Low bone status on calcaneal quantitative ultrasound among community screening attendees in rural northern Vietnam: a cross-sectional study

Abstract

Objective

To estimate the proportion with low bone status among adults aged 50 years and older attending a community bone health screening programme in rural northern Vietnam, using calcaneal quantitative ultrasound (QUS), and to examine associated factors.

Design

Community-based cross-sectional study.

Setting

Kieu Phu, a rural commune of Hanoi, Vietnam (population approximately 60 900), where a one-off community bone health screening programme was conducted between October and December 2025.

Participants

A total of 592 community-dwelling adults aged 50 years and older attending the screening programme.

Primary and secondary outcome measures

The primary outcome was low bone status, operationally defined as a QUS T-score ≤−2.5. Crude and adjusted ORs were estimated using logistic regression.

Results

The mean age of participants was 74.8 years (SD 7.7) and 58.4% were women. Overall, 355 of 592 participants had low bone status on calcaneal QUS (60.0%, 95% CI 56.0% to 63.8%). In the adjusted model, female sex showed the strongest association with low bone status (adjusted OR 8.71, 95% CI 5.77 to 13.15), and underweight participants had higher odds than those with a body mass index of 18.5–22.9 kg/m² (adjusted OR 2.11, 95% CI 1.15 to 3.87). Age ≥70 years, education level, former occupation, main income source, high sodium intake, hypertension and diabetes mellitus were not independently associated with low bone status.

Conclusions

Low bone status on calcaneal QUS was common among attendees of a rural community screening programme in northern Vietnam and was independently associated with female sex and underweight status. Because participants were screening attendees rather than a probability sample, these findings are not prevalence estimates for the wider rural population. Calcaneal QUS is not equivalent to dual-energy X-ray absorptiometry (DXA)-based diagnosis but may support early risk identification where DXA access is limited, provided screening is linked to accessible confirmatory assessment and referral.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• This study was conducted in a real-world, community-based screening setting in a rural commune of northern Vietnam, where access to dual-energy X-ray absorptiometry (DXA) is limited.

• The study included a relatively large sample of community-dwelling adults aged 50 years and older and used standardised field procedures for calcaneal quantitative ultrasound (QUS) and anthropometric assessment.

• Bone status was assessed using calcaneal QUS instead of DXA; therefore, the findings do not represent DXA-defined osteoporosis.

• Participants were recruited from a community screening programme rather than a probability sample of the commune population, limiting the representativeness and generalisability of the findings.

• Important bone-health covariates were not collected, including personal and parental fracture history, falls history, smoking, alcohol use, menopausal status, medication use including glucocorticoids, physical activity, use of walking aids, cognitive status and calcium or vitamin D intake. This may have resulted in residual confounding and prevented reliable estimation of individual fracture probability.

Background

Osteoporosis and fragility fractures are major contributors to morbidity, disability and healthcare burden in ageing populations.^{1 2} With rapid population ageing across Asia, the burden of osteoporotic fractures is expected to rise substantially, making early identification of poor bone health an increasingly important clinical and public health priority.^{3 4} In older adults, bone loss often progresses silently until a fracture occurs, by which time the consequences may include chronic pain, functional decline, loss of independence and reduced quality of life.¹ Population-based Vietnamese data indicate that osteoporosis affects approximately 27% of women and 13% of men aged 50 years and older, while 49% of women and 35% of men in this age group meet US National Osteoporosis Foundation criteria for treatment.⁵ Approximately 12% of community-dwelling Vietnamese adults aged 50 years and older have a prevalent vertebral fracture, with an incidence of 6.6 per 1000 person-years.⁶

Dual-energy X-ray absorptiometry (DXA) is the reference standard for measuring bone mineral density and underpins the conventional diagnosis of osteoporosis.^{1 7} However, access to DXA in community settings is limited by its cost, lack of portability and uneven availability, particularly in low-resource and rural areas.^{4 8} Across the Asia-Pacific region, limited DXA access and persistent treatment gaps continue to hinder timely diagnosis and management, including outside major urban centres in Vietnam.^{3 4} Recent regional audit data further illustrate these health-system constraints: across 22 Asia-Pacific countries and regions, DXA waiting times range from same-day access to nearly 1 year and scan costs range from approximately US$15 to US$175. Centralised fracture databases are available in only 10 of the 22 countries and regions, and fracture risk assessment tools were reported as not being routinely used in Vietnam.⁴

