Ageing, characterised by a progressive loss of physiological integrity leading to impaired function and increased vulnerability to death, is a major health concern for global populations.1 The older population (threshold defined by the World Health Organization as aged ≥60 years) in Hong Kong, one of the most rapidly ageing communities worldwide, is expected to increase from 2.28 million (30.5% of the total population) in 2023 to 3.31 million (40.4%) in 2046, with the steepest growth occurring in advanced old age.2 Major neurocognitive disorders (NCD) [also referred to as dementia] are the most common neurodegenerative disorders associated with ageing and exert a substantial impact on healthcare systems. A global projection estimated that the number of people living with dementia would increase almost threefold, from 57.4 million cases in 2019 to 152.8 million in 2050.3 In the most recent epidemiological study of dementia in Hong Kong conducted in 2008, over 8.9% of community-dwelling older adults had mild dementia, and 8.5% had mild cognitive impairment (MCI), a synonym for mild NCD.4 Given that these estimates are over a decade old, demographic changes (eg, higher educational attainment and evolving health conditions among older adults) warrant re-evaluation of prevalences and associated factors.5

Factors affecting cognitive decline are best understood through a life-course perspective. While genetic and early-life predisposing factors are not readily modifiable, mid- and late-life health and lifestyle factors are increasingly recognised as modulators of cognitive impairment. A 2024 Lancet review6 suggested that 14 potentially modifiable lifestyle and health factors, including cardiovascular risk, hearing and vision loss, air pollution, and mental health and lifestyle factors (eg, smoking, alcohol consumption, physical inactivity, social isolation, and depression), accounted for 45% of the population-attributable risk of dementia. Growing research interest has also focused on other potential risk factors, such as sleep, diet, dental disease, and frailty, which are also important determinants of cognitive health.6

Investigating secular trends in dementia is essential to understand the full spectrum of the condition in general populations and to identify risk factors across populations and life stages.7 Recent estimates from various populations have reported lower-than-expected prevalence rates, possibly due to improvements in education, environmental enrichment, and healthcare, with resulting reductions in cerebrovascular risk.6 8 In the local context, an updated prevalence study facilitates systematic evaluation of the evolving occurrence and modulating factors of cognitive decline from physical health to psychosocial perspectives.6 Such evaluation is essential for developing context-, culture-, and practice-tailored preventive strategies targeting the identified risk factors, as well as for optimising treatment and management.6 7 Particularly with advancing age and rising rates of physical and mental co-morbidities among older adults in Hong Kong, the burden of care and service demands related to cognitive impairment require more comprehensive assessment and practical guidance.

This study aimed to estimate the current prevalence of mild and major NCD in Hong Kong and to identify their multidimensional associated factors, based on the Hong Kong Mental Morbidity Survey for Older People (HKMMSOP). We also discuss how these findings may inform healthcare interventions for early risk modification, reduction of cognitive decline, and optimisation of care.

The HKMMSOP was a commissioned study funded by the Advisory Committee on Mental Health through the Health and Medical Research Fund. Conducted from January 2019 to January 2023, it was designed as a territory-wide, population-based, cross-sectional survey to examine the prevalence and modulating factors of NCD among older adults in Hong Kong. The study settings of HKMMSOP included both community households and long-term care (LTC) institutions. For clinical assessments, HKMMSOP adopted a two-phase design. Phase 1 interviews comprised cognitive assessments for mild and major NCD, evaluation of neuropsychiatric syndromes and functioning, as well as physical health and psychosocial measurements. Phase 2 involved clinician interviews for diagnostic assessment and subtyping of NCD.

Sample size was determined based on previous prevalence studies of dementia (2008)4 and common mental disorders (2010-2013)9 in Hong Kong. For an estimated NCD prevalence of approximately 2% among adults aged 60 to 74 years, 3012 participants were required to achieve a recommended precision of 0.005. For an estimated prevalence of approximately 5% among those aged 75 years or above, 1168 participants were required to achieve a recommended precision of 0.0125. The HKMMSOP ultimately recruited and completed assessments for 3560 participants aged 60 to 74 years and 1311 participants aged 75 years or above.

To recruit a representative sample of the older adult population in Hong Kong, we adopted a multi-stage random sampling method commonly used in household surveys. A random list of addresses (sampling frame) was generated by the Census and Statistics Department of the Hong Kong SAR Government, then stratified by geographical location and residential type (private versus public housing). For each address, an invitation letter introducing the survey and a consent form for assessment were enclosed. A telephone hotline, designated website, and email contact were provided for enquiries and to document refusals. Up to three invitation letters were sent within 6 months to improve recruitment success and reduce response bias; participants who responded to the third contact had higher rates of active employment compared with the rest of the sample (P<0.05).

