Abstract
Background
Hospital malnutrition is associated with adverse clinical outcomes, yet real-world evidence distinguishing nutritional risk screening from diagnostic assessment remains limited. This study evaluated a stepwise NRS2002-based screening and assessment pathway and examined associations between pathway-defined malnutrition and clinical outcomes in hospitalized patients.
Methods
This retrospective cohort study was conducted at Xishan People’s Hospital from January 2022 to June 2025. Adult patients admitted for > = 24 h who completed Nutritional Risk Screening 2002 (NRS2002) within 48 h were included. NRS2002 was the universal screening tool. Patients with NRS2002 > = 3 underwent diagnostic assessment according to a predefined decision tree: oncology patients received Patient-Generated Subjective Global Assessment (PG-SGA), non-oncology patients with BMI < 18.5 kg/m2 received Global Leadership Initiative on Malnutrition (GLIM) evaluation, and remaining at-risk patients received Subjective Global Assessment (SGA). Primary outcomes included hospital-acquired infection (HAI), 30-day readmission among discharged-alive patients, and in-hospital mortality.
Results
Among 2,856 patients (mean age 59.23 +/- 15.47 years; 52.17% male), 1,187 (41.56%) were identified as at nutritional risk by NRS2002. Three pathway groups were identified: NRS2002 < 3 (n = 1,669), NRS2002 > = 3 but not malnourished (n = 335), and malnourished (n = 852). The pathway-defined malnutrition proportion was 29.83% in the total cohort and 71.78% among NRS-positive patients. Malnourished patients had higher rates of HAI (14.67% vs. 5.29%; P < 0.001), 30-day readmission among discharged-alive patients (20.84% vs. 11.18%; P < 0.001), and in-hospital mortality (8.22% vs. 2.30%; P < 0.001). In multivariable models, malnutrition was associated with HAI (aOR = 2.19; 95% CI: 1.57–3.07), 30-day readmission among discharged-alive patients (aOR = 1.81; 95% CI: 1.38–2.37), and in-hospital mortality (aOR = 2.54; 95% CI: 1.59–4.04).
Conclusions
In this single-center retrospective cohort, malnutrition identified through a stepwise NRS2002-based assessment pathway was common and was associated with higher HAI, readmission, mortality, and length of stay. HIS-integrated nutrition software may support standardized documentation and workflow efficiency, but the present observational study did not directly evaluate software effectiveness or interventional care pathways.
Keywords: Malnutrition, NRS2002, SGA, PG-SGA, GLIM criteria, Hospital-acquired infection, Readmission, Mortality
Introduction
Malnutrition in hospitalized patients represents a significant yet frequently underrecognized clinical challenge, affecting approximately 20% to 50% of inpatients worldwide depending on population, setting, and diagnostic approach [1–3]. In China, reported nutritional risk rates vary across hospitals and patient groups [4–6]. Hospital malnutrition and nutritional risk are consistently associated with prolonged hospitalization, complications, readmission, higher costs, and mortality [7–9].
These adverse outcomes highlight the importance of early identification, but they also require careful terminology. Nutritional risk screening identifies patients who require further assessment, whereas malnutrition diagnosis requires a formal assessment process and severity grading. Treating screening positivity as equivalent to diagnostic malnutrition can overstate prevalence and obscure the clinical pathway.
The accurate identification of malnutrition therefore requires a two-step approach: first, validated nutritional risk screening; second, diagnostic assessment and severity grading among those at risk. NRS2002 is recommended by ESPEN and has been widely used for initial inpatient screening [10–12].
For patients identified as at nutritional risk, population-appropriate diagnostic assessment tools are needed. SGA is widely used in general medical and surgical populations [13]. PG-SGA is particularly suited to oncology because it captures cancer-specific symptoms and patient-reported intake barriers [14–16]. GLIM criteria provide a diagnostic framework requiring at least one phenotypic criterion and at least one etiologic criterion, and recent updates emphasize its role as a standardized malnutrition diagnosis framework rather than a screening tool [17–19].
Despite the availability of validated tools, implementation of systematic nutritional screening and assessment in routine care remains challenging. Integration of clinical nutrition software with the hospital information system (HIS) can standardize data capture by automatically retrieving demographics, laboratory results, diagnoses, and outcomes while storing structured screening and assessment records. In the present study, this integration was used as the data capture and workflow platform; it was not evaluated as an independent intervention or care pathway.
