary pain prevention programmes.

Patients with chronic obstructive pulmonary disease who exhibited high clinical complexity and multiple co-morbidities were another sub-population who typically exhibited high rehospitalisation rates and low utilisation of rehospitalisation-mitigating post-discharge services. This patient profile was used to guide the recruitment of study district residents outside of medical settings, enabling examination of the evolution of disease processes and hospitalisation trends among asymptomatic and symptomatic community residents. Together with the findings regarding musculoskeletal pain and health-related effects of the built environment, our work has provided the basis for a predictive AI platform that was commissioned by the Sham Shui Po District Office to support its social health surveillance and policy decision needs. Additionally, our work has been incorporated into an algorithm deployed at community diagnosis events hosted by the Sham Shui Po District Office and at events co-hosted by the Kwai Tsing Safe Community and Healthy City Association and the Kwai Tsing District Office.

The work described in the current editorial is made possible with the support of the Strategic Public Policy Research Funding Scheme (project No: S2019.A4.015.19S). The authors thank Dr Jingjing Guan's analytical leadership, Mr Sam Ching's data science management, Ms Olivia Lam's data analytics and visualisation, Ms Yilin He's data wrangling, and Ms Hilliary Yee's data collection. The authors are also deeply grateful for the partnertships of Health In Action and People Service Centre, who have granted us permission to analyse the data under their custodianship under strict confidentiality agreements that safeguard the anonymity of their clients while driving improvements in their respective services.

Concept or design: E Leung, A Lee, H Tsang.
Acquisition of data: E Leung.
Analysis or interpretation of data: E Leung, A Lee.
Drafting of the manuscript: E Leung, A Lee.
Critical revision of the manuscript for important intellectual content: All authors.

All authors had full access to the data, contributed to the study, approved the final version for publication, and take responsibility for its accuracy and integrity.

As editor and adviser of the journal, respectively, MCS Wong and E Leung were not involved in the peer review process. Other authors have disclosed no conflicts of interest.

This editorial is supported by the Strategic Public Policy Research Funding Scheme (project No.: S2019.A4.015.19S). The funder had no role in study design, data collection/analysis/interpretation or manuscript preparation.

1. Talias MA, Lamnisos D, Heraclides A. Data science and health economics in precision public health. Front Public Health 2022;10:960282. Crossref

2. Centers for Disease Control and Prevention and Health Resources and Services Administration. 2022. The social-ecological model: a framework for prevention. Available from: https://www.cdc.gov/violenceprevention/about/social-ecologicalmodel.html. Accessed 8 Dec 2023.

3. Noi E, Rudolph A, Dodge S. Assessing COVID-induced changes in spatiotemporal structure of mobility in the United States in 2020: a multi-source analytical framework. Int J Geogr Inf Sci 2022;36:585-616. Crossref

4. Glikson E, Woolley AW. Human trust in artificial intelligence: review of empirical research. Acad Manag Ann 2020;14:627-60. Crossref

Over the past decade, there has been an explosion in the development and use of digital technologies in health and health-related areas. In 2018, the 71st World Health Assembly Resolution on Digital Health demonstrated collective recognition of the contributions of digital technologies to improving health, reducing health inequalities, and enhancing healthcare services in the context of achieving the Sustainable Development Goals laid down by the United Nations.1 The increasing popularity of digital tools, wearable devices, information systems, and electronic resources in clinical practices and health services has resulted in unique opportunities to reshape healthcare in response to diverse existing and emerging health system challenges. Furthermore, technological innovations have been evolving at an unprecedented scale, transforming the ways in which medicine is practised. This transformation has created a range of opportunities for telemedicine and mobile health to transform service delivery, and for advanced ‘big data’ and artificial intelligence approaches to enhance evidence-based decision support. At the global level, the World Health Organization established an e-health vision in its Global Strategy on Digital Health 2020-2025, with strategic objectives and an action framework to support countries in various development contexts when expanding the implementation of digital health technologies.2

In this issue of the Hong Kong Medical Journal, two original articles report survey findings regarding the perception and acceptance of telemedicine, a service which is rapidly expanding to overcome distance barriers in healthcare delivery.3 4 Hung et al3 analysed the experiences of individuals who used telemedicine during the coronavirus disease 2019 (COVID-19) pandemic in Hong Kong. They found a high level of satisfaction with telemedicine consultations; users felt that such consultations were useful in disease diagnosis and management. Choi et al4 explored the values, concerns, and expectations associated with telemedicine among Hong Kong adults aged ≥60 years in two hypothetical scenarios: during a severe outbreak while under government-imposed lockdown, and after the COVID-19 pandemic. The results of both studies supported the use of high-quality telemedicine as a novel approach to enhance clinical consultations and patient education, while emphasising the need for government- and provider-level support to promote and expand services.

There has been global recognition of the power that digital health technologies (eg, telemedicine) have to exchange information for disease diagnosis, treatment, and prevention. For example, the use of telemedicine technologies to educate patients and train community care providers is included in an innovative stepwise approach recommended by the World Kidney Day Steering Committee to improve service affordability and access for patients with kidney disease and their care partners in low-resource settings.5 The increasing access to digital resources and growing popularity of electronic health records have substantially supported patient self-education and public advocacy regarding kidney disease awareness and learning, thereby bridging gaps in kidney health education and literacy.6 Mobile messaging applications and social media platforms, characterised by multi-channel information dissemination and knowledge sharing, also play key roles in meeting the need for community empowerment and public engagement through digital health communications.7