When households responded and included residents aged 60 years or above, we invited them for interviews; there could be one or more eligible residents per household. If individuals agreed to participate but were unable to provide complete information (eg, due to profound sensory deficits), data were obtained from their first-degree relatives and categorised as proxy. Households with no eligible participants (ie, all residents aged under 60 years) were invited to notify the research team of their ineligibility through convenient contact channels (email, text messages, or direct phone calls). From January 2019 to January 2023, a total of 39 772 invitation letters were sent to randomly generated addresses. Of these, 3352 households with 4369 eligible community-dwelling participants consented to and completed the survey. Eligibility and demographic characteristics were unknown for the remaining 36 420 households. The flow of participant sampling, recruitment, and assessment is depicted in the Figure.

To obtain representative statistics for people residing in LTC institutions, we adopted a two-stage cluster sampling method. Superintendents of 600 residential care homes for the elderly (ie, LTCs), randomly selected from the master list of registered old age homes in Hong Kong, were first approached for participation. When residential care homes agreed to assist with recruitment, eligible residents were randomly selected. Following institutional consent and consent from participants and/or their family members, assessments were completed for 503 residents from 51 registered old age homes across Hong Kong. The prison population was not included in this survey.

Phase 1 assessments were conducted by trained research assistants during visits to participants’ residential addresses or at the department’s research centre. Due to social distancing and infection control policies during the COVID-19 pandemic, phone assessments were offered as an alternative and were utilised by 28.7% of community participants. Proxy-based assessments were conducted for participants with profound physical or mental deficits (2.7% of community participants and 34.2% of LTC participants).

Socio-demographic information was collected, including age, gender, birthplace, housing type, education level, marital status, household composition, current employment status, family income, receipt of government financial subsidies, and religious affiliation.

Cognitive function was assessed using two locally validated tools: the Chinese Abbreviated MCI test10 and the Hong Kong version of Montreal Cognitive Assessment (HK-MoCA).11 12 For participants with moderate to severe major NCD in both domiciliary and institutional settings, an abridged version of the cognitive and mental state assessment was used based on the HK-MoCA 5-minute protocol. All interviewers were trained to administer the Clinical Dementia Rating (CDR) to each participant, and satisfactory concordance was achieved between interviewer-rated and clinician-rated results (Spearman’s correlation: 0.668; P<0.001).13 Neuropsychiatric symptoms were screened in all participants using the Neuropsychiatric Inventory Questionnaire.14

Physical health was assessed using questionnaires on chronic disease burden (Cumulative Illness Rating Scale [CIRS]),15 along with health screening measures including blood pressure, body mass index, oral health, grip strength, hearing and vision, postural balance, and sarcopenia (SARC-F: Strength, Assistance with walking, Rising from a chair, Climbing stairs, and Falls).16 Activities of daily living were evaluated using the Chinese version of the Disability Assessment for Dementia.17

Lifestyle factors were assessed using questionnaires covering smoking and alcohol consumption, physical and non-physical leisure activities, sleep quality (Pittsburgh Sleep Quality Index), use of drugs and vitamins, dietary intake of fruits and vegetables, loneliness, and quality of life.18 19 20 21

Among the 1020 participants whose scores crossed the threshold for mild or major NCD, clinician interviews were conducted for NCD subtyping, including assessment of cerebrovascular risk and neuroimaging. Of these, 457 participants (response rate: 44.8%) completed face-to-face structured assessments at the department’s research centre, where blood sampling facilities were available (Figure). For consenting participants, fasting lipid profiles, glycated haemoglobin, and apolipoprotein E4 genotypes were analysed. Diagnoses and subtypes of NCD were established according to the DSM-5 (Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition)22 through clinician assessment supplemented by laboratory results and structural magnetic resonance imaging of the brain.

Additionally, 201 participants with negative NCD screening results (CDR=0) were interviewed by research team psychiatrists to evaluate the sensitivity and specificity of Phase 1 assessments. The sensitivity of the Phase 1 screening tools for detecting mild or major NCD was 96.7%, while the specificity for correctly identifying participants without NCD was 81%. The positive predictive value was 90% and the negative predictive value was 93%.

Due to strict COVID-19 visitor policies in place during the fieldwork period (2022-2023), videoconference assessments were conducted with participants in some LTC facilities (62.6%). Individual health assessments were not performed because of infection control policies in effect during data collection in 2022.

In the two-phase prevalence study design, selection bias could potentially arise if heterogeneity in characteristics (eg, age, education level, family structure, physical and mental morbidity) existed between responders and non-responders. During the study invitation and Phase 1 assessment, demographic and clinical information for non-responders and non-completers was unavailable. Therefore, we compared Phase 2 completers (n=457) and non-completers (n=563) among participants with positive NCD screening results in Phase 1. Non-completers were more likely to be older (>75 years: 49.9% vs 36.1%; P<0.001), have a higher burden of chronic conditions (CIRS score: 4.8 ± 3.0 vs 4.4 ± 2.4; P=0.007), show greater levels of cognitive impairment (CDR ≥1: 20.5% vs 12.2%; P<0.001), exhibit more impaired daily functioning (score of the Chinese version of the Disability Assessment for Dementia: 90.6 ± 20.5 vs 96.0 ± 9.6; P<0.001), and have a prior diagnosis of dementia before joining the study (5.9% vs 1.6%; P=0.001). The potential impact of participant imbalance in Phase 2 on study findings is addressed in the Discussion section. No imputation was performed for missing data.

