Risk factors for community-acquired pressure injuries: Findings from a large-scale Asian case–control study

Abstract

Background

Pressure injuries, also known as pressure ulcers, remain a major global health problem, causing significant clinical, economic, and psychosocial burdens despite advances in prevention and treatment. While most research has focused on hospital-acquired cases, community-acquired pressure injuries are increasingly recognised, with prevalence in some settings exceeding hospital-acquired rates. Evidence from Asian healthcare systems is limited, yet region-specific factors such as ageing populations, caregiver dependence, and variable access to wound care highlight the urgent need for contextualised prevention strategies.

Objective

To identify the risk factors associated with pressure injuries that developed in the community among adults admitted to an acute care hospital.

Design

Retrospective, unmatched case–control study.

Setting

Acute tertiary hospital.

Participants

Electronic medical records from January 2021 to June 2022, which include 52,480 admissions (5772 cases and 46,708 controls).

Methods

The study was conducted using electronic medical records from January 2021 to June 2022. The cases were adults with community-acquired pressure injuries documented within 24 h of admission; controls had no pressure injuries on admission. Descriptive statistics and effect sizes were used to summarise the relationships between the risk factors and community-acquired pressure injuries. Statistically significant predictors were identified and included in the multivariable logistic regression, and interaction terms were tested for potential effect modification.

Results

Among 52,480 admissions, 5772 were cases and 46,708 were controls. The most common site of injury was the sacrum/coccyx (51%). On univariate analysis, cases were older, more likely to be unemployed, and had greater functional limitations. Moist skin and diaper use were more prevalent, along with poorer nutritional and inflammatory profiles, including lower serum albumin levels and higher C-reactive protein levels.

Independent predictors included unemployment, poor or inadequate nutrition, lower serum albumin, functional dependence, moist skin, hypertension, and a history of peptic ulcer disease. A significant interaction between albumin and sex indicated a modest sex-based difference in the association between albumin and pressure injury risk.

Conclusion

Community-acquired pressure injury risk is multifactorial, involving sociodemographic, functional, nutritional, and microclimate factors. Moisture exposure and mobility limitation were key predictors. A modest sex-based difference in the association between albumin and pressure injury risk was observed. Prevention strategies that integrate nutritional support, mobility optimisation, and moisture management may reduce the incidence of community-onset pressure injuries in ageing populations.

Registration

Not registered.

What is already known.

• •Community-acquired pressure injuries are under-recognised and poorly understood compared to hospital-acquired cases.
• •Nutritional vulnerability, reduced mobility, and moisture exposure have been suggested as contributing factors, but evidence remains limited.

Alt-text: Unlabelled box dummy alt text

What this paper adds.

• •This study identifies multifactorial predictors of community-acquired pressure injuries, including unemployment, poor nutrition, low serum albumin, functional dependence, moist skin, hypertension, and peptic ulcer disease.
• •It demonstrates a sex-based difference in the association between albumin levels and pressure injury risk.
• •Findings highlight the need for integrated prevention strategies combining nutritional support, mobility optimisation, and moisture management to reduce incidence in ageing populations.

Alt-text: Unlabelled box dummy alt text

1. Introduction

Pressure injuries, also known as pressure ulcers or bedsores, are defined as “localised damage to the skin and/or underlying tissue, usually over a bony prominence or related to a medical or other device, resulting from prolonged pressure or pressure in combination with shear” (Haesler, 2019). Despite advances in preventive care and treatment technologies, pressure injuries continue to occur frequently across care settings, imposing substantial clinical, economic, and psychosocial burdens on patients, caregivers, and healthcare systems (Kandula, 2025). Globally, more than one in ten hospitalised adults develops pressure injuries (Li et al., 2020), and the cost of treating hospital-acquired pressure injuries alone exceeded USD 26.8 billion in 2016 (Padula and Delarmente, 2019). In the United States, pressure injuries affect over 2.5 million individuals annually, cause >60,000 deaths, and cost USD 9–11 billion (Hughes, 2008). Similar burdens were reported in the United Kingdom (UK), with 700,000 cases per year, 27,000 associated deaths, and costs to the NHS of £1.4–£2.1 billion (Bennett et al., 2004). In Australia, the total cost of Pressure injuries in 2020 exceeded AUD 9.1 billion annually, of which AUD 3.6 billion was from direct treatment costs, alongside nearly AUD 500 million in productivity losses (Nghiem et al., 2022). In Singapore, patients with hospital-acquired Pressure injuries incur significantly higher costs and longer lengths of stay compared with those without (SGD 35,936 vs SGD 6266; 30 days vs 6 days; p < .0005) (Lim and Ang, 2017).

Besides economic burden, Pressure injuries inflict profound physical and psychological burdens. A meta-synthesis of qualitative studies described patient and caregiver experiences as “living with pain, smell, and sleep deprivation; living with fear, anxiety, worry; and loss of privacy, dignity, control and personal autonomy” (Burston et al., 2023). A meta-analysis of eight observational studies involving 5523 elderly patients found that those with Pressure injuries had nearly a twofold increase in risk of death. Advanced pressure injuries carried even greater mortality, with Stage 3–4 Pressure injuries associated with 2.5 times the mortality risk. Survival curve analyses confirmed this trend, with an estimated two times higher risk of death during follow-up periods of up to three years (Song et al., 2019).

While much of the literature and quality improvement efforts have focused on hospital-acquired pressure injuries, there is growing recognition of the importance of community-acquired pressure injuries. Community-acquired pressure injuries that develop outside of acute care settings, such as in patients’ homes, nursing homes, or other community environments, and are present upon admission to healthcare facilities (Aloweni et al., 2024a). Recent studies indicate that the prevalence of community-acquired pressure injuries may exceed that of hospital-acquired pressure injuries in some settings, with community-acquired cases accounting for a significant proportion of all pressure injuries identified in hospitals (Ding et al., 2022; Graves et al., 2020).

Community-acquired pressure injuries result from a multifactorial interaction between intrinsic and extrinsic influences. Intrinsic factors relate to individual characteristics such as age or physiological condition, which are often non-modifiable. In contrast, extrinsic factors stem from environmental or care-related conditions that are more modifiable to intervention and therefore key targets for prevention (Bluestein and Javaheri, 2008). Older age, poor mobility, multi-comorbidities and poor nutrition are associated with community-acquired pressure injuries (Aloweni et al., 2024b). Whereas sex or gender and cognitive impairment were not consistently identified as risk factors, their influence may be mediated through related variables such as nutritional status, frailty, or functional decline.