In addition to the collaborative efforts of healthcare professionals and scientists to navigate challenges arising during the COVID-19 pandemic, digital health technologies have significantly contributed to the widespread adoption of the quick-response code–based contact tracing system in many countries. In particular, the LeaveHomeSafe mobile application in Hong Kong has enabled the public to more accurately record the date and time of entering and exiting various locations.8 There has been a remarkable increase in the use of artificial intelligence—a cutting-edge computing science innovation—to inform diagnosis, prognosis, treatment, and triage decisions across clinical settings. As summarised in a recent scoping review, 66 artificial intelligence products and tools have been used in the healthcare response to COVID-19, including pulmonary evaluations, assessments of infection risk, personalised care recommendations, triage decisions, patient deterioration monitoring, and predictions of disease severity.9 Another scoping review specifically examined the cost savings, performance in improving health outcomes, workflow efficiency in treatment and diagnosis, local feasibility, user friendliness, and reliability and trust associated with the implementation of artificial intelligence in low- and middle-income countries.10 Innovations such as clinical decision support systems, treatment planning and triage assistants, and health chatbots have demonstrated the potential to strengthen healthcare systems.10

Regarding the management of arterial hypertension, which is the most important contributor to the global burden of disease, the 2023 European Society of Hypertension Guidelines recommend the use of internet-based, interactive digital interventions in home blood pressure monitoring to enhance the digital storage and transfer of home blood pressure data, and to facilitate evaluation of those data by healthcare professionals.11 Remote clinical management programmes based on standardised home blood pressure monitoring supported by automatic transmission via mobile applications, along with collaborations involving multiple healthcare providers in the context of team-based care, could help reduce nonadherence to antihypertensive treatment. Meta-analyses have shown that virtual care for hypertension, mediated by telemonitoring and smartphone applications, provides benefits such as better patient education, greater blood pressure reduction, and improved cardiovascular outcomes.11 A scientific statement from the American Heart Association has affirmed the utility of telehealth in risk factor modification, medication adherence, and symptom monitoring during the management of various cardiovascular diseases.12

Ophthalmology is another branch of medicine that has closely embraced new models of care to improve patient-physician interactions through digital health innovations, such as multipurpose mobile applications, community-based teleconsultation units, and medical chatbots for improved case triage.13 Additionally, the screening and management of diabetic retinopathy—a major complication of diabetes mellitus and leading cause of preventable blindness worldwide—has been augmented by advances in healthcare digitisation and increasing emphasis on telehealth initiatives. In primary care and community settings, deep learning–based artificial intelligence for automated image-recognition, combined with telemedicine programmes based on low-cost devices and remote interpretation, would enable greater population coverage and facilitate timely referral to ophthalmic specialists for the management of vision-threatening conditions.14

Digital infrastructure can also play a central role in efforts to support and expand research capacity. As accurate and reliable sources of research data, electronic patient record systems have been extensively used in epidemiological investigations of clinical manifestations, radiological characteristics, laboratory results, and biomarkers.15 16 17 18 The use of electronic clinical management systems for patient screening and data collection to identify socio-economic factors, as well as health-protective and health-damaging behaviours associated with quality of life and health outcomes, was demonstrated in a study of childhood cancer survivors in Hong Kong.19

Despite potential risks and challenges related to oversight, regulations, data protection, and privacy—the focus of stepwise capacity-building efforts and mitigation strategies—digital health innovations have been implemented worldwide. Considering the rapid growth and development of digital health technologies, the use of telemedicine, e-health, and artificial intelligence as integral components of routine health service delivery is revolutionising medicine and health; the greatest impacts involve management of the increasingly complex conditions and circumstances encountered in primary care.20 These innovations and advancements will benefit medical education, clinical practice, and healthcare delivery, thereby ensuring service quality, accessibility, and affordability. In terms of effectiveness, acceptability, and feasibility, studies with rigorously designed methodologies in various contexts are needed to formulate evidence-based recommendations regarding the use of digital health technologies.

All authors contributed to the editorial, approved the final version for publication, and take responsibility for its accuracy and integrity.

The authors have declared no conflicts of interest.

1. World Health Organization. WHO guideline: recommendations on digital interventions for health system strengthening. Geneva: World Health Organization; 2019.

2. World Health Organization. Global strategy on digital health 2020-2025. Geneva: World Health Organization; 2021.

3. Hung KK, Chan EY, Lo ES, Huang Z, Wu JC, Graham CA. User perceptions of COVID-19 telemedicine testing services, disease risk, and pandemic preparedness: findings from a private clinic in Hong Kong. Hong Kong Med J 2023;29:404-11. Crossref

4. Choi MC, Chu SH, Siu LL, et al. Telemedicine acceptance by older adults in Hong Kong during a hypothetical severe outbreak and after the COVID-19 pandemic: a cross-sectional cohort survey. Hong Kong Med J 2023;29:412-20. Crossref

5. Kalantar-Zadeh K, Li PK, Tantisattamo E, et al. Living well with kidney disease by patient and care partner empowerment: kidney health for everyone everywhere. Hong Kong Med J 2021;27:97-8. Crossref

6. Langham RG, Kalantar-Zadeh K, Bonner A, et al. Kidney health for all: bridging the gap in kidney health education and literacy. Hong Kong Med J 2022;28:106.e1-8. Crossref

7. Wang HH, Li YT, Wong MC. Leveraging the power of health communication: messaging matters not only in clinical practice but also in public health. Hong Kong Med J 2022;28:103-5. Crossref

8. Li VW, Wan TT. COVID-19 control and preventive measures: a medico-legal analysis. Hong Kong Med J 2021;27:224-5. Crossref

9. Mann S, Berdahl CT, Baker L, Girosi F. Artificial intelligence applications used in the clinical response to COVID-19: a scoping review. PLOS Digit Health 2022;1:e0000132. Crossref

10. Ciecierski-Holmes T, Singh R, Axt M, Brenner S, Barteit S. Artificial intelligence for strengthening healthcare systems in low- and middle-income countries: a systematic scoping review. NPJ Digit Med 2022;5:162. Crossref

11. Mancia G, Kreutz R, Brunström M, et al. 2023 ESH Guidelines for the management of arterial hypertension The Task Force for the management of arterial hypertension of the European Society of Hypertension Endorsed by the International Society of Hypertension (ISH) and the European Renal Association (ERA). J Hypertens 2023 Jun 21. Epub ahead of print.