Prevalence estimates of NCD were stratified by age-group, gender and recruitment setting. Weighted prevalence was calculated using sampling weights that reflected the proportion of participants in each stratum, with reference to the population distribution by age and gender, as well as community and LTC statistics (year-end population, 2022).23 24 Diagnoses included mild NCD and major NCD. For each disorder, prevalence was expressed as the rate per 100 persons, with corresponding 95% confidence interval (95% CI). Sampling errors were estimated using a bootstrap strategy implemented in Stata software (StataCorp, College Station [TX], US). One thousand samples were randomly drawn from the original dataset through sampling with replacement. The standard error of weighted prevalence was calculated as the square root of the sample variance across these resamples.

The psychosocial, physical health, and lifestyle correlates of NCD were evaluated. Crude and adjusted associations (odds ratios) between the prevalence of NCD and three categories of associated factors were examined: (1) demographic characteristics, including age, gender, years of education, marital status (married, cohabiting, widowed, divorced, separated or single), monthly household income (<HK$6000, HK$6000-14 999, HK$15 000-29 999, ≥HK$30 000), and housing type (owned or rented); (2) physical health status, including history of hypertension, postural balance test (pass or fail), corrected vision (normal or abnormal), corrected hearing (normal or abnormal), oral health problems (yes or no), sarcopenia (SARC-F score), and chronic disease burden (CIRS score); and (3) lifestyle habits, including smoking and alcohol consumption (never, former, current), fruit intake (≥two portions/day or not), vegetable intake (≥three portions/day or not),18 25 sleep quality (score of the Pittsburgh Sleep Quality Index21), regular participation (once or more per week) in aerobic, resistance, and intellectual activities,20 and social loneliness.19 Multivariable logistic regression analyses were conducted to identify independent associations between physical health and lifestyle factors and NCD, adjusting for confounders previously linked to NCD, including age, gender, years of education, marital status, and socio-economic status.5 6 8 All analyses were performed using SPSS (Windows version 22.0; IBM Corp, Armonk [NY], US), and differences were considered statistically significant at P<0.05.

Basic characteristics of participants living in the community and in LTC settings are presented in Table 1. The mean age was 69.6 ± 7.5 years (range, 60-105). Just over half of the respondents (53.6%) were born in Hong Kong. A large proportion had attained a secondary school education or above (66.1%) and were married or co-habiting (67.5%). Most older adults were economically inactive (retired, homemaker, or never worked) [78.1%].

Of the 503 participants residing in LTC facilities, 274 (54.5%) were women and the mean age was 80.3 ± 11.0 years (range, 60-106). More than half (54.6%) were born in Chinese Mainland; 30.4% had attained secondary education or above; and 50.4% were widowed, divorced or separated. Compared with the LTC sample, the community sample included a higher proportion of women (56.3% vs 54.5%) and younger-old adults (aged <75 years: 77.7% vs 32.6%), whereas the LTC sample included more men (45.5% vs 43.7%) and adults aged ≥75 years (67.4% vs 22.2%). After sample weighting by age and gender, both samples were comparable with the overall older population in Hong Kong (Table 1).

The distribution of CDR before sample weighting is presented by 5-year age intervals and gender in both community and LTC settings (Table 2). Older age was associated with a higher prevalence and greater severity of NCD, as measured by CDR (P<0.001). Montreal Cognitive Assessment scores stratified by gender and grouped into 5-year age intervals are also presented for community-dwelling participants (Table 3). Older age was associated with lower HK-MoCA scores (P<0.001), while men had higher HK-MoCA scores within 5-year age-groups >70 years (P<0.05).

The proportion of neuropsychiatric symptoms increased with worsening cognitive impairment. Among community-dwelling participants, the prevalence of psychotic symptoms rose from 2.1% in those with normal cognition to 6.3% in those with mild NCD, and 21.3% in those with major NCD (P<0.001). Similarly, the prevalence of depression and anxiety also increased with higher CDR scores (P<0.001). Sleep disturbances were common across all cognitive groups, affecting 32.2% of participants with normal cognition, 43.5% with mild NCD, and 40.2% with major NCD.

The unweighted prevalence of mild and major NCD is presented in online supplementary Table. As shown in Table 4, the weighted prevalence of mild NCD was 21.8% among community-dwelling older adults, 24.3% among those living in LTC facilities, and 21.8% overall in Hong Kong. For major NCD, prevalence was 7.4% in the community, significantly higher at 68.8% in LTC settings, and 9.7% overall.

Among Phase 2 community-dwelling participants with mild or major NCD, 22.0% met the DSM-5 criteria for Alzheimer’s disease, 23.7% had mixed vascular NCD and AD, and 43.5% had vascular NCD. Neurocognitive disorders due to Lewy body disease and frontotemporal lobar degeneration accounted for 1.2% and 0.5% of cases, respectively. The apolipoprotein E4 genotype was identified in 17% of cognitively normal participants, 19% of those with mild NCD, and 22% of those with major NCD. No significant differences in apolipoprotein E4 distribution were observed across cognitive function groups.