This study aimed to evaluate nutritional risk and pathway-defined malnutrition using an NRS2002-based stepwise screening-assessment approach integrated with electronic records, and to examine associations between malnutrition status and clinical outcomes including HAI, 30-day readmission, in-hospital mortality, and LOS.
Materials and methods
Study design and setting
This retrospective cohort study was conducted at Xishan People’s Hospital, a tertiary care facility with 924 beds in Wuxi, Jiangsu Province, China. The study period extended from January 1, 2022, to June 30, 2025. The study protocol was approved by the Institutional Review Board of Xishan People’s Hospital (Approval No.: xs2024js048). Due to the retrospective nature of the study and the use of de-identified data, the requirement for informed consent was waived.
Data source and integration
The hospital employs a clinical nutrition software system (Jianxi Cloud Smart Nutrition System V5.3.14) interfaced with the HIS. The interface automatically retrieves patient demographics, laboratory parameters, admission diagnoses, discharge status, and selected clinical outcomes. Nutritional screening results and diagnostic assessment data are entered by trained staff in structured fields and synchronized with the electronic medical record. The present study used this integrated database for retrospective data extraction; it did not compare outcomes before versus after software implementation.
Study population
The inclusion criteria were: (1) age > = 18 years; (2) hospital admission duration > = 24 h; and (3) completion of NRS2002 screening within 48 h of admission. The exclusion criteria were: (1) pregnancy; (2) admission for palliative or hospice care; (3) psychiatric admission precluding reliable assessment; (4) transfer from another acute care facility with admission > 7 days; (5) missing data for primary outcomes (> 20% of variables); and (6) repeated planned admissions for the same treatment cycle when identifiable, for which only the eligible index hospitalization was retained.
Nutritional screening and assessment protocol
The nutritional care pathway followed a standardized stepwise approach:
Step 1
Initial screening. NRS2002 served as the universal nutritional risk screening tool and was completed within 48 h of admission by trained nursing staff. NRS2002 incorporates disease severity, nutritional status (BMI, recent weight loss, and reduced food intake), and age > = 70 years. Recent weight loss and reduced intake were recorded from patient or caregiver report and cross-checked against available medical records, documented body weight, and dietitian assessment notes when present. Patients with NRS2002 scores > = 3 were classified as at nutritional risk [20].
Step 2: Comprehensive diagnostic assessment. Patients identified as at nutritional risk underwent comprehensive assessment by registered dietitians within 72 h. The diagnostic tool was selected according to a predefined hierarchy: oncology patients received PG-SGA; among non-oncology patients, those with BMI < 18.5 kg/m2 received GLIM evaluation; remaining at-risk patients received SGA. This hierarchy was used to avoid applying multiple diagnostic tools to the same hospitalization.
SGA categorized non-oncology general medical and surgical patients as well-nourished (SGA-A), moderately malnourished (SGA-B), or severely malnourished (SGA-C) [21].
PG-SGA was used for oncology patients because it incorporates cancer-specific symptoms, intake barriers, functional status, and patient-reported components; categorical ratings A, B, and C were mapped to well-nourished, moderate malnutrition, and severe malnutrition, respectively [22].
GLIM criteria were applied to non-oncology patients with BMI < 18.5 kg/m2. GLIM diagnosis required at least one phenotypic criterion (unintentional weight loss, low BMI, or reduced muscle mass) and at least one etiologic criterion (reduced food intake or assimilation, or disease burden/inflammation) [23]. Reduced muscle mass was assessed from dietitian physical examination and available clinical records rather than routine body-composition imaging. Disease burden/inflammation included active infection, cancer, recent major surgery or trauma, critical illness, organ failure, and other documented acute or chronic inflammatory conditions.
All assessors received standardized training on NRS2002, SGA, PG-SGA, and GLIM documentation before data collection. Questionable assessments were reviewed by a registered dietitian, and structured software fields were used to reduce free-text variability.
Outcome measures
Primary outcomes included:
Hospital-acquired infection (HAI): Defined according to Centers for Disease Control and Prevention criteria as any infection not present or incubating at admission that manifests ≥ 48 h after hospitalization [24].
30-day readmission: Any unplanned readmission to the same hospital within 30 days of discharge among patients discharged alive. Planned admissions for scheduled treatment were not counted as readmissions when identifiable.
In-hospital mortality: Death during the index hospitalization.
Secondary outcome
Length of stay (LOS), calculated as discharge date minus admission date.