12. Takahashi EA, Schwamm LH, Adeoye OM, et al. An overview of telehealth in the management of cardiovascular disease: a scientific statement from the American Heart Association. Circulation 2022;146:e558-68. Crossref

13. Tham YC, Husain R, Teo KY, et al. New digital models of care in ophthalmology, during and beyond the COVID-19 pandemic. Br J Ophthalmol 2022;106:452-7. Crossref

14. Vujosevic S, Limoli C, Luzi L, Nucci P. Digital innovations for retinal care in diabetic retinopathy. Acta Diabetol 2022;59:1521-30. Crossref

15. Wang Y, Luo S, Zhou CS, et al. Clinical and radiological characteristics of COVID-19: a multicentre, retrospective, observational study. Hong Kong Med J 2021;27:7-17. Crossref

16. Baig NB, Chan JJ, Ho JC, et al. Paediatric glaucoma in Hong Kong: a multicentre retrospective analysis of epidemiology, presentation, clinical interventions, and outcomes. Hong Kong Med J 2021;27:18-26. Crossref

17. Tam EM, Kwan YK, Ng YY, Yam PW. Clinical course and mortality in older patients with COVID-19: a cluster-based study in Hong Kong. Hong Kong Med J 2022;28:215-22. Crossref

18. Kwok CC, Wong WH, Chan LL, et al. Effects of primary granulocyte-colony stimulating factor prophylaxis on neutropenic toxicity and chemotherapy dose delivery in Chinese patients with breast cancer who received adjuvant docetaxel plus cyclophosphamide chemotherapy: a retrospective cohort study. Hong Kong Med J 2022;28:438-46. Crossref

19. Cheung YT, Yang LS, Ma JC, et al. Health behaviour practices and expectations for a local cancer survivorship programme: a cross-sectional study of survivors of childhood cancer in Hong Kong. Hong Kong Med J 2022;28:33-44. Crossref

20. Wang HH, Li YT, Wong MC. Strengthening attributes of primary care to improve patients’ experiences and population health: from rural village clinics to urban health centres. Hong Kong Med J 2022;28:282-4. Crossref

Type 2 diabetes mellitus (DM) and hypertension combined were responsible for over 200 million DALYs (disability-adjusted life years) worldwide, representing the sum of years of life lost to premature mortality plus years lived with disability arising from these two highly prevalent conditions.1 2 Population-based preventive measures (eg, promotion of healthy lifestyles and targeted screening of at-risk individuals) followed by early team-based intervention have been effective strategies for reducing the associated morbidity, mortality, and healthcare burden.3 4 In countries with mature primary healthcare systems, these services are led by family doctors (ie, general practitioners) who partner with multidisciplinary healthcare teams to provide personalised, continuous, and comprehensive holistic care for all community residents.5 6 7

Hong Kong has a treatment-oriented healthcare system, in which 90% of hospital-based services are provided by the public sector8 and approximately 50% of public general outpatient services are used to manage DM and hypertension.9 Population-based preventive care initiatives, such as anti-smoking campaigns, the Colorectal Cancer Screening Programme (CRCSP) and the Vaccination Subsidy Scheme (VSS), have been successful but sporadic; they have limited potential to empower participants to pursue healthy living over life course. There is no structured cardiovascular disease screening programme in the public sector for most at-risk citizens without their own family doctors,10 except for older adults aged ≥65 years. Consequently, the public healthcare system has been heavily strained by the increasing prevalence of chronic diseases such as hypertension and DM, along with complications resulting from delayed diagnosis.

To improve healthcare system sustainability and overall population health, the Government recognises the urgent need to establish a prevention-oriented primary healthcare system. The Primary Healthcare Blueprint, issued in December 2022, highlighted key areas of development needed to address gaps in preventive care, continuity of care, and community participation. An important strategy is to implement the ‘Family Doctor for All’ concept by establishing a family doctor registration system: the Primary Care Register (PCR).11 The PCR is intended to build a recognition system for doctors who are committed to providing comprehensive, continuous, and holistic care to patients in the community; such care ranges from preventive services to chronic disease management.

Private doctors have been providing approximately 70% of episodic outpatient care in the community.8 However, the health advocacy potential of family doctors was not fully recognised by the public until the expansion of family medicine training in 2004. Beginning in 2013, the Government established a Primary Care Directory (PCD) to recruit primary care doctors (ie, any doctor in the private sector who was committed to providing primary care) for participation in various prevention-based programmes, such as the VSS and the CRCSP. To promote the community participation necessary for desired health improvements, the Government subsidised participants for each consultation conducted within these programmes. Because PCD registration was a prerequisite for receipt of related subsidies, PCD doctors were often engaged by the programme-based funding model. However, this model does not promote continuity of care if participants choose to consult PCD doctors only for specific preventive care services, or if the PCD doctors choose to only provide the specific preventive care services to the scheme participants during consultation.

To encourage community-based management of stable chronic diseases and reduce the service burden within the public healthcare sector, the Hospital Authority (HA) implemented the General Outpatient Clinic Public-Private Partnership Programme as an outsourcing service, beginning in 2014.12 However, the working relationship was unilateral. As the funder, the HA would purchase doctors’ services for specific tasks at prices determined through a bidding process; as service providers, PCD doctors would charge fees based on the service agreement. Benefits to patients were not considered in this programme. Similarly, doctors’ efforts to deliver holistic care beyond the scope of the service agreement were not appreciated or supported. Both models regarded PCD doctors as Government agents for service delivery and contributed to the fragmentation of care. Thus, these models failed to encourage the establishment of long-term doctor-patient partnerships necessary to enhance overall population health through patientcentric care.