In unadjusted analyses, older age, fewer years of education, being widowed, divorced or separated, living in rented housing, and lower household income were significantly associated with a higher prevalence of mild NCD. These factors were subsequently controlled for in the logistic regression analyses.

After controlling for the above demographic confounders, hypertension, diabetes mellitus, history of stroke, poor postural balance, higher SARC-F scores, visual or hearing impairment, and dental problems were associated with significantly higher adjusted ORs for mild NCD (P<0.05). Physical exercise (mind-body and resistance), engagement in intellectual activities, and better subjective sleep quality were associated with lower adjusted ORs for mild NCD (P<0.05). In contrast, individuals with mild NCD had significantly higher loneliness scores (P<0.05) [Table 5].

Similar demographic risk factors, such as older age, female gender, fewer years of education, being widowed, divorced, separated or never married, living in rented housing, and lower household income level were associated with a higher prevalence of major NCD. These factors were controlled for in the logistic regression analyses (Table 5).

After controlling for demographic confounders, hypertension, diabetes mellitus, history of stroke, poor postural balance, abnormal vision or hearing, edentulism, high SARC-F scores (≥2), and multiple co-morbidities (≥4 chronic diseases) were associated with higher adjusted ORs for major NCD (P<0.05) [Table 5].

Less loneliness, participation in mind-body physical exercise, engagement in intellectual activities, and consumption of three portions of vegetables or more per day were associated with a lower likelihood of major NCD after adjustment for demographic confounders (P<0.05). Poor sleep quality was associated with a higher risk of major NCD in unadjusted analyses (Table 5).

A family history of dementia was reported by 27% of cognitively normal participants, 25% of those with mild NCD, and 23% of those with major NCD. Pearson Chi squared test showed no significant differences across groups.

Self-reported service utilisation among community-dwelling participants with NCD was assessed in 488 individuals who completed the Phase 2 assessment. Participants with mild and major NCD reported higher use of inpatient, accident and emergency, and outpatient services compared with those with normal cognition. Notably, participants with major NCD reported significantly more psychiatric (15.7%) and neurological (23.0%) consultations in the preceding 3 months than those with normal cognition (2.5% and 2.7%, respectively) and those with mild NCD (6.3% and 4.2%, respectively) [P<0.001].

As population ageing accelerates, the Asia-Pacific region is projected to experience more than a threefold increase in the number of people living with dementia over the next three decades, rising from 23 million in 2015 to 71 million by 2050.26 This territory-wide, population-based study provides up-to-date prevalence estimates of NCD among adults aged ≥60 years in Hong Kong in 2022. The prevalence of mild NCD was 21.8% (21.8% in the community; 24.3% in LTC settings), while the prevalence of major NCD was 9.7% (7.4% in the community; 68.8% in LTC settings).26

A nationwide population-based study conducted in Chinese Mainland in 2020 reported overall age- and gender-adjusted prevalences of 15.5% (95% CI=15.2-15.9) for MCI and 6.0% (95% CI=5.8-6.3) for dementia.5 The relatively higher prevalence observed in Hong Kong may be related to differences in population demographics and healthcare systems. First, the present study included older adults residing in residential care homes in both recruitment and prevalence estimates. The higher proportion of older adults living in LTC facilities in Hong Kong (3.7% of the older population)27 compared with Chinese Mainland (<1%),28 combined with the high prevalence of major NCD among LTC residents (68.8%), contributed to an increased overall prevalence of major NCD in Hong Kong. Second, even after excluding LTC residents, the prevalence of NCD among community-dwelling older adults in Hong Kong remained higher. This difference may be attributed to longer life expectancy (83.7 vs 78.6 years) and an age structure characterised by a greater proportion of the oldest-old (population aged ≥80 years: 5.3% vs 2.3%) in Hong Kong29 30 compared with the Chinese Mainland. Nevertheless, comparison of age-specific prevalence rates revealed a lower prevalence of major NCD in younger-old groups in Hong Kong29 30 (age 60-69 years: 1.0% vs 2.9%; age 70-79 years: 4.7% vs 8.4%5), but a higher prevalence among the oldest-old (age ≥80 years: 33.7% vs 16.1%5). In contrast, the prevalence of mild NCD was consistently higher across all age-groups in Hong Kong (age 60-69 years: 14.5% vs 11.8%; age 70-79 years: 28.8% vs 19.2%; age ≥80 years: 33.1% vs 25.0%).5 These patterns may reflect better management of cardiovascular diseases and other risk factors (eg, lower rates of smoking and alcohol consumption) among the younger-old population, the approximately twofold higher proportion of the oldest-old among older adults (age ≥85 years: 10.7% vs 4.9%5), and longer survival following dementia onset in Hong Kong.