Calcaneal quantitative ultrasound (QUS) offers a potential approach to earlier bone health assessment in settings where DXA access is constrained. QUS is portable, radiation-free, relatively inexpensive and feasible for use in large-scale or community-based screening programmes.^9–11 Previous studies have shown that calcaneal QUS parameters are associated with bone status and fracture risk, and recent evidence suggests that QUS may be a promising tool for bone health assessment in older populations, especially where access to DXA is constrained.^{9 12 13} Nevertheless, QUS should be interpreted cautiously. Although it may support early case finding and risk stratification, it is not equivalent to DXA-based diagnosis.^{7 14} Vietnamese research has shown discordance between instrument-derived and population-derived T-scores, indicating that instrument-generated QUS T-scores should not be used uncritically to diagnose osteoporosis.¹⁴ More recent Vietnamese evidence has also shown limited diagnostic agreement between calcaneal QUS and DXA, reinforcing its role as a screening rather than diagnostic tool.⁷

Evidence on bone health assessed through community-based tools in Vietnam remains limited. Earlier Vietnamese studies using heel ultrasound suggested that low bone status is common, and that age, female sex and anthropometric characteristics are important correlates, but contemporary data from rural community settings remain scarce.^{14 15} This gap is important because rural and peri-urban populations may face both rapid demographic ageing and persistent barriers to specialist assessment.⁴ In these contexts, community screening may provide a feasible opportunity to identify individuals who may benefit from further evaluation, counselling or referral.^{10 16}

Kieu Phu, a rural commune in northern Vietnam, provides a relevant setting to examine community-based bone health screening where access to DXA and specialist assessment is limited. Calcaneal QUS may provide a practical first-line approach to identify older adults with poorer bone health profiles and groups who may benefit from further assessment or targeted prevention.^{7 9 10} Findings from this setting may help inform local bone health education, risk-prioritised prevention and referral pathways in similar rural communities.

Therefore, this study aimed to estimate, among adults aged 50 years and older attending a community bone health screening programme in a rural commune of northern Vietnam, the proportion with low bone status identified using calcaneal QUS, and to examine factors associated with low bone status.

Methods

Study design and setting

This community-based cross-sectional study was conducted in Kieu Phu commune, a rural commune of Hanoi, Vietnam. Kieu Phu covers approximately 34.5 km² and has a population of about 60 900 residents.¹⁷ Primary healthcare is delivered by the commune health station, which refers patients to district-level services and, for specialist investigation, to city or central hospitals in Hanoi. Bone densitometry is not always available at commune or district level.

Field-based screening and data collection were conducted from October to December 2025 as part of a one-off, time-limited community bone health screening programme coordinated with the commune health station.

Study population and recruitment

Community-dwelling adults aged 50 years and older who were permanent residents of Kieu Phu commune were eligible for inclusion. Additional inclusion criteria were the ability to communicate with the research team and to provide written informed consent. For this study, ability to communicate was defined as being able to hear, understand and respond to the interviewer-administered questionnaire in Vietnamese. As the questionnaire was administered orally by trained staff, no minimum level of reading or writing ability was required.

Individuals with severe hearing or speech impairment, or apparent cognitive impairment that precluded reliable responses, were not eligible. Formal cognitive or hearing assessments were not undertaken; these eligibility criteria were assessed pragmatically by trained interviewing staff at enrolment. Individuals were also excluded if they were unable to undergo calcaneal QUS because of severe foot deformity, an active heel wound or another condition preventing valid heel assessment. Individuals who declined participation were not enrolled.

Participants were recruited through announcements made by the commune health station and local community channels before scheduled screening sessions. All eligible individuals who presented to the screening programme during the data collection period were invited consecutively to participate. Written informed consent was obtained from all participants before data collection. A total of 592 participants completed screening and were included in the analytic sample. The number of individuals assessed but not enrolled was not separately recorded because recruitment was integrated into the administrative processes of the community screening programme. No formal sample size calculation was performed; the analytic sample comprised all eligible participants who completed screening during the study period.