In contrast, the planned PCR recognises the robust potential of family doctors. Under the PCR, each citizen will be paired with their preferred family doctor; each paired family doctor will be the only doctor eligible to receive any subsidy allocated to the paired patient, including existing programme-based subsidies (eg, the VSS, the CRCSP, and the General Outpatient Clinic Public-Private Partnership Programme) and any future initiatives to support primary healthcare. In addition to subsidies, efforts to optimise holistic care require community-based multidisciplinary team support for family doctors. Community drug formularies will be established to ensure that family doctors have access to common medications at affordable prices, which will facilitate long-term patient management. Community nurses and allied health professionals at District Health Centres will empower patients in leading healthy lifestyles and managing their own health. To encourage best practices, family doctors who have fulfilled their preventive and chronic disease care obligations, such as the provision of seasonal influenza vaccination, will be rewarded through additional payments.

The Chronic Disease Co-Care Pilot Scheme (the Scheme) targeting DM and hypertension, which will be launched in November 2023, represents the prototype service model under the planned PCR. People aged ≥45 years without a known diagnosis of DM or hypertension will be eligible for enrolment in a subsidised screening programme consisting of laboratory tests and a medical consultation with a paired family doctor registered in the current PCD. Healthy participants will be offered education regarding a healthy lifestyle and the opportunity to undergo repeat screening every 3 years. Participants with prediabetes, DM, and/or hypertension will receive subsidised care, including laboratory investigations for chronic disease monitoring, from their paired family doctor and a multidisciplinary team in the community.

To support PCD doctors in this new role, the Scheme incorporates seven key components that will shape the future primary healthcare system when the Primary Healthcare Commission (PHC) is established in 2024. These components include: (1) pairing of family doctors with participants; (2) District Health Centres and their services; (3) subsidised multidisciplinary services in the community; (4) protocol-driven bi-directional referral with designated medical specialist clinics under the HA; (5) pay-for-performance incentives for both participants and family doctors; (6) community drug formularies to ensure that family doctors have access to common medications at affordable prices; and (7) uniform data sharing in the Electronic Health Record Sharing System platform with participant consent. An important objective of the Scheme is to attract and build a pool of future PCR family doctors who agree with our vision and are committed to delivering quality primary care for our fellow citizens. The ultimate goals of the Scheme are to shift population-level health-seeking behaviour from treatment-oriented to prevention-focused, to encourage shared responsibility for personal health at an affordable cost, to enable family doctors to maintain continuity of care, and to improve health for all.

The Scheme establishes a new framework for primary healthcare involving family doctors and community services. Upon establishment of the PHC, a quality assurance system will be constructed to guide clinical practice among healthcare professionals via service quality standards and reference frameworks. The PHC will also monitor the performance of subsidised services and drive continuous quality improvement through pay-forperformance incentives. Additional subsidy tiers will be established based on clinical complexity and doctors’ qualifications. Hopefully, the Scheme will encourage more doctors to enrol in the PCD, and work as Family Doctors to provide continuous, comprehensive, and patient-centric care to all citizens.

All authors had full access to the data, contributed to the study, approved the final version for publication, and take responsibility for its accuracy and integrity.

The authors have declared no conflict of interest.

This editorial received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

1. Safiri S, Karamzad N, Kaufman JS, et al. Prevalence, deaths and disability-adjusted-life-years (DALYs) due to type 2 diabetes and its attributable risk factors in 204 countries and territories, 1990-2019: results from the Global Burden of Disease Study 2019. Front Endocrinol (Lausanne) 2022;13:838027. Crossref

2. World Health Organization. The Global Health Observatory: blood pressure/ hypertension. 2023. Available from: https://www.who.int/data/gho/indicator-metadata-registry/imr-details/3155. Accessed 18 Aug 2023.

3. ElSayed NA, Aleppo G, Aroda VR, et al. 3. Prevention or delay of type 2 diabetes and associated comorbidities: Standards of Care in Diabetes—2023. Diabetes Care 2023;46(Suppl. 1):S41-8. Crossref

4. Carey RM, Muntner P, Bosworth HB, Whelton PK. Prevention and control of hypertension: JACC Health Promotion Series. J Am Coll Cardiol 2018;72:1278-93. Crossref

5. Government of Canada. Primary health care. 2015. Available from: https://www.canada.ca/en/health-canada/services/primary-health-care.html. Accessed 18 Aug 2023.

6. Australian Institute of Health and Welfare. Primary health care in Australia. 2021. Available from: https://www.aihw.gov.au/reports/primary-health-care/primary-health-care-in-australia/contents/about-primary-health-care. Accessed 18 Aug 2023.

7. National Health Service, the United Kingdom Government. Primary care. 2020. Available from: https://www.longtermplan.nhs.uk/areas-of-work/primary-care/. Accessed 18 Aug 2023.

8. Schooling CM, Hui LL, Ho LM, Lam TH, Leung GM. Cohort profile: ‘children of 1997’: a Hong Kong Chinese birth cohort. Int J Epidemiol 2012;41:611-20. Crossref

9. Health Bureau, Hong Kong SAR Government. Legislative Council Brief. Primary Healthcare Blueprint. 2023. Available from: https://www.legco.gov.hk/yr2023/english/brief/hb202301_20230120-e.pdf. Accessed 18 Aug 2023.

10. Census and Statistics Department, Hong Kong SAR Government. Thematic Household Survey Report No. 74. 2021. Available from: https://www.censtatd.gov.hk/en/data/stat_report/product/C0000015/att/B11302742021XXXXB0100.pdf. Accessed 18 Aug 2023.

11. Health Bureau, Hong Kong SAR Government. Primary Healthcare Blueprint. 2022. Available from: https://www.primaryhealthcare.gov.hk/en/. Accessed 29 Aug 2023.

12. Electronic Health Record Sharing System, Hong Kong SAR Government. Public-Private Partnership Programmes (PPP). 2020. Available from: https://www.ehealth.gov.hk/en/whats-new/partnership/public-private-partnership-programmes.html. Accessed 1 Sep 2023.