The challenge of preventing pressure injuries in the community is further compounded by an ageing population, variability in health system design, and uneven access to preventive or wound care services across Asia. Studies from China and Southeast Asia highlight unique patterns of contribution, including differences in healthcare infrastructure, the availability of specialised wound care teams, reimbursement policies, and caregiver training (Ding et al., 2022; Graves et al., 2020; Shi et al., 2025). Barriers identified in China include limited provider knowledge, resource constraints, insufficient multidisciplinary collaboration, and funding limitations (Shi et al., 2025). In Singapore and other parts of Asia, reliance on family members or foreign domestic helpers for daily care further shapes early detection, prevention practices, and escalation pathways (Aloweni et al., 2024a; Sari et al., 2025).

Although the global epidemiology of pressure injuries has been extensively studied in Western contexts, evidence from Asian healthcare systems remains sparse. Differences in socio-demographic factors, home care infrastructure, reimbursement policies, and caregiver dependence create unique patterns of risk that may not be fully captured by existing hospital-based or Western-derived models. In Singapore and other parts of Asia, the rapidly ageing population and reliance on family or domestic caregivers pose distinct challenges for early recognition and prevention of community-onset wounds. Consequently, region-specific data are essential to contextualize global wound prevention strategies and ensure their applicability to Asian care systems. This study addresses this evidence gap by identifying modifiable risk factors for community-acquired pressure injury in an Asian acute-care setting. Therefore, this study aims to identify risk factors associated with community-acquired pressure injuries among adults admitted to an acute care hospital.

2. Methods

This retrospective, unmatched case–control study was conducted and reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines (Von Elm et al., 2007). Data were extracted from the electronic medical records of patients admitted to a single acute care hospital in Singapore between January 2021 to June 2022 (18 months). Cases were defined as patients admitted to Singapore General Hospital with a community-acquired pressure injury of any stage, as classified by the NPIAP (Haesler, 2019), and documented in the wound care record within 24 h of admission. Controls were patients admitted during the same period who had no pressure injury on admission. An unmatched case–control design was used to allow inclusion of a broad range of candidate variables. Potential confounding was addressed using multivariable logistic regression, with adjustment for key demographic and clinical factors.

The inclusion criteria were: (1) patients aged ≥18 years; (2) admitted between January 2021 and June 2022; and (3) confirmed to have a community-acquired pressure injury (cases) or no pressure injury (controls) on admission. The exclusion criteria were: (1) admission for childbirth; (2) admission for burns; (3) admission solely for observation in the emergency department; (4) neonates in the neonatal unit; (5) patients transferred from another hospital; (6) patients with hospital-acquired pressure injuries attributed to previous hospitalisations; and (7) patients who died during the index hospitalisation.

Patients transferred from other healthcare institutions were excluded because their medical records were not accessible, and the community-acquired status of pressure injuries could not be reliably verified. For patients with multiple admissions during the study period, only the earliest admission was included to avoid duplication, ensure independence of observations, and minimise bias from later admissions that may reflect disease progression.

In total, 50 variables were extracted and examined. Variables in the following domains were analysed: demographic and socioeconomic characteristics; hospitalisation data; lifestyle factors; functional status; nutritional and inflammatory status; microclimate (skin moisture); elimination status; fall-related data; cognitive and behavioural status; comorbidities; and medication use (Table 1). The selection of these variables was guided by existing literature on established and emerging risk factors for pressure injury development, ensuring the inclusion of both patient-related and clinical contributors relevant to community-acquired pressure injury.

Table 1.

Domains of the variables extracted.

Domains	Variables Extracted	Source / Notes
Demographic & Socioeconomic	Age, Sex, Working status	Admission records
Hospitalisation Data	Admitting discipline, Repeat admissions, Length of stay	Medical records
Nursing Assessments	Braden Scale score, Mobility and ambulation status, Continence (urinary, bowel, diaper use), Moisture (Braden subscale), Nutritional risk (MUST), Recent weight loss, Condition on arrival, Level of consciousness, Behaviour on admission	Nursing Care Record (NCR)
Nutritional / Laboratory Results	Body Mass Index (BMI)*, Serum albumin, Haemoglobin, C-reactive protein (CRP)	Laboratory data and dietitian records
Comorbidities	Diabetes mellitus, Hypertension, Chronic kidney disease, Cancer, Asthma, Stroke, Myocardial infarction, Atrial fibrillation, Hyperlipidaemia, Vascular disease, Urinary incontinence, Peptic ulcer	Medical history and ICD coding
Medication Use	Analgesics, Immunosuppressants, Anti-infectives	Medication chart
Fall-related Data	Morse Fall Scale (score and risk category), History of falls, Frequency of falls	Nursing assessment and incident reports

Note:.

^⁎Body Mass Index (BMI) was based on Singapore reference values.

Nutritional status was assessed using the Malnutrition Universal Screening Tool (MUST). A score of ≥2 was defined as high risk of malnutrition (poor/inadequate nutrition), while scores of 0–1 were classified as low to moderate risk (adequate/good nutrition). Serum albumin was analysed as a continuous variable. The normal reference range in our institution is 40–51 g/L, and values below 40 g/L are considered indicative of hypoalbuminemia.

Functional status variables were derived from routine nursing assessments documented in the electronic medical record at the time of admission. Ambulation status was classified based on the level of assistance required for mobility and categorised as independent, requires assistance, or dependent as part of the routine admission assessment conducted by trained nursing staff.

Skin moisture status was derived from routine nursing assessments documented in the electronic medical record at admission. This included the moisture subscale of the Braden Scale, supplemented by clinical observations recorded in the nursing care record. Skin moisture was categorised as rarely moist, occasionally moist, or very moist based on the documented assessment. In addition, elimination-related indicators such as diaper use were included to reflect sustained exposure to moisture. All variables were extracted at admission and verified to represent the patient’s condition within the first 24 h.

2.1. Sample size calculation

Community-acquired pressure injury is a relatively uncommon condition, with incidence rates of approximately 2–4% in clinical practice (Chen et al., 2020). To optimise statistical power, a 1:8 case–control ratio was adopted, as higher ratios provide additional precision for rare outcomes, with minimal gains beyond 1:4 (Katki et al., 2023). A 20% exposure prevalence among controls was assumed. This estimate was supported by published data indicating that 20–50% of hospital inpatients present with malnutrition on admission (Cass and Charlton, 2022), a common risk factor for pressure injury.

Sample size estimation using G*Power (z test for two independent proportions, two-tailed, α = 0.05, power = 0.90, case–control ratio = 1:8) indicated that 606 cases and 4851 controls (total = 5457) were required to detect an odds ratio of 1.40 (Faul et al., 2007). During the study period (January 2021 to June 2022), 52,480 admissions met eligibility criteria, comprising 5772 community-acquired pressure injury cases and 46,708 controls, exceeding the minimum requirement and ensuring robust statistical power for multivariable analyses.