Compared with other developed Asian economies, including Japan, South Korea, Taiwan, and Singapore, the prevalence of dementia in Hong Kong is also relatively high, particularly among those aged 80 years or above (online supplementary Fig).31 Although the high proportion of oldest-old individuals partly contributes to this observation (Hong Kong: 10.7%, Japan: 15.2%, Taiwan: 7.7%, Singapore: 6.6%, South Korea: 6.5%), another important explanation may be the substantial burden of cerebrovascular disease, which is associated with a higher prevalence of dementia (online supplementary Fig).31 In Hong Kong, cerebrovascular risk factors strongly contribute to dementia cases (eg, 43.5% vs 26.7% in Chinese Mainland5), whereas among the oldest-old population, cerebral small vessel disease is highly prevalent.

In comparison with the 2008 community-based prevalence study in Hong Kong,4 a decrease in the overall prevalence of mild dementia (mild stage of major NCD) was observed (from 5.4% in 2008 to 3.8% in 2022 [Table 4]), with reductions noted across all age-groups. This decline may be attributed to improved management of physical health risk factors, enabling more older adults to remain within the mild NCD range. Contributing factors may include higher educational attainment, reduced smoking and alcohol consumption, and better control of cerebrovascular disease.

Regarding mild NCD (MCI or CDR=0.5), the prevalences of very mild dementia (5.8%) and MCI (23.8%) reported in 20084 are not directly comparable to the prevalence of mild NCD observed in HKMMSOP (21.8%). First, different cognitive screening tools were employed. The 2008 study utilised the CMMSE and AMT,4 whereas HKMMSOP adopted the HK-MoCA as the primary screening instrument. The HK-MoCA was designed to be more sensitive in detecting early executive dysfunction associated with vascular or non-Alzheimer’s pathology.12 In practice, the HK-MoCA12 demonstrates comparable sensitivity (97%) but higher specificity (81% vs 72%) and negative predictive value (93% vs 81%) compared with the CMMSE.4 Second, the diagnostic threshold for mild NCD, labelled as very mild dementia (CDR=0.5) in the 2008 study,4 has shifted over the past decade in favour of an early-detection paradigm. Advances in screening accuracy and diagnostic algorithms may have contributed to a higher detection rate of mild NCD in HKMMSOP. Finally, due to methodological limitations in the 2008 study,4 comparisons across the full spectrum of cognitive impairment are restricted. The CDR assessments were only conducted among participants who screened positive for cognitive impairment and proceeded to Phase 2 clinician evaluation. In contrast, CDR scores in HKMMSOP were determined for all participants during Phase 1 by trained research assistants and subsequently corroborated by experienced psychiatrists.

Finally, optimisation of dementia risk management, along with changes in the population age structure, may also help explain the increased concentration of moderate and severe cases among the oldest-old participants in HKMMSOP.

This study identified several risk factors associated with NCD, many of which were common to both mild and major NCD. These included increasing age, lower educational attainment, being widowed, divorced or separated, and poorer socio-economic status as indicated by living in rented housing and reporting a lower household income. These findings are consistent with previous studies.5 8 32 Female gender was associated with major NCD, but not with mild NCD, among older adults in Hong Kong.

Regarding physical health conditions, in addition to cardiovascular disease, poorer postural balance, higher sarcopenia scores, visual and hearing impairment, and oral health problems were associated with the presence of NCD. A greater number of co-morbidities was associated with major NCD, while poor sleep quality was associated with an increased risk of mild NCD. With respect to potentially modifiable lifestyle factors, consuming three portions of vegetables or more per day was associated with a lower prevalence of major NCD. Regular physical exercise, engagement in intellectual activities, and lower levels of loneliness were associated with a reduced prevalence of NCD. The cross-sectional associations observed between potentially modifiable factors and NCD in this study may enrich the existing evidence base and provide converging directions for future research into causal relationships. These findings may also inform policy development aimed at the dementia prevention worldwide.

The findings of HKMMSOP should be interpreted in light of several limitations. First, this was a cross-sectional study; therefore, causal relationships between NCD and associated factors cannot be inferred. Second, sampling bias is an inherent limitation in prevalence studies, as individuals with an existing diagnosis and ongoing treatment may be less likely to participate (due to reduced activity levels resulting in underestimation) or more motivated to enrol, potentially leading to overestimation). Hard-to-reach populations may also have been underrepresented due to factors such as poor physical or cognitive function, limited mobility, or the absence of family caregivers. The COVID‑19 pandemic likely exacerbated sample bias in this study: the household response rate was merely 8.4%, and the sample over‑represented women and younger‑old adults, potentially underestimating NCD severity among the oldest‑old.

Third, participants who did not attend Phase 2 assessments were older and had a greater burden of physical morbidity. The main reasons for non-participation among individuals with positive screening results in Phase 1 included ‘assessment centre too far from home’, ‘too old or too frail’, ‘no accompanying caregiver’, and ‘no perceived necessity’. While Phase 1 assessments demonstrated satisfactory positive and negative predictive values for NCD diagnosis (90% and 93%, respectively), differences in participation profiles between those who did and did not complete Phase 2 may have influenced prevalence estimates.