In 1942, TL Terry was the first to report a condition that he termed ‘retrolental fibroplasia’, which developed in premature infants with low birth weight (BW)—this condition is now known as retinopathy of prematurity (ROP).1 Insufficient retinal vasculature development can lead to abnormal blood vessel growth, typically at the junction of the peripheral vascular and avascular retina. Subsequent fibrosis onset can result in retinal detachment and fibrovascular mass formation behind the crystalline lens (ie, retrolental fibroplasia). Oxygen therapy contributes to ROP onset.2 In the vasoconstrictive phase, oxygen can inhibit retinal vascularisation and suppress the production of vascular endothelial growth factor (VEGF). During the vasoproliferative phase, increased VEGF levels can cause neovascularisation and retinal blood vessel dilatation. Meticulous control of hyperoxia (arterial oxygen saturation >92%-93%) and avoidance of fluctuations in arterial oxygen saturation could prevent severe ROP.3

Appropriate treatment can protect against ROP-related blindness. Treatment largely depends on the location (zone) and severity of neovascularisation (stage), as well as a confirmed need for treatment. Historically, ROP was initially managed by avascular retina–targeted cryotherapy to reduce ischaemic drive. In the Cryotherapy for ROP (CRYO-ROP) study, adverse outcomes (retinal detachment, macula fold or retrolental mass) were reduced by almost 50% in eyes that received cryotherapy.4 In the 2000s, laser photocoagulation largely replaced cryotherapy as conventional treatment. The Early Treatment for ROP (ETROP) trial established standard treatment recommendations for type 1 (treatment-warranted) ROP: zone II ROP, stage 2 or 3 with plus disease, and zone 1 ROP, stage 3 with or without plus disease.5 Since then, the intravitreal injection of anti-VEGF drugs, including bevacizumab, ranibizumab, and aflibercept, has gained broad acceptance in the treatment of ROP; laser is still a popular option for primary therapy as well as rescue therapy (eg, in cases of disease reactivation and persistent avascular retina). The Bevacizumab Eliminates the Angiogenic Threat of ROP (BEAT-ROP) study showed promising results when bevacizumab was used in the treatment of stage 3 ROP; the retreatment rate was 4%, compared with 22% in the laser group.6 The ranibizumab compared with laser therapy for the treatment of infants born prematurely with ROP (RAINBOW) study also revealed excellent treatment success in 80% of infants receiving ranibizumab, compared with 66% of infants receiving laser therapy.7 Late complications, such as high myopia (-5 dioptres or worse), were less frequent after ranibizumab (5%) than after laser therapy (16%).8 Systemic complications did not differ between groups; the incidences of motor and hearing problems were similar.8 However, anti-VEGF therapy is not a panacea for ROP; reactivation or delayed progression of peripheral retina vascularisation may occur after injection.9 Therefore, recent ROP treatment guidelines from The Royal College of Ophthalmologists recommend close monitoring after anti-VEGF injection therapy.10

This new paradigm of ROP treatment requires an update to the classification of ROP. The International Classification of ROP, Third Edition (ICROP3) refined classification metrics such as posterior zone II, notch, and subcategorisation of stage 5; it also recognised the existence of a continuous spectrum of vascular abnormalities (ie, from normal to plus disease).11 The term ‘aggressive ROP’ replaced the term ‘aggressive-posterior ROP’ because of increasing awareness of aggressive ROP onset in larger infants, which extends beyond the posterior retina in regions of limited resources.

Modern advances in neonatal care have greatly improved premature infant survival. However, this improvement has led to an increase in ROP incidence, especially in middle-income countries (eg, India and China).12 In less developed countries or remote areas, telemedicine is increasingly important for ROP screening. Fundus photographs can be taken by nurses or technicians; screening can then be conducted remotely by ophthalmologists who specialise in ROP. This approach avoids the physical stress and financial cost involved in transporting high-risk infants; it also minimises screening delays. The Stanford University Network for Diagnosis of ROP (SUNDROP), a telemedicine-based screening initiative covering six satellite neonatal intensive care units in northern California of the United States (US), has screened 608 infants over 6 years. Its screening sensitivity of 100% and specificity of 99.8% are comparable with bedside clinical examination.13 Furthermore, the use of deep learning and federated learning for automatic diagnosis of ROP is under extensive investigation and may be important in future clinical management.14 15 Technical, medicolegal, regulatory, and financial aspects require consideration.

Local investigators have provided valuable data regarding the incidence and visual outcomes of ROP in Hong Kong. From 2007 to 2012, the incidences of ROP and type 1 ROP were 18.5% and 3.7%, respectively, among 513 infants at Caritas Medical Centre.16 Incidences were similar at Queen Mary Hospital in 2013 (16.9% and 3.4%, respectively).17 However, incidences at Prince of Wales Hospital were higher (31% and 4.5%, respectively) among 754 infants from 2007 to 2012.18 This discrepancy may be related to an increase in premature infant survival.18 In a study of 14 infants with type 1 ROP, one (7%) developed retinal detachment, nine (64%) developed amblyopia, and nine (64%) developed strabismus.19

Because ROP is a leading preventable cause of childhood blindness, screening protocol adherence is essential. The 2022 United Kingdom (UK) ROP screening protocol recommends examination of all infants born at gestational age (GA) ≤31 weeks and 6 days or with BW <1501 g.20 These thresholds differ from the US screening protocol (GA ≤30 weeks and 0 days or BW ≤1500 g).21 Because of the GA difference, fewer infants would be screened using the US protocol. This modified screening approach would reduce stress on premature infants, limit systemic absorption of dilating eye drops, and eventually lower medical costs. Currently, most hospitals under the Hospital Authority follow the UK protocol.

In this issue of the Hong Kong Medical Journal, Iu et al22 evaluated whether the use of the US protocol could reduce the number of infants screened without compromising the type 1 ROP detection sensitivity. The authors reviewed the clinical records of premature infants screened at Prince of Wales Hospital from 2009 to 2018; they found that if the US protocol had been followed, the number of infants requiring screened would have decreased by 21.1%. Using the US protocol, the investigators found that only 1.7% of cases would have been missed; all missed cases would have been mild ROP that did not require treatment.