2.2. Data extraction process

Data for this case–control study were extracted from the hospital’s electronic health records (Sunrise Clinical Manager), encompassing nursing care records, multidisciplinary clinical notes, and laboratory results. Pressure injury data were obtained from wound care documentation. Extraction was performed by authorised personnel independent of the study team through the SingHealth–IHiS Electronic Health Intelligence System (eHints). All datasets were de-identified before release to investigators.

Eligible patient records were included based on predefined inclusion and exclusion criteria. The extracted dataset comprised demographic, clinical, functional, nutritional, comorbidity, medication, and fall-related variables (Table 1). Data cleaning included merging overlapping variables and performing cross-source triangulation (e.g., verifying mobility across nursing and physiotherapy notes).

2.3. Ethical considerations

This study was approved by the SingHealth Centralised Institutional Review Board (Reference No 2022/2406) and conducted in accordance with the ethical principles outlined in the Declaration of Helsinki. Additional institutional approval was obtained from the Chief Medical Board before data extraction from the electronic medical record system. As this was a retrospective study utilising de-identified data, the Institutional Review Board waived the requirement for informed consent. All data were de-identified by an authorised third party before analysis and stored on secure, password-protected institutional servers with restricted access.

2.4. Data analysis

Descriptive statistics were used to summarise patient characteristics. Continuous variables were expressed as means with standard deviations or medians with interquartile ranges, depending on normality, and categorical variables as frequencies and percentages. The normality of continuous variables was assessed using the Shapiro–Wilk test and visual inspection of histograms and Q–Q plots. Between-group comparisons were performed using independent t-tests or Mann–Whitney U tests, as appropriate, and categorical variables were compared using the chi-square test.

Missing data patterns were examined, and variables with >25% missingness were excluded from analysis (Dong and Peng, 2013). Serum albumin and C-reactive protein were examined due to their clinical relevance despite moderate missingness (approximately 40–60%). Serum albumin was retained in the multivariable model, whereas C-reactive protein was excluded due to substantial missing data. The proportion of missing data for all variables is presented in Supplementary Table A. Missingness in these laboratory variables was considered clinically driven rather than random, as tests were performed at the physician's discretion rather than routinely for all patients. In clinical practice, serum albumin is not routinely measured unless clinically indicated. When patients were already identified as at risk of malnutrition based on the MUST, physicians typically did not order additional biochemical tests. Complete case analysis was applied in the multivariable modelling. A sensitivity analysis excluding serum albumin was conducted to assess the robustness of the findings, given its level of missingness.

Of the 50 candidate variables, those demonstrating at least a small effect size or informed by prior literature were selected for multivariable modelling. This approach ensured that clinically relevant variables were retained, even if they were not statistically significant in univariate analysis. Effect sizes were determined using Cohen’s d for continuous variables and Cramer’s V for categorical variables (Cohen, 1988; Cohen, 1992). Multivariable logistic regression was then performed to examine the association between the selected risk factors and community-acquired pressure injury. The final model was derived using the Backward Likelihood Ratio method, with variables sequentially removed based on statistical significance (p > .05) to achieve a parsimonious model while retaining clinically meaningful predictors. Model performance was assessed using the Hosmer–Lemeshow goodness-of-fit test (p = .801) and Nagelkerke R² (0.215).

Multicollinearity was assessed (all VIF < 2.5; condition index < 15), indicating no concerning intercorrelations among predictors, and no major deviations from linearity were observed for continuous variables. Adjusted odds ratios (ORs) with 95% confidence intervals (CIs) were reported, with statistical significance set at p < .05 (two-tailed). All analyses were conducted using IBM SPSS Statistics for Windows, Version 26.0 (IBM, 2017).

3. Results

3.1. Univariate analyses

A total of 52,480 patient records were extracted from the EMR system. Of these, 5772 were identified as cases with community-acquired pressure injury, and 46,708 were controls without pressure injury. The most common site of pressure injury was the sacrum/coccyx (51%; n = 1453) (Table 2).

Table 2.

PI characteristics in this cohort.

PI site		Counts	% of Total		Cumulative %
Sacral/Coccyx		1453	51.10%		51.10%
Buttock/ Gluteal		257	9.00%		60.20%
Trochanter		228	8.00%		68.20%
Foot/Toes/Heels		210	7.40%		75.60%
Shoulder/ Scapula		153	5.40%		81.00%
Elbow		104	3.70%		84.60%
Shin Area		80	2.80%		87.40%
Malleolus		74	2.60%		90.00%
Spine		69	2.40%		92.50%
Facial Area		42	1.50%		93.90%
Ear		37	1.30%		95.20%
Knee		37	1.30%		96.60%
Rib Cage Area		24	0.80%		97.40%
Lateral Aspect Wrist		19	0.70%		98.10%
Occiput/ Lateral Aspect of the Head		17	0.60%		98.70%
Penile/ Pubic Area		16	0.60%		99.20%
Thigh Area		13	0.50%		99.70%
Clavicle Area		9	0.30%		100.00%
Frequencies of the PI stage
PI stage	Counts			% of Total		Cumulative %
Stage 1	2472			42.80%		42.80%
Stage 2	2307			40.00%		82.80%
Stage 3	253			4.40%		87.20%
Stage 4	97			1.70%		88.90%
MDRPI	70			1.20%		90.10%
DTPI	417			7.20%		97.30%
Unstageable	156			2.70%		100.00%

Note: MDRPI: Medical-device related Pressure Injury; DTPI: Deep Tissue Pressure Injury.

Compared with controls, patients with community-acquired pressure injury were significantly older (72.7 ± 17.1 vs 63.3 ± 17.9 years; p < .001, d = 0.53) and more often unemployed (80.4% vs 61.2%; p < .001, V = 0.14). Functional limitations were also prominent: dependence in ambulation was observed in 24.1% of community-acquired pressure injury patients compared with 10.4% of controls (p < .001, V = 0.18), and reduced mobility showed a similar pattern (p < .001, V = 0.21). Marked differences were seen in microclimate and elimination factors; very moist skin (47.3% vs 1.3%) and diaper use (46.6% vs 15.3%) were substantially more frequent among community-acquired pressure injury cases (both p < .001; V = 0.17 and 0.25, respectively).