Patient and public involvement

This study was conducted as part of a community bone health screening programme organised in collaboration with the local health station in response to local service needs. Patients and members of the public were not involved in the design of the research question, the selection of outcome measures, the statistical analysis plan or the interpretation of the findings. They were also not involved in the writing of the manuscript or planning for dissemination of the findings.

Data collection procedures

Data were collected during a single screening visit using a structured interviewer-administered questionnaire and standardised physical measurements. Trained healthcare staff administered the questionnaire and performed physical measurements according to the study protocol. The questionnaire captured sociodemographic information, including age, sex, education level, former occupation and main income source. It also captured self-reported chronic conditions, including hypertension, diabetes mellitus, gastric disease, coronary artery disease, arthritis, chronic obstructive pulmonary disease, cancer and other comorbidities.

Participants were additionally asked about habitual dietary patterns using brief self-report screening items. These included indicators of high sodium intake, low sodium intake, a plant-based or high-fibre dietary pattern and a high-fat or high-lipid dietary pattern. These indicators reflected self-reported dietary tendencies and were not mutually exclusive; participants could report more than one.

Anthropometric measurements

Height and weight were measured using standard procedures with participants wearing light clothing and no shoes. Body mass index (BMI) was calculated as weight in kilograms divided by height in metres squared. BMI was categorised using commonly applied Asian cutoffs as underweight, <18.5 kg/m²; 18.5–22.9 kg/m²; and ≥23.0 kg/m².¹⁸

Bone status assessment

Bone status was assessed using calcaneal QUS with the SONOST 3000 device (OsteoSys Co., Seoul, Republic of Korea). The device measures speed of sound and broadband ultrasound attenuation and automatically generates a QUS-derived T-score and a bone quality index.¹⁹

Measurements were performed at the right calcaneus with the participant in a seated position, following the manufacturer’s operating instructions. A coupling gel was applied to ensure adequate acoustic transmission. Each measurement took approximately 15 s. Device calibration was performed daily before data collection using the manufacturer-supplied phantom. If a measurement was flagged as technically unstable, the participant was repositioned and the measurement repeated. One valid measurement per participant was retained for analysis.¹⁹

Outcome definition

The primary outcome was low bone status on calcaneal QUS, operationally defined as a QUS T-score ≤−2.5.¹⁰ This screening threshold was informed by previous Vietnamese and Asian studies using QUS-derived cut-offs and was not intended to represent DXA-defined osteoporosis.^{7 14}

For descriptive purposes, QUS T-scores were also summarised as normal (T-score ≥−1.0), intermediate low bone status (T-score <−1.0 to −2.49), and low bone status (T-score ≤−2.5). For regression analyses, the outcome was analysed as a binary variable: low bone status (QUS T-score ≤−2.5) versus no low bone status (QUS T-score >−2.5).¹⁰

Covariates

Variables considered in the analysis were selected based on clinical relevance, previous literature, and availability within the dataset. These included age group, sex, BMI category, education level, former occupation, main income source, high sodium intake, hypertension and diabetes mellitus.^{13 15 20}

Age was presented in prespecified age groups for descriptive analysis and comparison by sex. For multivariable regression, age was entered as a dichotomous variable (≥70 vs <70 years), consistent with the prespecified analytic framework of the study.

Data management and quality assurance

Data were entered electronically at the point of collection using Google Forms and exported to Google Sheets for review and cleaning. Range and logic checks were undertaken before analysis to identify implausible values and coding inconsistencies. The final analytic dataset was de-identified and stored with access restricted to authorised study team members.

Missing data were minimal. Height was missing for 10 of 592 participants (1.7%), and weight was missing for 11 of 592 participants (1.9%). Missing values were imputed using the sample mean for the respective anthropometric measure before BMI calculation. Given the small proportion of missing data, simple mean imputation was used for the primary analysis.

Statistical analysis

Continuous variables were summarised using means and SDs. Categorical variables were presented as counts and percentages.