However, conventional screening protocols have their own limitations, primarily because they are solely based on GA and BW. Many potentially unnecessary examinations are conducted to identify the approximately 20% of infants requiring treatment. To avoid unnecessary examinations, investigators are developing new screening algorithms with multiple clinical parameters (eg, postnatal weight gain and hydrocephalus status). Examples of these screening algorithms include WINROP, PINT-ROP, CHOP ROP, ROPScore, CO-ROP, OMA-ROP, G-ROP, STEP-ROP, and DIGIROP.23 24 Various studies are underway to validate these new algorithms. The G-ROP criteria appear promising; they demonstrated greater sensitivity and specificity than the US protocol for US infants.25 Although there is emerging evidence that up to 50% of eye examinations may be avoidable, it remains challenging to utilise the new screening algorithms in Hong Kong; postnatal weight gain is required to calculate these scores, and such data may not be readily available in our region. Until these new screening algorithms are satisfactorily validated, they are unlikely to replace conventional screening criteria. However, now may be the best time for neonatologists and ophthalmologists in Hong Kong to begin preparing for the new era of ROP by updating classification, screening, and treatment protocols.

All authors contributed to the editorial, approved the final version for publication, and take responsibility for its accuracy and integrity.

The authors have declared no conflicts of interest.

1. Terry TL. Fibroblastic overgrowth of persistent tunica vasculosa lentis in infants born prematurely: II. Report of cases—clinical aspects. Trans Am Ophthalmol Soc 1942;40:262-84.Crossref

2. Shah PK, Prabhu V, Karandikar SS, Ranjan R, Narendran V, Kalpana N. Retinopathy of prematurity: past, present and future. World J Clin Pediatr 2016;5:35-46. Crossref

3. Saugstad OD. Oxygen and retinopathy of prematurity. J Perinatol 2006;26 Suppl 1:S46-50. Crossref

4. Palmer EA. Results of U.S. randomized clinical trial of cryotherapy for ROP (CRYO-ROP). Doc Ophthalmol 1990;74:245-51. Crossref

5. Munro M, Maidana DE, Chan RV. Raising the bar in retinopathy of prematurity treatment. Can J Ophthalmol 2021;56:149-50. Crossref

6. Mintz-Hittner HA, Kennedy KA, Chuang AZ; BEAT-ROP Cooperative Group. Efficacy of intravitreal bevacizumab for stage 3+ retinopathy of prematurity. N Engl J Med 2011;364:603-15. Crossref

7. Stahl A, Lepore D, Fielder A, et al. Ranibizumab versus laser therapy for the treatment of very low birthweight infants with retinopathy of prematurity (RAINBOW): an open-label randomised controlled trial. Lancet 2019;394:1551-9. Crossref

8. Marlow N, Stahl A, Lepore D, et al. 2-year outcomes of ranibizumab versus laser therapy for the treatment of very low birthweight infants with retinopathy of prematurity (RAINBOW extension study): prospective follow-up of an open label, randomised controlled trial. Lancet Child Adolesc Health 2021;5:698-707. Crossref

9. Chan JJ, Lam CP, Kwok MK, et al. Risk of recurrence of retinopathy of prematurity after initial intravitreal ranibizumab therapy. Sci Rep 2016;6:27082. Crossref

10. The Royal College of Ophthalmologists. Clinical Guidelines. Treating Retinopathy of Prematurity in the UK. 2022. Available from: https://www.rcophth.ac.uk/wp-content/uploads/2022/03/Treating-Retinopathy-of-Prematurity-in-the-UK-Guideline.pdf. Accessed 3 Jul 2023.

11. Chiang MF, Quinn GE, Fielder AR, et al. International Classification of Retinopathy of Prematurity, Third Edition. Ophthalmology 2021;128:e51-68.

12. Shah PK, Ramya A, Narendran V. Telemedicine for ROP. Asia Pac J Ophthal (Phila) 2018;7:52-5.

13. Wang SK, Callaway NF, Wallenstein MB, Henderson MT, Leng T, Moshfeghi DM. SUNDROP: six years of screening for retinopathy of prematurity with telemedicine. Can J Ophthalmol 2015;50:101-6. Crossref

14. Gensure RH, Chiang MF, Campbell JP. Artificial intelligence for retinopathy of prematurity. Curr Opin Ophthalmol 2020;31:312-7. Crossref

15. Lu C, Hanif A, Singh P, et al. Federated learning for multicenter collaboration in ophthalmology: improving classification performance in retinopathy of prematurity. Ophthalmol Retina 2022;6:657-63. Crossref

16. Yau GS, Lee JW, Tam VT, et al. Incidence and risk factors of retinopathy of prematurity from 2 neonatal intensive care units in a Hong Kong Chinese population. Asia Pac J Ophthalmol (Phila) 2016;5:185-91. Crossref

17. Iu LP, Lai CH, Fan MC, Wong IY, Lai JS. Screening for retinopathy of prematurity and treatment outcome in a tertiary hospital in Hong Kong. Hong Kong Med J 2017;23:41-7. Crossref

18. Chow PP, Yip WW, Ho M, Lok JY, Lau HH, Young AL. Trends in the incidence of retinopathy of prematurity over a 10-year period. Int Ophthalmol 2019;39:903-9. Crossref

19. Lok JY, Yip WW, Luk AS, Chin JK, Lau HH, Young AL. Visual outcome and refractive status in first 3 years of age in preterm infants suffered from laser-treated type 1 retinopathy of prematurity (ROP): a 6-year retrospective review in a tertiary centre in Hong Kong. Int Ophthalmol 2018;38:163-9. Crossref

20. Royal College of Paediatrics and Child Health. UK Screening of Retinopathy of Prematurity Guideline. 2022. Available from: https://www.rcpch.ac.uk/sites/default/files/2022-12/FC61116_Retinopathy_Guidelines_14.12.22.pdf. Accessed 3 Jul 2023.