Nutritional and inflammatory vulnerability was evident, with a higher prevalence of poor/inadequate nutrition (19.5% vs 8.1%; p < .001, V = 0.13), lower mean serum albumin (27.6 ± 8.8 vs 29.3 ± 9.4 g/L; p < .001, r = 0.13), and higher C-reactive protein (67.7 ± 73.4 vs 50.6 ± 67.3 mg/L; p < .001, r = 0.12), while haemoglobin did not differ significantly (p = .93, d = –0.01). A higher proportion of community-acquired pressure injury patients were classified as high fall risk (21.7% vs 11.1%; p < .001, V = 0.12) and had chronic conditions such as hypertension (63.7% vs 48.0%; p < .001, V = 0.10), peptic ulcer (65.1% vs 44.6%; p < .001, V = 0.13), diabetes mellitus (67.7% vs 55.0%; p < .001, V = 0.08), prior myocardial infarction (41.2% vs 29.1%; p < .001, V = 0.08), and cancer (29.7% vs 17.7%; p < .001, V = 0.10). Variables reflecting downstream healthcare utilisation (e.g., length of stay, admitting discipline) were excluded from modelling as they represent outcomes rather than antecedent risk factors. Detailed univariate comparisons are presented in Table 3.

Table 3.

Univariable comparisons of demographic and admission-related variables between CAPI cases and controls (N = 52,480).

Variables	All (n=52,480)	Control No PI (n=46,708)	Cases PI (n=5772)	p-value	ES	ES interpret
Age, mean ± SD	63.37 ± 18.08	63.34 ± 17.93	72.70 ± 17.05	<0.001	d = 0.525	M
Length of stay (LOS), mean ± SD	6.42 ± 11.79	5.68 ± 9.80	12.33± 21.18	<0.001	r = 0.229	S
No. of admissions, mean ± SD	1.42 ± 0.95	1.38 ± 0.89	1.74 ± 1.30	<0.001	r = 0.160	S
Albumin (g/L), mean ± SD	27.86 ± 8.21	29.34 ± 9.36	27.63 ± 8.82	<0.001	r = 0.126	S
CRP (mg/L), mean ± SD	80.34 ± 92.62	50.64 ± 67.26	67.67 ± 73.36	<0.001	r = 0.118	S
Hb (g/dL), mean ± SD	8.22 ± 2.57	8.97 ± 2.94	8.94 ± 2.62	0.928	d=−0.008	NA
Sex, n (%)
Male	22,410 (42.7%)	20,285 (43.4%)	2125 (36.8%)	<0.001	V = 0.042	VS
Female	30,070 (57.3%)	26,423 (56.6%)	3647 (63.2%)
Working status, n (%)
Employed	3798 (35.9%)	3484 (38.8%)	314 (19.6%)	<0.001	V = 0.143	SM
Unemployed	6794 (64.1%)	5505 (61.2%)	1289 (80.4%)
Discipline, n (%)
Medical	29,593 (56.4%)	25,612 (54.8%)	3981 (69.0%)	<0.001	V = 0.089	S
Surgical	22,887 (43.6%)	21,096 (45.2%)	1791 (31.0%)
Repeated admission, n (%)
Yes	12,889 (24.6%)	10,716 (22.9%)	2173 (37.6%)	<0.001	V = 0.107	S
No	39,591 (75.4%)	35,992 (77.1%)	3599 (62.4%)
Condition on arrival, n (%)
Alert & Conscious	49,214 (97.6%)	44,018 (98.1%)	5196 (92.1%)	<0.001	V = 0.130	S
Confused & agitated	354 (0.7%)	249 (0.6%)	105 (1.9%)
Drowsy & Uncommunicative	840 (1.6%)	502 (1.1%)	338 (6.0%)
Level of consciousness on arrival, n (%)
Conscious	47,641 (95.2%)	42,785 (96.4%)	4856 (86.1%)	<0.001	V = 0.154	S
Drowsy & Uncommunicative	1683 (3.4%)	1111 (2.5%)	572 (10.1%)
Restless & confused	710 (1.4%)	495 (1.1%)	215 (3.8%)
Behaviour on admission, n (%)
Cooperative	44,419 (97.9)	39,579 (98.3%)	4840 (94.7%)	<0.001	V = 0.079	VS
Restless	299 (0.7%)	220 (0.5%)	79 (1.5%)
Withdrawn	53 (0.1%)	35 (0.1%)	18 (0.4%)
Uncooperative	623 (1.4%)	447 (1.1%)	176 (3.4%)
Mobility status, n (%)
Independent	25,448 (50.2%)	22,253 (49.4%)	3195 (56.1%)	<0.001	V = 0.208	M
Requires assistance	2276 (4.5%)	1479 (3.3%)	797 (14.0%)
Chairbound	22,640 (44.6%)	21,100 (46.9%)	1540 (27.1%)
Bedbound	346 (0.7%)	187 (0.4%)	159 (2.8%)
Ambulation status, n (%)
Independent	32,932 (64.9%)	30,550 (67.8%)	2382 (41.3%)	<0.001	V = 0.180	M
Requires assistance	11,735 (23.1%)	9797 (21.7%)	1938 (34.0%)
Dependent	6076 (12.0%)	4703 (10.4%)	1373 (24.1%)
Use of walking aid, n (%)
Yes	9124 (48.1%)	7571 (46.2%)	1553 (60.9%)	<0.001	V = 0.101	S
No	9830 (51.9%)	8833 (53.8%)	997 (39.1%)
Physiotherapy follow-up, n (%)
Yes-On-going	24,385 (96.5%)	20,660 (96.2%)	3725 (98.2%)	<0.001	V = 0.039	VS
No-Previously on PT	896 (3.5%)	826 (3.8%)	70 (1.8%)
History of smoking, n (%)
Never smoked	41,655 (79.4%)	37,332 (79.9%)	4323 (74.9%)	<0.001	V=−0.027	VS
Ever smoked	5586 (10.6%)	5148 (11.0%)	438 (7.6%)
History of drinking alcohol, n (%)
Never drank	56 (4.3%)	53 (4.8%)	3 (1.6%)	.048	V = 0.055	NA
Ever drank	1238 (95.7%)	1054 (95.2%)	184 (98.4%)
Nutrition status, n (%)
Poor - Inadequate	3905 (9.5%)	2920 (8.1%)	985 (19.5%)	<0.001	V = 0.128	SM
Excellent- Adequate	37,265 (90.5%)	33,204 (91.9%)	4061 (80.5%)
Recent weight loss, n (%)
Yes	1426 (2.7%)	1185 (2.5%)	241 (4.2%)	<0.001	V = 0.030
No	47,431 (90.4%)	42,105 (90.1%)	5326 (92.3%)
BMI, n (%)
Underweight	3841 (9.2%)	3006 (14.5%)	835 (17.1%)	<0.001	V = 0.106	S
Normal	19,094 (45.6%)	16,816 (39.5%)	2278 (46.5%)
Overweight	11,896 (28.4%)	10,790 (29.2%)	1106 (22.6%)
Obese	7051 (16.8%)	6373 (17.2%)	678 (13.8%)
Skin Moisture, n (%)
Rarely moist skin	32,418 (63.9%)	29,929 (66.5%)	2489 (43.7%)	<0.001	V = 0.171	M
Occasionally moist skin	17,376 (34.3%)	14,509 (32.2%)	2867 (50.4%)
Very moist skin	920 (1.8%)	584 (1.3%)	336 (5.9%)
Bowel status, n (%)
Normal	48,477 (92.4%)	43,054 (92.2%)	5423 (94.0%)	<0.001	V = 0.021	VS
Constipation	1130 (2.2%)	955 (2.0%)	175 (3.0%)
Diarrhoea	417 (0.8%)	362 (0.8%)	55 (1.0%)
Urinary status, n (%)
Self	33,012 (62.9%)	30,749 (62.9%)	2263 (39.2%)	<0.001	V = 0.254	SM
Catheter	6325 (12.1%)	5703 (12.2%)	622 (10.8%)
Diaper	9854 (18.8%)	7163 (15.3%)	2691 (46.6%)
Fall risk, n (%)
Low	18,101 (34.5%)	16,805 (36.0%)	1296 (22.5%)	<0.001	V = 0.123	SM
Moderate	25,810 (49.2%)	22,724 (48.7%)	3086 (53.5%)
High	6415 (12.2%)	5162 (11.1%)	1253 (21.7%)
Frequent fall, n (%)
Yes	2222 (4.7%)	1839 (4.4%)	383 (7.0%)	<0.001	V = 0.041	VS
No	45,472 (95.3%)	40,419 (95.6%)	5053 (93.0%)
Fall history, n (%)
Yes	7390 (14.7%)	6157 (13.8%)	1233 (21.9%)	<0.001	V = 0.072	S
No	42,916 (85.3%)	38,516 (86.2%)	4400 (78.1%)
Medication, n (%)
Anti-infectives	29,391 (56.0%)	25,324 (54.2%)	4067 (70.5%)	<0.001	V = 0.102	S
Analgesic	26,873 (51.2%)	23,770 (50.9%)	3103 (53.8%)	<0.001	V = 0.018	VS
Immunosuppressants	1478 (2.8%)	1217 (2.6%)	261 (4.5%)	<0.001	V = 0.036	VS
Comorbidities, n (%)
Cancer	10,003 (19.1%)	8289 (17.7%)	1714 (29.7%)	<0.001	V = 0.095	S
Asthma	9497 (18.1%)	8173 (17.5%)	1324 (22.9%)	<0.001	V = 0.044	VS
Hypertension	26,112 (49.8%)	22,436 (48.0%)	3676 (63.7%)	<0.001	V = 0.098	S
Diabetes Mellitus	29,611 (56.4%)	25,702 (55.0%)	3909 (67.7%)	<0.001	V = 0.080	S
Hyperlipidaemia	11,984 (22.8%)	10,533 (22.6%)	1451 (25.1%)	<0.001	V = 0.019	VS
Chronic Kidney Disease	4580 (8.7%)	3777 (8.1%)	803 (13.9%)	<0.001	V = 0.065	VS
History of Urinary Incontinence	2482 (4.7%)	2277 (4.9%)	205 (3.6%)	<0.001	V = 0.020	VS
History of Peptic Ulcer	24,591 (46.9%)	20,835 (44.6%)	3756 (65.1%)	<0.001	V = 0.128	SM
History of Acute Myocardial Infarction	15,963 (30.4%)	13,583 (29.1%)	2380 (41.2%)	<0.001	V = 0.083	S
History of Stroke	3581 (6.8%)	3095 (6.6%)	486 (8.4%)	<0.001	V = 0.022	VS
History of Vascular Disease	1114 (2.1%)	959 (2.1%)	155 (2.7%)	.002	V = 0.014	VS
History of Atrial Fibrillation	2443 (4.7%)	2007 (4.3%)	436 (7.6%)	<0.001	V = 0.048	VS