The proportion of participants with low bone status was estimated overall and stratified by age group and sex. Participant characteristics were compared between those with and without low bone status using Pearson’s χ² test for categorical variables.

Logistic regression models were used to estimate crude and adjusted ORs with 95% CIs for factors associated with low bone status. The multivariable model included age group, sex, BMI category, education level, former occupation, main income source, high sodium intake, hypertension and diabetes mellitus. BMI 18.5–22.9 kg/m² was the reference BMI category.

Adjusted predicted probabilities of low bone status across BMI categories were estimated from the multivariable logistic regression model using marginal effects and presented graphically. Statistical significance was defined as a two-sided p value below 0.05. All analyses were performed using Stata V.17 (StataCorp, College Station, Texas, USA).

Ethics approval

The study was approved by the Institutional Review Board (IRB) of the Vietnam University of Traditional Medicine (IRB-VN01.014) and conducted in accordance with the principles of the Declaration of Helsinki. Written informed consent was obtained from all participants before enrolment, following provision of information about the study objectives, procedures and data confidentiality. Collected data were de-identified and used solely for research purposes.

Reporting guideline

This study is reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement.²¹

Results

Participant characteristics

A total of 592 adults aged 50 years and older who attended the community screening programme were included in the analysis (table 1). The mean age was 74.8 years (SD 7.7), with 449 participants (75.8%) aged 70 years or older. Women accounted for 58.4% of the sample. The mean BMI was 21.4 kg/m² (SD 3.0); 85 participants (14.4%) were underweight and 155 (26.2%) had a BMI of ≥23.0 kg/m².

Table 1. Characteristics of participants attending the community bone health screening programme (n=592).

Characteristics	All attendees (n=592), n (%)
Age ≥70 years	449 (75.8)
Sex
Male	246 (41.6)
Female	346 (58.4)
BMI <18.5 kg/m²	85 (14.4)
BMI 18.5–22.9 kg/m²	352 (59.5)
BMI ≥23.0 kg/m²	155 (26.2)
Education level
No formal schooling	61 (10.3)
Primary school	277 (46.8)
Secondary school	201 (34.0)
High school and above	53 (9.0)
Former occupation
Farmer	491 (82.9)
Civil servant/Public official	53 (9.0)
Other	48 (8.1)
Main income source
No income	112 (18.9)
Social allowance	363 (61.3)
Pension/Other income	117 (19.8)
Dietary pattern indicators
High sodium intake	67 (11.3)
Plant-based/High-fibre diet	248 (41.9)
Low sodium intake	160 (27.0)
High-fat/High-lipid diet	52 (8.8)
Comorbidities
Hypertension	235 (39.7)
Gastric disease	99 (16.7)
Diabetes mellitus	80 (13.5)
Coronary artery disease	70 (11.8)
Arthritis	59 (10.0)
Chronic obstructive pulmonary disease	38 (6.4)
Cancer	18 (3.0)
Other	88 (14.9)
Calcaneal QUS
Low bone status (QUS T-score ≤−2.5)	355 (60.0)

Data are presented as n (%) of 592 screening attendees. Mean age, mean BMI, and mean QUS T-score are reported in the text. Dietary pattern indicators were self-reported and were not mutually exclusive; participants could report more than one dietary tendency. Comorbidities were self-reported and were not mutually exclusive.

BMI, body mass index; QUS, quantitative ultrasound.

Most participants had primary or secondary school education, while 61 (10.3%) had no formal schooling. Most were former farmers (82.9%), and 61.3% reported social allowance as their main source of income. Hypertension was the most frequently reported comorbidity (39.7%), followed by gastric disease (16.7%) and diabetes mellitus (13.5%).

The mean calcaneal QUS T-score was −2.41 (SD 0.79). Overall, 355 of 592 participants had low bone status on calcaneal QUS (60.0%, 95% CI 56.0% to 63.8%).

Proportion of low bone status by age group and sex

The proportion of participants with low bone status on calcaneal QUS varied by age group and sex (figure 1). Among women, the proportion increased from 50.0% in those aged 50–59 years to 86.8% in those aged 80 years and older. Among men, the proportion was consistently lower across age groups, ranging from 22.2% to 35.4%.