21. Fierson WM; American Academy of Pediatrics Section on Ophthalmology; American Academy of Ophthalmology; American Association for Pediatric Ophthalmology and Strabismus; American Association of Certified Orthoptists. Screening examination of premature infants for retinopathy of prematurity. Pediatrics 2018;142:e20183061. Crossref

22. Iu LP, Yip WW, Lok JY, et al. Comparison of United Kingdom and United States screening criteria for detecting retinopathy of prematurity in Hong Kong. Hong Kong Med J 2023;29:330-6. Crossref

23. Iu LP, Yip WW, Lok JY, et al. Prediction model to predict type 1 retinopathy of prematurity using gestational age and birth weight (PW-ROP). Br J Ophthalmol 2023;107:1007-11. Crossref

24. Lundgren P, Stoltz Sjöström E, Domellöf M, et al. WINROP identifies severe retinopathy of prematurity at an early stage in a nation-based cohort of extremely preterm infants. PLoS One 2013;8:e73256. Crossref

25. Binenbaum G, Tomlinson LA, de Alba Campomanes AG, et al. Validation of the postnatal growth and retinopathy of prematurity screening criteria. JAMA Ophthalmol 2020;138:31-7. Crossref

Countries around the world are experiencing widespread challenges in health workforce expansion to manage the health implications of dramatic changes in demographic, socio-economic, epidemiological, climatic, and technological factors. These changes require health providers to demonstrate increasing flexibility and creativity, along with a more proactive approach in terms of addressing the interactions among diverse factors associated with health and healthcare in an ever-changing environment.1 However, traditional didactic methodologies have been widely utilised to emphasise the central role of the teacher in knowledge transfer and learning practices mainly via planned lectures with large amounts of theoretical content in a fixed environment. This approach offers limited opportunities for students to practise and share their knowledge, hindering the development of adaptability to meet the growing demand for lifelong learning.

In this issue of the Hong Kong Medical Journal, Ng et al2 evaluated the effectiveness of online micromodule teaching in knowledge transfer within the urology subspecialty among medical students without prior exposure to urology practice. The ‘flipped classroom’ demonstrated similar efficacy in knowledge transfer, as measured by pre-intervention and post-intervention multiple-choice questions and objective structured clinical examinations, compared with the traditional didactic lecture model.2 The findings suggest that the adoption of micromodules as a ‘flipped classroom’ component can maximise time for practical training and experience sharing between clinicians and medical students. This approach incorporating the use of digital media echoes previous research which highlighted the need for urology training innovation because of the impact of the COVID-19 (coronavirus disease 2019) pandemic.3 Despite the non-inferiority trial design used in their study, the efforts by Ng et al2 to create a culture of autonomy and establish a self-paced learning environment have demonstrated the substantial potential of online digital learning for improving student engagement and sustaining knowledge development.

In healthcare and clinical practice, knowledge acquisition is particularly important for both health professionals and the general population. Recent studies have identified widening gaps in health knowledge, awareness, and practice in the fields of hepatology and nephrology.4 5 Additionally, diverse educational resources for health advocacy and self-learning have become available because of the growing popularity of electronic material, combined with increasing access to digital technologies and social media platforms.5 The expansion of internet-based channels has led to broader education outreach, as reflected in a large-scale survey among >3000 respondents who reported regular access to digital platforms for rapid communication of health-related information.6

The World Health Organization has identified five key domains for interventions to transform and enhance the education available to health professionals: education and training institutions, accreditation and regulation, financing, monitoring and evaluation, and governance.7 Online learning (e-learning) has been highlighted as an innovative teaching and learning strategy that can support the establishment of institutions with sufficient strength to produce the desired quantity and quality of health professionals in both high-income and resource-limited settings.7 The ‘flipped classroom’ model is gaining popularity as an innovative teaching technique. In contrast to the primarily passive listening approach involved in traditional direct-instruction classroom lectures, students in ‘flipped classrooms’ receive digital learning material (eg, pre-recorded video lectures, podcasts, narrated presentations, and other internet-based material) prior to the traditional in-class session. This approach permits ‘in-class’ time to be used for knowledge consolidation and application through student-centred learning activities such as group discussion, peer projects, problem-solving exercises, and individualised assessments of student understanding. A literature review identified the many opportunities presented by digital technologies, which include (but are not limited to) more effective use of traditional ‘class’ time, greater diversity of learning materials, and additional opportunities to revitalise the learning process.8 For example, the integration of virtual patients and clinical simulation scenarios offers students unique learning opportunities to consolidate practical skills via digitally enhanced clinical practice.8 9 This approach may be ideal for the reformation of medical school curricula to address social stigma associated with various diseases (eg, human immunodeficiency virus/acquired immunodeficiency syndrome,10 mental illness, and cancer11) by incorporating interventions that involve experiential and affective teaching components.

The inclusion of digital innovations in education and pedagogy reform enhances clinical competencies among students, while creating environments for resilience building. There is a need to manage physicians’ increasing clinical responsibilities that arise from rapid progress in health systems, in addition to the growing demand for medical and translational research conducted in clinical settings.12 Longer working hours can lead to significant work-life imbalance and greater risks of dissatisfaction and depression among physicians. A territory-wide survey revealed a high prevalence of burnout among training and practising physicians in Hong Kong.13 In the United States, the implementation of an innovative ‘flipped classroom’ mindfulness training programme significantly reduced physician burnout, emotional exhaustion, and depersonalisation among both residents and faculty.14 The design of online modules focused on mind-body skills training, combined with interactive discussion sessions, has demonstrated efficacy in terms of increasing resilience, thereby improving the provision of calm and compassionate care.14