Finally, HKMMSOP was conducted during periods when Hong Kong was affected by various phases of the COVID-19 pandemic. Infection control measures adversely impacted participant recruitment (low household response rate: 8.4%) and interview arrangements. Surveys involving residents of care homes and hostels were particularly restricted due to stringent lockdown policies. As a result, only a limited number of in-person, telephone, or online assessments could be conducted with older participants. Most information was obtained from family caregivers or formal carers within the respective institutions, which may have influenced the reliability and validity of the assessment instruments.

Despite these limitations, the findings of this study remain valuable for informing future clinical practice, public health interventions, and research priorities. First, the HKMMSOP revealed that nearly one in five older adults in Hong Kong had mild NCD. This pattern is likely not unique to Hong Kong and may also apply to other Asian metropolitan cities characterised by increasing life expectancy and a high burden of physical co-morbidities among older populations. Mild NCD represents an at-risk state with variable clinical trajectories. In a 5-year prospective study of Chinese older adults with MCI, approximately 30% progressed to dementia, while others either remained stable or improved to normal cognitive function.33 Given the spectrum from normal cognition to mild and major NCD encountered in primary care, prevention and timely intervention should address a broad range of associated health factors, particularly common and modifiable ones operating across the life course of cognitive health, such as educational attainment, socio-economic status, sensorimotor function, physical exercise, and intellectual activity.7 34 Although the cross-sectional design of the HKMMSOP does not permit causal inference, early intervention and management of these modifiable health risks may help reduce progression to major NCD, improve quality of life and functional capacity among individuals living with MCI, and yield meaningful clinical and economic benefits.

Within Hong Kong, the development of District Health Centres provides an opportunity to support primary care providers in planning screening and early intervention programmes for cognitive and mental health. Optimal management of cardiovascular disease and related risk factors from midlife is essential as cerebral small vessel disease contributes to—and may play a causal role in—a substantial proportion of vascular and mixed dementia cases.35 Additionally, sensory function, oral health, and musculoskeletal integrity should be emphasised and integrated into primary healthcare screening. Lifestyle interventions also warrant attention, as clinicians may promote cognitive benefits through various forms of physical exercise and intellectual activities, as well as interventions targeting sleep hygiene and social connectedness.25 36 37 38

Second, considering the high prevalence of major NCD (dementia) among the oldest-old population, an integrated medico-social support system should be established. Among participants with major NCD in the HKMMSOP, utilisation rates of specialist services (psychiatric, neurological, and psychological outpatient care) were far from optimal (<40%). This gap may substantially hinder the timely treatment and management of cognitive or behavioural complications, thereby increasing family, economic, and societal burdens.39 At present, primary care consultations specifically addressing cognitive decline are not well established in Hong Kong. While not all individuals with NCD require specialist medical attention, the observed 60% service gap underscores the importance of strengthening primary care management to optimise cognitive function in the community.39 Given that NCD comprises heterogeneous neurodegenerative conditions, appropriate and tiered medical assessments and interventions across both primary and specialist settings play a critical role in accurate subtype diagnosis, personalised manageme

Third, approximately 70% of residents in LTC facilities were affected by major NCD. Coordinated efforts and the integration of multidisciplinary care are essential to recognise and address the complex cognitive, physical, and mental health needs of individuals living in LTC facilities, as well as those of their caregivers.

Finally, considering the limitations of this cross-sectional prevalence study, further research is warranted in the following two areas: (1) investigation of potentially modifiable health and lifestyle factors for healthy cognitive ageing through longitudinal, cohort, and clinical trial designs to elucidate causal relationships; and (2) focused evaluation of cognitive impairment among the oldest-old population, individuals with complex socio-medical conditions (eg, hard-to-reach groups, those living alone, those with high comorbidity burden, or limited access to health and social resources), and residents of LTC facilities.

The HKMMSOP study provides updated estimates of NCD prevalence among community-dwelling adults aged 60 years or above in Hong Kong during the COVID-19 pandemic period. The weighted prevalence of major NCD in Hong Kong was estimated at 9.7%, with the greatest increase observed among the oldest-old. Given the steep rise in the proportion of the population reaching advanced age, the total number of people living with major NCD is expected to continue increasing over the coming decade. Approximately one in five adults aged 60 years or above had mild NCD. Public health education for older adults should focus on optimising the management of chronic medical and cerebrovascular diseases, promoting regular physical exercise, correcting sensory impairments, and encouraging active engagement in intellectual and social enrichment activities. Equally important, the complex needs of older adults residing in residential care homes should not be overlooked—nearly seven in ten LTC residents were affected by major NCD.

Concept or design: LCW Lam, WC Chan, ATC Lee, AWT Fung, SL Ma, CPW Cheng, ST Cheng, FHY Lai, BHK Yip, SYS Wong.
Acquisition of data: LCW Lam, WC Chan, ATC Lee, Z Huo, VC Lin.
Analysis or interpretation of data: LCW Lam, WC Chan, ATC Lee, Z Huo, VC Lin.
Drafting of the manuscript: All authors.
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.