Note: ES = effect size. For t-tests we report Cohen’s d; for Mann–Whitney U we report effect-size r (Z/√N); for χ² we report Cramér’s V. ES interpretation based on Cohen (1988): VS = very small, SM = small to moderate, M = moderate, L = large. p < .05 considered significant.

Additional analyses examining CAPI severity showed that patients with more severe injuries were more likely to have poorer ambulation status and poorer nutritional status. The proportion of severe CAPI increased from 7.7% among patients who were independent in ambulation to 17.0% among those requiring assistance and 30.7% among those who were dependent (p < .001). Similarly, severe CAPI was more common among patients with poor/inadequate nutrition compared with those with adequate/excellent nutrition (30.2%vs 14.5%, p < .001) (Supplementary Table B).

3.2. Multivariable analyses

In the final Backward stepwise logistic regression, independent predictors of community-acquired pressure injury included unemployment (adjusted OR [aOR] 1.50, 95% CI 1.14–1.98), poor/inadequate nutrition (aOR 1.92, 95% CI 1.44–2.55), lower serum albumin (per 1 g/L: aOR 0.95, 95% CI 0.94–0.97), functional dependence (vs independent: requires assistance aOR 2.25, 95% CI 1.92–2.64; dependent aOR 2.87, 95% CI 2.38–3.47), moist skin (vs rarely moist: occasionally moist aOR 1.31, 95% CI 1.06–1.62; very moist aOR 3.02, 95% CI 1.65–5.53), hypertension (aOR 1.92, 95% CI 1.44–2.56), and history of peptic ulcer (aOR 2.13, 95% CI 1.69–2.67). The model fit was acceptable (Hosmer–Lemeshow p = .801). Full estimates are provided in Table 4.

Table 4.

Unadjusted and adjusted odds ratios (OR) for predictors of CAPI from multivariable logistic regression (adjusted for age and sex).

Predictor (Reference1)	Unadjusted OR (95% CI)	p-value	Adjusted OR (95% CI)	p-value
Age (per year ↑)	1.03 (1.03–1.04)	<0.001	– (ns, removed)	–
Sex (Female1 vs Male)	0.76 (0.72–0.80)	<0.001	– (ns, removed)	–
Working status		<0.001		<0.001
(Unemployed vs Employed1)	2.60 (2.28–2.96)	<0.001	1.50 (1.14–1.98)	0.004
Nutrition (Poor vs Adequate/Good1)	2.76 (2.55–2.99)	<0.001	1.92 (1.44–2.55)	<0.001
Albumin (per 1 g/L ↑)	0.98 (0.98–0.99)	<0.001	0.95 (0.94–0.97)	<0.001
Ambulation		<0.001		<0.001
Requires assistance vs Independent1	2.54 (2.38–2.71)	<0.001	2.25 (1.92–2.64)	<0.001
Dependent vs Independent	3.74 (3.48–4.03)	<0.001	2.87 (2.38–3.47)	<0.001
Skin moisture		<0.001		<0.001
Occasionally moist vs Rarely moist1	2.38 (2.24–2.52)	<0.001	1.31 (1.06–1.62)	0.013
Very moist vs Rarely moist1	6.92 (6.01–7.96)	<0.001	3.02 (1.65–5.53)	<0.001
Hypertension (Yes vs No1)	1.90 (1.79–2.01)	<0.001	1.92 (1.44–2.56)	<0.001
History of Peptic Ulcer (Yes vs No1)	2.31 (2.19–2.45)	<0.001	2.13 (1.69–2.67)	<0.001

Note. OR = odds ratio; CI = confidence interval; ns = non-significant. Predictors in the unadjusted model were controlled for age and sex. Statistical significance was set at p < .05. Reference categories ¹: Female; Employed; Independent ambulation; Rarely moist skin; Adequate/Good nutrition; No hypertension; No peptic ulcer.