Among the various innovative components of the ‘flipped classroom’ model, multimedia tools with digital elements contribute to greater improvements in visualisation and student engagement throughout the teaching and learning process.15 A combination of interactive text, graphics, sound, animation, and video delivered by electronic means may be appropriate for children with intellectual disabilities—such children have an increased risk of infection because their limited cognitive ability hinders absorption and retention of health knowledge.16 A study conducted in Hong Kong showed that the development of multimedia visualisation teaching strategies with visual prompts (eg, lyrics and posters) helped the target population to learn proper hand-washing procedures.16

From the perspective of health communication, health awareness among patients (who are enhanced through effective physician-patient education) and clinical skills among physicians have equal importance in terms of ensuring excellent care.17 With respect to eye health, a sustained school-family partnership is critical for achieving the desired goal of ‘Vision for Everyone’.18 Advances in digital communication to share, disseminate, and amplify health messages—to target audiences and the wider community—have key roles in promoting universal eye health and preventing avoidable blindness. Digital technologies are also expected to play major roles in out-of-class settings where the communication of health knowledge between school teachers and students’ parents via digital routes (eg, instant messengers) may have long-term effects on students’ abilities to learn and maintain healthy behaviours.19

In recent decades, dramatic advances in digital technologies (eg, mobile computing, artificial intelligence, blockchain, virtual reality, and augmented reality) have facilitated widespread exploration of digital innovations in clinical practice and public health.20 Digitally enhanced learning has become a key driver of health system changes that can empower patients, physicians, and students. Therefore, the expansion of digitally enhanced learning practices should be encouraged and supported, both within and across medical specialities, to generate evidence that can guide education and pedagogy reform in response to the changing environment and health profiles in the post–COVID-19 era.

All authors contributed to the editorial, approved the final version for publication, and take responsibility for its accuracy and integrity.

The authors have declared no conflict of interest.

1. Gill D, Whitehead C, Wondimagegn D. Challenges to medical education at a time of physical distancing. Lancet 2020;396:77-9. Crossref

2. Ng CF, Lim K, Yee CH, Chiu PK, Teoh JY, Lai FP. Time for change? Feasibility of introducing micromodules into medical student education: a randomised controlled trial. Hong Kong Med J 2023;29:208-13. Crossref

3. Yee CH, Wong HF, Tam MH, et al. Effect of SARS and COVID-19 outbreaks on urology practice and training. Hong Kong Med J 2021;27:258-65. Crossref

4. Chan HL, Wong GL, Wong VW, Wong MC, Chan CY, Singh S. Questionnaire survey on knowledge, attitudes, and behaviour towards viral hepatitis among the Hong Kong public. Hong Kong Med J 2022;28:45-53. Crossref

5. Langham RG, Kalantar-Zadeh K, Bonner A, et al. Kidney health for all: bridging the gap in kidney health education and literacy. Am J Nephrol 2022;53:87-95. Crossref

6. Tam VC, Tam SY, Khaw ML, Law HK, Chan CP, Lee SW. Behavioural insights and attitudes on community masking during the initial spread of COVID-19 in Hong Kong. Hong Kong Med J 2021;27:106-12. Crossref

7. World Health Organization. Transforming and Scaling Up Health Professionals’ Education and Training: World Health Organization Guidelines 2013. Geneva: World Health Organization; 2013.

8. Forde C, OBrien A. A literature review of barriers and opportunities presented by digitally enhanced practical skill teaching and learning in health science education. Med Educ Online 2022;27:2068210. Crossref

9. Kononowicz AA, Woodham LA, Edelbring S, et al. Virtual patient simulations in health professions education: systematic review and meta-analysis by the Digital Health Education Collaboration. J Med Internet Res 2019;21:e14676. Crossref

10. Tam G, Wong NS, Lee SS. Serial surveys of Hong Kong medical students regarding attitudes towards HIV/AIDS from 2007 to 2017. Hong Kong Med J 2022;28:223-9. Crossref

11. Cheung YT, Yang LS, Ma JC, et al. Health behaviour practices and expectations for a local cancer survivorship programme: a cross-sectional study of survivors of childhood cancer in Hong Kong. Hong Kong Med J 2022;28:33-44. Crossref

12. Wang HH, Chen L, Ding H, Huang J, Wong MC. Scientific research on COVID-19 conducted in Hong Kong in 2020. Hong Kong Med J 2021;27:244-6. Crossref

13. Kwan KY, Chan LW, Cheng PW, Leung GK, Lau CS. Burnout and well-being in young doctors in Hong Kong: a territory-wide cross-sectional survey. Hong Kong Med J 2021;27:330-7. Crossref

14. Moffatt-Bruce SD, Nguyen MC, Steinberg B, Holliday S, Klatt M. Interventions to reduce burnout and improve resilience: impact on a health system’s outcomes. Clin Obstet Gynecol 2019;62:432-3. Crossref

15. Abdulrahaman MD, Faruk N, Oloyede AA, et al. Multimedia tools in the teaching and learning processes: a systematic review. Heliyon 2020;6:e05312. Crossref

16. Lee RL, Leung C, Chen H, Tong WK, Lee PH. Five-step hand hygiene programme for students with mild intellectual disability: abridged secondary publication. Hong Kong Med J 2022;28 Suppl 3:41-2.

17. Wang HH, Li YT, Wong MC. Leveraging the power of health communication: messaging matters not only in clinical practice but also in public health. Hong Kong Med J 2022;28:103-5. Crossref

18. Du K, Huang J, Guan H, Zhao J, Zhang Y, Shi Y. Teacher-to-parent communication and vision care-seeking behaviour among primary school students. Hong Kong Med J 2022;28:152-60. Crossref

19. Burton MJ, Ramke J, Marques AP, et al. The Lancet Global Health Commission on Global Eye Health: vision beyond 2020. Lancet Glob Health 2021;9:e489-551. Crossref

20. Budd J, Miller BS, Manning EM, et al. Digital technologies in the public-health response to COVID-19. Nat Med 2020;26:1183-92. Crossref