Figure 1. Proportion of participants with low bone status on calcaneal quantitative ultrasound (QUS) by age group and sex.

Comparison of participants with and without low bone status

Table 2 compares participant characteristics according to low bone status on calcaneal QUS. Participants with low bone status were more often women, and more often underweight, than participants without low bone status. The distributions of education level, former occupation and main income source also differed between the two groups. Participants with low bone status more frequently had no formal or primary education, reported farming as their former occupation and received social allowance as their main source of income.

Table 2. Comparison of participant characteristics by low bone status on calcaneal quantitative ultrasound (QUS).

Factors	No low bone status (n=237)	Low bone status (n=355)	P value
Age ≥70 years	173 (73.0)	276 (77.7)	0.190
Sex
Male	170 (71.7)	76 (21.4)	<0.001
Female	67 (28.3)	279 (78.6)
BMI classification
BMI <18.5 kg/m²	22 (9.3)	63 (17.7)	<0.001
BMI 18.5–22.9 kg/m²	132 (55.7)	220 (62.0)
BMI ≥23.0 kg/m²	83 (35.0)	72 (20.3)
Education level
No formal schooling	11 (4.6)	50 (14.1)	<0.001
Primary school	100 (42.2)	177 (49.9)
Secondary school	96 (40.5)	105 (29.6)
High school and above	30 (12.7)	23 (6.5)
Former occupation
Farmer	179 (75.5)	312 (87.9)	<0.001
Civil servant/Public official	26 (11.0)	27 (7.6)
Other	32 (13.5)	16 (4.5)
Main income source
No income	42 (17.7)	70 (19.7)	0.006
Social allowance	133 (56.1)	230 (64.8)
Pension/Other income	62 (26.2)	55 (15.5)
High sodium intake	31 (13.1)	36 (10.1)	0.270
Hypertension	94 (39.7)	141 (39.7)	0.990
Diabetes mellitus	34 (14.3)	46 (13.0)	0.630

Data are presented as n (column %). P values are from Pearson’s χ2 test.

BMI, body mass index.

No statistically significant differences were observed between participants with and without low bone status with respect to age group, high sodium intake, hypertension or diabetes mellitus.

Factors associated with low bone status

Results from logistic regression analyses are presented in table 3. In crude analyses, female sex, lower BMI, lower educational attainment and lower main income source were associated with higher odds of low bone status.

Table 3. Factors associated with low bone status on calcaneal quantitative ultrasound (QUS): logistic regression analysis.

Factors	Crude OR (95% CI)	P value	Adjusted OR (95% CI)	P value
Age ≥70 years	1.29 (0.88 to 1.89)	0.186	1.30 (0.81 to 2.10)	0.280
Sex
Male	1.00 (Ref)		1.00 (Ref)
Female	9.31 (6.37 to 13.62)	<0.001	8.71 (5.77 to 13.15)	<0.001
BMI category
BMI 18.5–22.9 kg/m²	1.00 (Ref)		1.00 (Ref)
BMI <18.5 kg/m²	1.72 (1.01 to 2.92)	0.046	2.11 (1.15 to 3.87)	0.016
BMI ≥23.0 kg/m²	0.52 (0.35 to 0.76)	0.001	0.69 (0.44 to 1.09)	0.114
Education level
High school and above	1.00 (Ref)		1.00 (Ref)
No formal schooling	5.93 (2.54 to 13.86)	<0.001	1.49 (0.54 to 4.17)	0.443
Primary school	2.31 (1.27 to 4.19)	0.006	1.11 (0.50 to 2.43)	0.799
Secondary school	1.43 (0.78 to 2.62)	0.253	0.75 (0.35 to 1.60)	0.455
Former occupation
Civil servant/Public official	1.00 (Ref)		1.00 (Ref)
Farmer	1.68 (0.95 to 2.96)	0.074	0.82 (0.38 to 1.76)	0.605
Other	0.48 (0.21 to 1.08)	0.076	0.42 (0.16 to 1.11)	0.081
Main income source
Pension/Other income	1.00 (Ref)		1.00 (Ref)
No income	1.88 (1.11 to 3.18)	0.019	1.48 (0.77 to 2.84)	0.245
Social allowance	1.95 (1.28 to 2.97)	0.002	1.30 (0.76 to 2.21)	0.334
High sodium intake	0.75 (0.45 to 1.25)	0.270	0.57 (0.30 to 1.08)	0.087
Hypertension	1.00 (0.72 to 1.40)	0.989	0.95 (0.53 to 1.71)	0.870
Diabetes mellitus	0.89 (0.55 to 1.43)	0.628	1.05 (0.69 to 1.60)	0.825