Sensitivity analysis excluding serum albumin yielded similar findings, with key predictors including working status, nutritional status, ambulation status, skin moisture, hypertension, and history of peptic ulcer remaining statistically significant.

3.3. Interaction analyses

Interaction analyses were conducted after confirming that model assumptions, including the absence of multicollinearity, were met. Three two-way interaction terms (Age × Albumin, Albumin × Nutrition, and Albumin × Sex) were tested by introducing multiplicative terms into the multivariable logistic regression models to examine potential effect modification. These interactions were selected as age, nutritional status, and sex are known to influence serum albumin concentrations and the underlying nutritional–inflammatory balance. Albumin levels decline with age due to reduced dietary intake, chronic disease, and frailty (Don and Kaysen, 2004), while older adults are at increased risk of malnutrition through sarcopenia and multimorbidity (Norman et al., 2021). Sex-related differences in body composition and inflammatory profiles also contribute to variation in albumin levels (Weaving et al., 2016).

The Age × Albumin and Albumin × Nutrition interactions were not statistically significant, indicating that the association between serum albumin and community-acquired pressure injury was consistent across age and nutritional strata (Table 5, Table 6, Table 7). In contrast, a statistically significant Albumin × Sex interaction was observed (interaction OR = 1.03, 95% CI 1.00–1.05, p = .034) (Table 7). This suggests that the association between serum albumin and pressure injury risk differs slightly by sex, with a modest variation in the strength of the protective effect.

Table 5.

Logistic regression predicting CAPI with age × albumin interaction.

Predictor (reference)	OR	95% CI		P-value
		Lower	Upper
Sex (Male)	1.05	0.82	1.36	0.694
Working Status (Unemployed)	1.50	1.14	1.98	0.004
Nutrition (Poor/Inadequate)	1.92	1.44	2.55	<0.001
Skin Moisture: Rarely				<0.001
Skin Moisture: Occasionally	1.31	1.06	1.62	0.014
Skin Moisture: Very	3.02	1.65	5.53	<0.001
Hypertension	1.92	1.44	2.56	<0.001
Peptic Ulcer History	2.13	1.69	2.67	<0.001
Age Centred	1.01	1.00	1.02	0.003
Albumin Centred	0.95	0.94	0.97	<0.001
Age x Albumin interaction	1.00	1.00	1.00	0.197

Note: OR = Odds Ratio. CI = Confidence Interval. Bolded interaction term indicates main effect of interest. Significant predictors at p < .05.

Table 6.

Logistic regression predicting CAPI with albumin × nutrition interaction.

Predictors (reference)	OR	95% CI		p-value
		Lower	Upper
Sex (Male)	1.07	0.83	1.38	0.603
Age	1.01	1.00	1.02	0.004
Working Status (Unemployed)	1.50	1.14	1.97	0.004
Skin Moisture: Rarely				<0.001
Skin Moisture: Occasionally	1.31	1.06	1.62	0.013
Skin Moisture: Very	3.02	1.66	5.52	<0.001
Nutrition (Poor/Inadequate)	1.86	1.40	2.48	<0.001
Hypertension	1.91	1.43	2.55	<0.001
Peptic Ulcer History	2.14	1.70	2.68	<0.001
Albumin Centred	0.95	0.94	0.97	<0.001
Albumin x Nutrition interaction	1.03	0.99	1.06	0.139

Note: OR = Odds Ratio. CI = Confidence Interval. Bolded interaction term indicates main effect of interest. Significant predictors at p < .05.

Table 7.

Logistic regression predicting CAPI with albumin × sex interaction.

Predictors (reference)	OR	95% CI		p-value
		Lower	Upper
Age	1.01	1.00	1.02	0.010
Sex(Male)	1.07	0.83	1.37	0.603
Working Status (Unemployed)	1.51	1.14	1.98	0.004
Skin Moisture: Rarely				<0.001
Skin Moisture: Occasionally	1.31	1.06	1.63	0.012
Skin Moisture: Very	2.95	1.62	5.39	<0.001
Nutrition (Poor/Inadequate)	1.93	1.45	2.56	<0.001
Peptic Ulcer History	2.10	1.67	2.63	<0.001
Hypertension	1.95	1.46	2.59	<0.001
Albumin Centred	0.92	0.88	0.95	<0.001
Albumin x Sex interaction	1.03	1.00	1.05	0.034

Note: OR = Odds Ratio. CI = Confidence Interval. Bolded interaction term indicates main effect of interest. Significant predictors at p < .05.

4. Discussion

This study examined the clinical, functional, nutritional, and sociodemographic determinants of community-acquired pressure injuries using a large electronic health record dataset. Multiple domains influenced the risk of community-acquired pressure injuries, underscoring its multifactorial pathogenesis. Independent predictors included unemployment, poor or inadequate nutrition, hypoalbuminemia, impaired mobility, hypertension, peptic ulcer disease, and excessive skin moisture. Interaction analyses also revealed a modest sex-based difference in the association between albumin and pressure injury risk.

Patients with comorbid conditions such as hypertension and peptic ulcer may require closer monitoring, as these may reflect underlying physiological vulnerability. These findings highlight the importance of a multifaceted approach to CAPI prevention that addresses nutritional, functional, and skin-related factors. The following sections discuss these findings in relation to existing evidence and their implications for community-based prevention and care.

4.1. Sociodemographic factors

Age, sex, and employment status were significant correlates in univariate analysis, but only unemployment remained independently predictive after adjustment (OR = 1.52, 95% CI [1.05–2.20]). The initial effects of age and sex appeared mediated through clinical and functional vulnerabilities such as immobility, malnutrition, and multimorbidity.

The relationship between age and PI risk is well established; ageing skin exhibits reduced elasticity, perfusion, and reparative capacity (Coleman et al., 2013; Kirkland-Khyn et al., 2019). Similar findings from community studies confirm that frail, multimorbid older adults account for most community-acquired pressure injury cases (Corbett et al., 2017; Ding et al., 2022). The age effect diminished after adjustment, similar to prior studies (Chung et al., 2023), suggesting that chronological age is a proxy for accumulated physiological deficits rather than an independent risk factor.