Covariates
The following variables were extracted from the HIS: age, sex, BMI, primary diagnosis (categorized by department), Charlson Comorbidity Index (CCI), serum albumin, hemoglobin, and NRS2002 score.
Statistical analysis
Continuous variables were expressed as mean ± standard deviation (SD) for normally distributed data or median with interquartile range (IQR) for skewed distributions. Categorical variables were presented as frequencies and percentages. Normality was assessed using the Shapiro-Wilk test.
Two-group comparisons were performed using Student’s t test or the Mann-Whitney U test for continuous variables, as appropriate, and chi-square or Fisher exact tests for categorical variables. For the reviewer-requested three-group analysis (NRS2002 < 3, NRS2002 > = 3 but not malnourished, and malnourished), one-way ANOVA or Kruskal-Wallis tests were used for continuous variables and chi-square tests were used for categorical outcomes.
Multivariable logistic regression models were constructed for HAI, 30-day readmission, and in-hospital mortality. The adjustment set was specified a priori based on clinical relevance and included age > = 65 years, sex, CCI > = 2, department, and albumin. Albumin was retained as a marker of nutritional and inflammatory status, but it may also lie partly on the causal pathway between illness severity and malnutrition; results adjusted for albumin should therefore be interpreted conservatively. Results were expressed as adjusted odds ratios (aOR) with 95% CIs.
Sensitivity analyses included restriction to NRS2002-positive patients and a discharged-alive analytic set for 30-day readmission. To address potential time-at-risk bias for HAI, infection density per 1,000 patient-days was calculated as a descriptive sensitivity measure. Missing data were evaluated descriptively; because the provided analysis variables had no missing values after exclusions, complete-case models were used and multiple imputation was not performed.
All statistical analyses were performed using SPSS version 26.0 (IBM Corp., Armonk, NY) and R version 4.2.1 (R Foundation for Statistical Computing, Vienna, Austria). A two-sided P value < 0.05 was considered statistically significant.
Results
Patient characteristics
During the study period, 3,412 patients were screened for eligibility. After excluding 556 patients (163 for age < 18 years, 98 for admission < 24 h, 142 for incomplete NRS2002 screening, 89 for palliative care admission, 38 for psychiatric diagnosis, and 26 for missing outcome data), 2,856 patients were included in the final analysis (Fig. 1). The mean age was 59.23 ± 15.47 years, with 1,490 patients (52.17%) being male. The mean BMI was 23.48 ± 4.18 kg/m². Patient distribution across departments was as follows: internal medicine (n = 914; 32.00%), general surgery (n = 628; 21.99%), oncology (n = 514; 18.00%), orthopedics (n = 343; 12.01%), intensive care unit (ICU) (n = 229; 8.02%), and others (n = 228; 7.98%).
Fig. 1.
Study flow diagram showing patient enrollment and the stepwise nutritional screening-assessment pathway. Of 3,412 screened patients, 2,856 met inclusion criteria. NRS2002 identified 1,187 (41.56%) patients at nutritional risk and 1,669 (58.44%) patients with NRS2002 < 3. Among at-risk patients, 852 (71.78%) were diagnosed with malnutrition and 335 (28.22%) were classified as not malnourished after full assessment. The final well-nourished group therefore comprised 2,004 patients (1,669 NRS-negative plus 335 NRS-positive but not malnourished)
Nutritional screening and assessment results
Of the 2,856 patients screened using NRS2002, 1,187 (41.56%) were identified as being at nutritional risk (NRS2002 > = 3), whereas 1,669 (58.44%) had NRS2002 < 3. Among the 1,187 at-risk patients, 758 (63.86%) underwent SGA, 312 (26.29%) underwent PG-SGA, and 117 (9.85%) underwent GLIM evaluation. Following diagnostic assessment, 852 patients were classified as malnourished and 335 NRS-positive patients were classified as not malnourished. Thus, the pathway-defined malnutrition proportion was 29.83% of the total cohort and 71.78% among NRS-positive patients. Baseline characteristics stratified by nutritional status are presented in Table 1.
Table 1.