Adjusted odds ratios were estimated from a multivariable logistic regression model including age group, sex, BMI category, education level, former occupation, main income source, high sodium intake, hypertension, and diabetes mellitus.

BMI, body mass index.

In the multivariable model, female sex showed the strongest association with low bone status (adjusted OR 8.71, 95% CI 5.77 to 13.15). Underweight participants also had higher adjusted odds of low bone status than those with a BMI of 18.5–22.9 kg/m² (adjusted OR 2.11, 95% CI 1.15 to 3.87). Participants with a BMI of ≥23.0 kg/m² had lower adjusted odds than those in the reference BMI category, but the association was not statistically significant (adjusted OR 0.69, 95% CI 0.44 to 1.09).

In the prespecified adjusted model, age ≥70 years was not independently associated with low bone status after adjustment (adjusted OR 1.30, 95% CI 0.81 to 2.10). Education level, former occupation, main income source, high sodium intake, hypertension and diabetes mellitus were also not independently associated with low bone status.

Adjusted predicted probability of low bone status by BMI category

Marginal effects derived from the multivariable logistic regression model showed an inverse gradient between BMI and the adjusted probability of low bone status (figure 2). The adjusted predicted probabilities were 72.3% among underweight participants, 59.9% among those with a BMI of 18.5–22.9 kg/m², and 53.4% among those with a BMI of ≥23.0 kg/m².

Figure 2. Adjusted predicted probability of low bone status on calcaneal quantitative ultrasound by body mass index (BMI) category.

Discussion

In this community-based cross-sectional study, low bone status on calcaneal QUS was common among adults aged 50 years and older who attended a rural community bone health screening programme in northern Vietnam. Female sex was the factor most strongly associated with low bone status, and underweight participants had higher adjusted odds than those with a BMI of 18.5–22.9 kg/m². These findings identify women and underweight adults as important higher-risk groups within this screened population.

Our findings are broadly consistent with previous evidence from Vietnam and other Asian settings showing that bone status assessed by heel ultrasound is generally lower in women and declines with age, with body size also playing an important role.^{12 13 15} Notably, regional heel ultrasound data from seven Asian countries showed a particularly steep age-related decline among Vietnamese women, with mean T-scores falling below −2.5 at approximately 71–75 years of age.¹² Given the mean age of 74.8 years in our sample, the high proportion of attendees with low bone status is consistent with this regional pattern. However, the observed 60.0% should not be compared directly with population-based Vietnamese estimates of DXA-defined osteoporosis or vertebral fracture.^{5 6} Our estimate was derived from calcaneal QUS in a self-selected screening sample. Instrument-derived QUS T-scores may also overclassify low bone status in Vietnamese adults, and diagnostic agreement between calcaneal QUS and DXA is limited.^{7 14} The outcome should therefore be interpreted as low bone status detected by calcaneal QUS, not DXA-defined osteoporosis.

The strong association with female sex is consistent with the established pattern of greater bone loss in later life among women.^{12 13 15} In descriptive analyses, the proportion with low bone status increased markedly across age groups among women but remained substantially lower among men. Age ≥70 years was not independently associated with low bone status after adjustment. This may partly reflect the concentration of the sample in the older age range and the use of a dichotomous age variable, which reduced variation available for detecting a more gradual age-related association. The relationship with BMI was also consistent across analyses. Underweight participants had higher adjusted odds of low bone status, and adjusted predicted probabilities decreased across increasing BMI categories. Low BMI may therefore provide a simple clinical marker for prioritising further bone health assessment during community screening.^{13 15 20}