Sex differences were inconsistent. Although females formed a slightly larger proportion of community-acquired pressure injury cases, sex did not independently predict risk after adjustment. Previous reports diverge; some cite a higher female risk due to lower lean mass and longevity (Jaul et al., 2018; Johansen et al., 2015), while others attribute increased male risk to differential fat distribution, occupational exposure, and care patterns (Worsley et al., 2016; Lichterfeld-Kottner et al., 2020). These mixed findings indicate that sex modifies, rather than determines, community-acquired pressure injury risk through physiological and contextual mechanisms.

Unemployment, by contrast, remained a robust independent predictor, highlighting how socioeconomic disadvantage increases vulnerability. Financial strain, limited caregiver support, and reduced access to dietetic or equipment resources heighten pressure injury risk (Singh and Shoqirat, 2021; Latimer et al., 2019). In this study, unemployment likely reflected lower socioeconomic position and poorer access to preventive care. Integrating social determinants of health into community-acquired pressure injury screening is therefore crucial.

Overall, while age and sex represent biological correlates, socioeconomic disadvantage is a modifiable factor. Effective prevention should combine clinical screening (mobility, nutrition) with social support such as subsidised equipment, caregiver training, and early community nursing engagement.

4.2. Nutritional factors

Nutrition emerged as a key determinant of community-acquired pressure injury. Poor or inadequate nutritional status was associated with an approximately twofold increase in risk, while each 1 g/L increase in serum albumin was associated with a 5% reduction in risk. These findings align with evidence that inadequate protein and micronutrient intake impairs collagen synthesis, perfusion, and wound repair (Chen et al., 2023; Langer et al., 2024). Hypoalbuminemia, which was common among affected patients, reflects both nutritional and inflammatory depletion (Li et al., 2025) and has been associated with pressure injury progression (Alderden et al., 2018; Elsorady and Nouh, 2023). Serum albumin may therefore function as a combined marker of nutritional reserve and systemic inflammation, with lower levels indicating reduced protein availability and impaired tissue integrity. Elevated C-reactive protein observed among patients with community-acquired pressure injury further supports the presence of an underlying inflammatory-catabolic state contributing to albumin decline.

A statistically significant interaction between serum albumin and sex was observed, indicating a modest sex-based difference in the association between albumin and pressure injury risk. While the underlying mechanisms are not fully established, this may relate to sex-related differences in body composition and protein reserves, with males generally having greater muscle mass and protein stores (Weaving et al., 2016). In addition, variations in inflammatory and metabolic responses between males and females may influence how serum albumin reflects underlying physiological status (Kahlert et al., 2017). However, evidence examining sex-specific effects of serum albumin in pressure injury risk remains limited, and this finding should be interpreted cautiously.

These findings highlight nutritional vulnerability as a modifiable target for prevention. Routine screening using tools such as the MUST or Mini Nutritional Assessment–Short Form (MNA-SF), coupled with early dietetic referral and caregiver education, may support timely identification of at-risk individuals. Monitoring nutritional status, including serum albumin trends, may further aid in early risk stratification, particularly among patients with multiple comorbidities.

4.3. Functional factors

Functional dependence was one of the strongest predictors of community-acquired pressure injury. Patients requiring assistance or fully dependent had two- to threefold higher odds than independent individuals. This confirms that mobility limitation remains a core risk factor for pressure injury across settings (Jaul et al., 2018; Coleman et al., 2013). Limited movement prolongs tissue loading and shear, leading to ischemia and cellular deformation (Gefen et al., 2022).

Functional decline often coexists with sarcopenia, malnutrition, and multimorbidity, each compounding vulnerability to community-acquired pressure injury (Faxén-Irving et al., 2021). Thus, immobility should be viewed not in isolation but as part of an interdependent physiological triad encompassing muscle loss, poor perfusion, and systemic inflammation. Physiotherapy follow-up and walking-aid use were not independently significant, indicating that the degree of mobility, not therapy access, is the critical determinant.

From a prevention standpoint, early physiotherapy referral, assistive device optimisation, and caregiver training in safe repositioning should be prioritised. Embedding mobility screening into community and transitional care workflows could identify at-risk individuals early, before skin breakdown occurs. Preserving mobility should therefore be viewed as a central strategy in community-acquired pressure injury prevention alongside nutritional and moisture management.

4.4. Comorbidities & medication-related factors

Among chronic conditions, hypertension and peptic ulcer disease remained independent predictors after adjustment, while diabetes, stroke, and chronic kidney disease did not. Chronic hypertension may impair tissue perfusion through endothelial dysfunction and arterial stiffness (Lv et al., 2025; De Azevedo Macena et al., 2017; Kim, 2023). Although some systematic reviews report inconsistent associations (Huang et al., 2024), this finding may reflect differences in study populations and care settings. Most prior studies have focused on hospital-acquired pressure injuries, whereas the present study examined community-acquired cases. In this context, hypertension may contribute to impaired microcirculation prior to hospitalisation, increasing susceptibility to tissue breakdown in already vulnerable individuals.

Peptic ulcer disease may contribute through systemic inflammation, nutrient malabsorption, and frequent NSAID use, all of which can impair tissue repair (Frykberg and Banks, 2015). Although analgesic and anti-infective use were more common among patients with community-acquired pressure injury, these variables were excluded from modelling due to confounding and collinearity with inflammatory markers. C-reactive protein was examined descriptively as a marker of systemic inflammation but was not included in the multivariable model due to substantial missing data (Dong and Peng, 2013).

These findings highlight the interplay between vascular, metabolic, and inflammatory stressors in the pathogenesis of community-acquired pressure injury. Integrating cardiovascular, gastrointestinal, and nutritional considerations into community nursing assessments may support earlier identification of high-risk individuals (El-Tallawy et al., 2021).

Findings from this study are broadly consistent with evidence from Asian settings. A home care study in Japan reported that malnutrition was a strong predictor of both pressure injury development and severity (Iizaka et al., 2010). Similarly, a large multicentre study in China identified low serum albumin, immobility, and incontinence as key factors associated with increased pressure injury risk among hospitalised patients (Liu et al., 2019). The consistency of these findings across both community and hospital contexts suggests that nutritional vulnerability and functional dependence are important determinants of pressure injury risk in Asian populations.

In contrast, the identification of hypertension and peptic ulcer disease as independent predictors differs from some existing evidence. This discrepancy may be explained by differences in care context, as prior studies have largely focused on hospital-acquired pressure injuries, whereas the present study examined community-acquired cases.