All authors have disclosed no conflicts of interest.

The authors thank the participants, their family members, and staff of the participating long-term care homes for the generous support they offered. Special thanks are extended to the Hospital Authority psychogeriatric teams for encouraging selected old age homes to participate.

Part of this study was presented as oral presentation at the 2024 International Congress on Neuropsychiatry in Melbourne, Australia, 27-29 October 2024.

This commissioned study was funded by the Health and Medical Research Fund of the Hong Kong SAR Government (Ref No.: MHS-P1 Part 3). The funder had no role in study design, data collection, analysis, interpretation, or manuscript preparation.

This research was approved by the Survey and Behavioural Research Ethics Committee (Ref No.: SBRE 18-628) and the Clinical Research Ethics Committee of The Chinese University of Hong Kong, Hong Kong (Ref No.: CREC NTEC CUHK 2018-0529). Written consent was obtained from each participant or their first-degree relative (for those with profound cognitive impairments or sensory deficits) before joining the study.

1. López-Otín C, Blasco MA, Partridge L, Serrano M, Kroemer G. The hallmarks of aging. Cell 2013;153:1194–217. Crossref

2. Census and Statistics Department, Hong Kong SAR Government. Hong Kong Population Projections 2022-2046. Aug 2023. Available from: https://www.censtatd.gov.hk/en/data/stat_report/product/B1120015/att/B1120015092023XXXXB01.pdf. Accessed 17 Mar 2025.

3. GBD 2019 Dementia Forecasting Collaborators. Estimation of the global prevalence of dementia in 2019 and forecasted prevalence in 2050: an analysis for the Global Burden of Disease Study 2019. Lancet Public Health 2022;7:e105-25. Crossref

4. Lam LC, Tam CW, Lui VW, et al. Prevalence of very mild and mild dementia in community-dwelling older Chinese people in Hong Kong. Int Psychogeriatr 2008;20:135-48. Crossref

5. Jia L, Du Y, Chu L, et al. Prevalence, risk factors, and management of dementia and mild cognitive impairment in adults aged 60 years or older in China: a cross-sectional study. Lancet Public Health 2020;5:e661-71. Crossref

6. Livingston G, Huntley J, Liu KY, et al. Dementia prevention, intervention, and care: 2024 report of the Lancet standing Commission. Lancet 2024;404:572-628. Crossref

7. Brayne C, Davis D. Making Alzheimer’s and dementia research fit for populations. Lancet 2012;380:1441-3. Crossref

8. Livingston G, Huntley J, Sommerlad A, et al. Dementia prevention, intervention, and care: 2020 report of the Lancet Commission. Lancet 2020;396:413-46. Crossref

9. Fung AW, Chan WC, Wong CS, et al. Prevalence of anxiety disorders in community dwelling older adults in Hong Kong. Int Psychogeriatr 2017;29:259-67. Crossref

10. Lam LC, Tam CW, Lui VW, et al. Screening of mild cognitive impairment in Chinese older adults—a multistage validation of the Chinese abbreviated mild cognitive impairment test. Neuroepidemiology 2008;30:6-12. Crossref

11. Wong A, Law LS, Liu W, et al. Montreal Cognitive Assessment: one cutoff never fits all. Stroke 2015;46:3547-50. Crossref

12. Yeung PY, Wong LL, Chan CC, Leung JL, Yung CY. A validation study of the Hong Kong version of Montreal Cognitive Assessment (HK-MoCA) in Chinese older adults in Hong Kong. Hong Kong Med J 2014;20:504-10. Crossref

13. Morris JC, Ernesto C, Schafer K, et al. Clinical Dementia Rating training and reliability in multicenter studies: the Alzheimer’s Disease Cooperative Study experience. Neurology 1997;48:1508-10. Crossref

14. Kaufer DI, Cummings JL, Ketchel P, et al. Validation of the NPI-Q, a brief clinical form of the Neuropsychiatric Inventory. J Neuropsychiatry Clin Neurosci 2000;12:233-9. Crossref

15. Conwell Y, Forbes NT, Cox C, Caine ED. Validation of a measure of physical illness burden at autopsy: the Cumulative Illness Rating Scale. J Am Geriatr Soc 1993;41:38-41. Crossref

16. Malmstrom TK, Morley JE. SARC-F: a simple questionnaire to rapidly diagnose sarcopenia. J Am Med Dir Assoc 2013;14:531-2. Crossref

17. Mok CC, Siu AM, Chan WC, Yeung KM, Pan PC, Li SW. Functional disabilities profile of Chinese elderly people with Alzheimer’s disease—a validation study on the Chinese version of the Disability Assessment for Dementia. Dement Geriatr Cogn Disord 2005;20:112-9. Crossref

18. Lee AT, Richards M, Chan WC, Chiu HF, Lee RS, Lam LC. Lower risk of incident dementia among Chinese older adults having three servings of vegetables and two servings of fruits a day. Age Ageing 2017;46:773-9. Crossref