The association with underweight status may also be relevant in the wider nutritional context of rural older adults in Vietnam. Previous research among rural Vietnamese adults aged 60 years and older reported that 7.3% were malnourished and 50.2% were at risk of malnutrition.²² National data have also shown low calcium intake among Vietnamese women, with mean intake of approximately 506 mg/day,^{23 24} compared with the national recommendation of 900–1000 mg/day for women aged 50 years and older.²⁵ Our study did not directly assess nutritional status or quantitative calcium intake, so these findings provide context rather than evidence of an explanatory pathway. Socioeconomic characteristics may also influence diet, access to assessment and treatment, but these pathways were not examined in this study. Household income, household assets and validated measures of socioeconomic deprivation were not collected; therefore, receipt of social allowance should not be interpreted as a direct measure of poverty.

From a practice and public health perspective, community-based calcaneal QUS may have value as an initial risk-identification tool where DXA access is limited.^{9 10} However, its value depends on screening being linked to appropriate counselling, confirmatory assessment and referral. A two-step QUS-to-DXA pathway may be relevant in rural settings, but its feasibility and cost-effectiveness require local evaluation. Modelling from higher-income settings suggests that QUS used to select individuals for DXA can be cost-effective under some assumptions, with results depending on the relative costs and accessibility of QUS and DXA.²⁶ Regional data also show substantial variation in DXA costs and waiting times across the Asia-Pacific region.⁴ These economic findings cannot be assumed to transfer directly to Vietnam. Local studies are needed to assess screening costs, referral completion, confirmatory DXA, treatment uptake and the overall budget impact of community-based pathways.

This study has several strengths, including its real-world community screening setting, relatively large sample and standardised field procedures. Several limitations should also be considered. First, the cross-sectional design precludes causal inference. Second, participants were attendees of a community screening programme rather than a probability sample of the commune population, a sampling approach also used in other campaign-based QUS screening studies.¹⁶ This limits representativeness and means that the observed proportion with low bone status should not be interpreted as population prevalence. We were also unable to quantify the number of individuals assessed for eligibility, excluded, or declining participation because these counts were not separately recorded during the community screening programme. This further limits our ability to characterise potential selection into the analytic sample. Because a complete sampling frame and age-specific data for non-attendees were unavailable, we could not determine whether the lower representation of adults aged 50–69 years reflected the commune age structure, differential screening uptake, or both. Third, bone status was assessed using calcaneal QUS rather than DXA, and QUS-based classification should not be considered interchangeable with DXA-defined osteoporosis.^{7 14}

Fourth, several important bone-health and fracture-risk covariates were not collected, including personal and parental fracture history, falls history, smoking, alcohol use, menopausal status and age at menopause, medication use including glucocorticoids, physical activity, use of walking aids and calcium or vitamin D intake.¹ Cognitive status was also not formally assessed among enrolled participants. The absence of these variables may have resulted in residual confounding and prevented estimation of individual fracture probability using clinical risk factors. Fifth, dietary indicators were based on brief self-reported items rather than a validated dietary instrument or quantitative intake measure. Recall bias and misclassification are therefore possible. Finally, missing anthropometric values were imputed using the sample mean. Although missingness was low, mean imputation may reduce variability and is less robust than multiple imputation.

Future studies should evaluate community screening in more representative samples, incorporate key clinical fracture-risk factors, and compare calcaneal QUS with DXA in similar rural settings. Implementation research should also examine screening uptake, referral completion, treatment initiation and cost-effectiveness.

Conclusion

Among adults aged 50 years and older attending a community bone health screening programme in rural northern Vietnam, low bone status on calcaneal QUS was common. Female sex and underweight status were independently associated with low bone status. Calcaneal QUS should not be interpreted as equivalent to DXA-based diagnosis of osteoporosis, but it may support early risk identification in resource-limited community settings. These findings may help inform targeted bone health education, further assessment and referral, particularly for women and underweight adults. For community screening to contribute meaningfully to fracture prevention, screening should be linked to accessible confirmatory assessment, appropriate counselling and referral for treatment when clinically indicated.

Data availability statement

Data are available upon reasonable request.