4.5. Microclimate & elimination-related factors

Microclimate regulation and elimination function were prominent extrinsic predictors of community-acquired pressure injuries. Patients with very moist skin were over three times more likely to develop community-acquired pressure injuries, and diaper use independently increased risk. Excessive moisture compromises the skin barrier, increasing friction, maceration, and shear susceptibility (Kottner et al., 2018; Faust et al., 2022; Mervis and Phillips, 2019).

Diaper use and incontinence likely amplify risk by sustaining skin overhydration, trapping heat, and exposing tissue to irritants such as ammonia (Bostan et al., 2019). The coexistence of incontinence-associated dermatitis and pressure injuries, mediated by friction and moisture rather than pressure alone, has been similarly reported by Beeckman et al. (2014). Importantly, the persistence of both skin moisture and diaper use as independent predictors even after adjusting for mobility suggests that microclimate management is a distinct preventive domain (Beeckman et al., 2014).

Clinically, proactive moisture control, including frequent product changes, breathable materials, pH-balanced cleansers, and barrier creams, is essential. Environmental measures such as adequate ventilation and humidity regulation further help maintain skin integrity. These findings reinforce the importance of maintaining an optimal skin microclimate for community-acquired pressure injury prevention and underscore the need to integrate this into community care protocols alongside functional and nutritional interventions.

From a clinical perspective, these findings highlight key priorities for targeted intervention. Nutritional optimisation, mobility support, and effective moisture management represent the most actionable areas for prevention, given their strong associations with community-acquired pressure injury and feasibility in routine care. Integrating these domains into community and transitional care pathways may support earlier identification of at-risk individuals and reduce preventable pressure injuries.

5. Limitations

This study has several limitations. The retrospective case–control design relied on routinely collected electronic health records, which limited control over data completeness and accuracy. Although we triangulated information across nursing, medical, physiotherapy, and dietetics records, some diagnoses (e.g., liver disease, dementia) were inconsistently documented and could not be reliably extracted.

Laboratory data such as Hb, were missing for a substantial proportion of patients, preventing a full assessment of anaemia as a potential risk factor. Missingness in laboratory variables was likely clinically driven rather than random, potentially introducing bias despite the use of complete-case analysis. Variables with >25% missing data were excluded, potentially introducing residual confounding.

Pressure injury identification was based on electronic medical record documentation and may be subject to diagnostic variability or misclassification. Caregiver-related and environmental factors, such as caregiving capacity, caregiver knowledge, and home care conditions, were not captured because they are not routinely recorded in the electronic medical record. Only employment status was available as a proxy for socioeconomic factors. As such, the influence of social and caregiving contexts on community-acquired pressure injury risk could not be fully explored.

The retrospective design also prevents the establishment of causal relationships. Although multivariable models accounted for known confounders, residual confounding from unmeasured variables cannot be ruled out. Misclassification bias may have occurred because the identification of community-acquired pressure injury within 24 h depended on the accuracy of clinical documentation. In addition, we could not identify patients who developed pressure injuries as part of “skin failure” in the dying process (Berlowitz and Levine, 2025); this may have introduced unmeasured bias, even though patients who died during hospitalisation were excluded.

Finally, the study’s single-centre design limits generalisability. The findings may not fully represent patterns in primary care, long-term care, or home-based settings where many community-onset wounds originate. The EMR data also lacked temporal details, such as the timing or stage progression of community-acquired pressure injury before admission. Moreover, the dataset did not capture caregiver practices, home care environments, or community service use, factors known to influence wound development.

Future research should build on these findings through multi-centre prospective studies to validate the observed interaction between serum albumin and sex, and to better establish temporal relationships between risk factors and community-acquired pressure injuries. Incorporating caregiver-related and home-environment variables into future studies may help develop more comprehensive risk prediction models. In addition, intervention studies are needed to evaluate the effectiveness of community-based prevention strategies targeting modifiable risk factors such as nutrition, mobility, and skin moisture management.

6. Conclusions

This large single-centre study identified independent associations between community-acquired pressure injury on admission and several modifiable risk factors, including socioeconomic disadvantage (unemployment), functional dependence, nutritional vulnerability (poor or inadequate nutrition and lower albumin), impaired microclimate control (moist skin, diaper use), and chronic disease (hypertension and peptic ulcer). These findings suggest that community-acquired pressure injury reflects the combined effects of social, functional, metabolic, and vascular vulnerabilities rather than a purely localised skin condition.

From a clinical perspective, prevention strategies should consider several key priorities. These include embedding brief functional screening and targeted mobility support; implementing systematic nutritional screening (e.g., MUST or MNA-SF) with early dietetic referral and albumin monitoring; standardising moisture and incontinence management using appropriate products and skin protection; and integrating socioeconomic assessment to facilitate access to equipment, caregiver support, and community nursing follow-up. The observed Albumin × Sex interaction suggests that sex may be considered when interpreting albumin levels, although this finding requires further validation.

This study contributes region-specific evidence to the global understanding of pressure injury prevention. While most large-scale data originates from North America, Europe, and Australia, evidence from Asian populations remains limited despite rapid population ageing. Findings from Singapore’s integrated acute–community care context highlight the importance of considering both biological and contextual factors in understanding the risk of community-acquired pressure injury.

Future research should prospectively validate these predictors in multi-centre Asian cohorts, further explore the clinical relevance of the Albumin × Sex interaction, and evaluate community-based preventive interventions targeting mobility, nutrition, and microclimate within the broader context of social determinants of health.

Ethical approval

Approved by the SingHealth Centralised Institutional Review Board (Reference No 2022/2406). Informed consent was waived for this retrospective analysis of de-identified data.

Declaration of generative AI and AI-assisted technologies in the manuscript preparation process

During the preparation of this work, the author(s) used AI-assisted tools (e.g., Grammarly) to improve the clarity, readability, and language of the manuscript. The authors reviewed and edited all content generated by these tools and take full responsibility for the content of the published article.

Funding

This study received funding from the Wound Care Innovation for the Tropics Programme (Project number H17/01/a0/0BB9), an industry alignment grant under A*STAR and the Skin Research Institute of Singapore, Biomedical Sciences Institutes, Singapore.

CRediT authorship contribution statement

Fazila Aloweni: Writing – original draft, Project administration, Formal analysis, Data curation, Methodology, Conceptualization. Kee Chen Elaine Siow: Writing – review & editing, Supervision, Conceptualization. Nanthakumahrie Gunasegaran: Writing – review & editing. Gek Hsiang Lim: Writing – review & editing, Methodology, Formal analysis. Shin Yuh Ang: Supervision, Writing – review & editing. Truls Østbye: Writing – review & editing, Supervision, Methodology, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

The data underlying this study is not publicly available.