19. Leung GT, de Jong Gierveld J, Lam LC. Validation of the Chinese translation of the 6-item De Jong Gierveld Loneliness Scale in elderly Chinese. Int Psychogeriatr 2008;20:1262-72. Crossref

20. Leung GT, Fung AW, Tam CW, et al. Examining the association between late-life leisure activity participation and global cognitive decline in community-dwelling elderly Chinese in Hong Kong. Int J Geriatr Psychiatry 2011;26:39-47. Crossref

21. Tsai PS, Wang SY, Wang MY, et al. Psychometric evaluation of the Chinese version of the Pittsburgh Sleep Quality Index (CPSQI) in primary insomnia and control subjects. Qual Life Res 2005;14:1943-52. Crossref

22. American Psychiatric Association, DSM-5 Task Force. Diagnostic and Statistical Manual of Mental Disorders: DSM-5. 5th Edition. American Psychiatric Publishing; 2013. Crossref

23. Census and Statistics Department, Hong Kong SAR Government. Population by sex and age group 2022. Available from: https://www.censtatd.gov.hk/en/web_table.html?id=1A. Accessed 10 Jul 2025.

24. Census and Statistics Department, Hong Kong SAR Government. Thematic Report: Older Persons. Feb 2023. Available from: https://www.censtatd.gov.hk/en/data/stat_report/product/B1120118/att/B11201182021XXXXB0100.pdf. Accessed 23 Dec 2024.

25. Lee AT, Richards M, Chan WC, Chiu HF, Lee RS, Lam LC. Intensity and types of physical exercise in relation to dementia risk reduction in community-living older adults. J Am Med Dir Assoc 2015;16:899.e1-7. Crossref

26. Alzheimer’s Disease International. Dementia in the Asia Pacific Region. Nov 2014. Available from: https://www.alzint.org/resource/dementia-in-the-asia-pacific-region/. Accessed 18 Mar 2026.

27. Census and Statistics Department, Hong Kong SAR Government. Hong Kong 2021 Population Census—Thematic Report. Feb 2023. Available from: https://www.censtatd.gov.hk/en/data/stat_report/product/B1120118/att/B11201182021XXXXB0100.pdf. Accessed 17 Mar 2026.

28. Ministry of Civil Affairs, National Working Committee on Aging. 2024 National Bulletin on the Development of Aging Affairs. Jul 2025. Available from: https://www.mca.gov.cn/n152/n165/c1662004999980006089/part/21508.pdf. Accessed 17 Mar 2026.

29. Statista. Share of population aged 60 and older in China from 1950 to 2020 with forecasts until 2100. Available from: https://www.statista.com/statistics/251529/share-of-persons-aged-60-and-older-in-the-chinese-population/. Accessed 17 Dec 2024.

30. Census and Statistics Department, Hong Kong SAR Government. Hong Kong Population Census—Summary Findings. Available from: https://www.censtatd.gov.hk/en/data/stat_report/product/B1120106/att/B11201062021XXXXB01.pdf. Accessed 17 Dec 2024.

31. Institute for Health Metrics and Evaluation, University of Washington. Global Burden of Disease Results 2021. Available from: https://vizhub.healthdata.org/gbd-results/. Accessed 2 Mar 2025.

32. Kalaria RN, Maestre GE, Arizaga R, et al. Alzheimer’s disease and vascular dementia in developing countries: prevalence, management, and risk factors. Lancet Neurol 2008;7:812-26. Crossref

33. Wong CH, Leung GT, Fung AW, Chan WC, Lam LC. Cognitive predictors for five-year conversion to dementia in community-dwelling Chinese older adults. Int Psychogeriatr 2013;25:1125-34. Crossref

34. Flodgren GM, Berg RC. Primary and secondary prevention interventions for cognitive decline and dementia [internet]. Available from: https://pubmed.ncbi.nlm.nih.gov/29553642/. Accessed 18 Mar 2026.

35. Lam BY, Cai Y, Akinyemi R, et al. The global burden of cerebral small vessel disease in low- and middle-income countries: a systematic review and meta-analysis. Int J Stroke 2023;18:15-27. Crossref

36. Lam LC, Tam CW, Lui VW, et al. Modality of physical exercise and cognitive function in Hong Kong older Chinese community. Int J Geriatr Psychiatry 2009;24:48-53. Crossref

37. Lam LC, Chau RC, Wong BM, et al. A 1-year randomized controlled trial comparing mind body exercise (Tai Chi) with stretching and toning exercise on cognitive function in older Chinese adults at risk of cognitive decline. J Am Med Dir Assoc 2012;13:568.e15-20. Crossref

38. Lee AT, Luo Y, Huo Z, Shi L, Chu WC, Lam LC. Effect of increasing cognitive activity participation on default mode network in older adults with subjective cognitive decline: a randomised controlled trial. EBioMedicine 2024;102:105082. Crossref

39. Huo Z, Yip BH, Lee AT, et al. Healthcare utilization and economic costs of neurocognitive disorders in community-dwelling older Chinese adults: A comparison with 9 Asian economies. J Alzheimers Dis 2025;107:515-28. Crossref