Baseline characteristics by nutritional status
| Variable | Well-nourished (n = 2,004) | Malnourished (n = 852) | P value |
|---|---|---|---|
| Demographics | |||
| Age, years (mean ± SD) | 56.18 ± 15.82 | 66.42 ± 12.36 | < 0.001 |
| Age ≥ 65 years, n (%) | 612 (30.54) | 478 (56.10) | < 0.001 |
| Male, n (%) | 1,048 (52.30) | 442 (51.88) | 0.836 |
| Anthropometric data | |||
| BMI, kg/m² (mean ± SD) | 24.56 ± 3.72 | 20.94 ± 4.28 | < 0.001 |
| BMI < 18.5 kg/m², n (%) | 89 (4.44) | 186 (21.83) | < 0.001 |
| Laboratory parameters | |||
| Albumin, g/L (mean ± SD) | 38.24 ± 4.86 | 31.56 ± 6.24 | < 0.001 |
| Hemoglobin, g/L (mean ± SD) | 124.36 ± 18.42 | 106.28 ± 22.86 | < 0.001 |
| Comorbidity | |||
| CCI (mean ± SD) | 1.62 ± 1.34 | 2.86 ± 1.72 | < 0.001 |
| CCI ≥ 2, n (%) | 724 (36.13) | 546 (64.08) | < 0.001 |
| Department, n (%) | |||
| Internal medicine | 672 (33.53) | 242 (28.40) | < 0.001 |
| General surgery | 476 (23.75) | 152 (17.84) | |
| Oncology | 280 (13.97) | 234 (27.46) | |
| Orthopedics | 265 (13.22) | 78 (9.15) | |
| ICU | 138 (6.89) | 91 (10.68) | |
| Others | 173 (8.63) | 55 (6.46) | |
| NRS2002 score | |||
| Score ≥ 3, n (%) | 335 (16.72) | 852 (100.00) | < 0.001 |
BMI, body mass index; CCI, Charlson Comorbidity Index; ICU, intensive care unit; NRS2002, Nutritional Risk Screening 2002; SD, standard deviation
Three-group screening-assessment analysis
To distinguish nutritional risk screening from malnutrition diagnosis, patients were additionally categorized as NRS2002 < 3 (n = 1,669), NRS2002 > = 3 but not malnourished after assessment (n = 335), and malnourished (n = 852). HAI rates were 4.85%, 7.46%, and 14.67%, respectively. Among discharged-alive patients, 30-day readmission rates were 11.19%, 11.18%, and 20.84%, respectively. In-hospital mortality rates were 2.52%, 1.19%, and 8.22%, respectively, and median LOS was 6.0, 6.0, and 10.0 days, respectively (all overall P < 0.001; Table 2).
Table 2.
Clinical outcomes across the three screening-assessment pathway groups
| Outcome | NRS2002 < 3 (n = 1,669) | NRS2002 > = 3 but not malnourished (n = 335) | Malnourished (n = 852) | P value |
|---|---|---|---|---|
| HAI, n (%) | 81 (4.85) | 25 (7.46) | 125 (14.67) | < 0.001 |
| 30-day readmission among discharged alive, n/N (%) | 182/1,627 (11.19) | 37/331 (11.18) | 163/782 (20.84) | < 0.001 |
| In-hospital mortality, n (%) | 42 (2.52) | 4 (1.19) | 70 (8.22) | < 0.001 |
| LOS, median (IQR), days | 6.0 (4.0–9.0) | 6.0 (4.0–10.0) | 10.0 (6.0–16.0) | < 0.001 |
HAI, hospital-acquired infection; IQR, interquartile range; LOS, length of stay; NRS2002, Nutritional Risk Screening 2002
Malnutrition prevalence by department
The pathway-defined malnutrition proportion varied across departments (P < 0.001), with oncology (234/514; 45.53%) and ICU (91/229; 39.74%) showing the highest values (Fig. 2). Because assessment tools were selected according to patient group and clinical characteristics, department-level and tool-specific proportions should not be interpreted as direct comparisons of diagnostic tool performance. Among at-risk patients, the proportions classified as malnourished were 75.00% for PG-SGA, 89.74% for GLIM, and 67.68% for SGA; the GLIM estimate reflects preselection of patients with BMI < 18.5 kg/m2.
Fig. 2.
Pathway-defined malnutrition proportion by department. Values should be interpreted cautiously because diagnostic assessment tools were selected according to patient group and clinical characteristics
Clinical outcomes
Table 3 summarizes clinical outcomes stratified by nutritional status. Hospital-acquired infection. The overall HAI rate was 8.09% (231/2,856). Malnourished patients had higher HAI rates compared with well-nourished patients (14.67% vs. 5.29%; P < 0.001). In the multivariable model, malnutrition was associated with HAI (aOR = 2.19; 95% CI: 1.57–3.07; P < 0.001). The most common HAIs were pneumonia (3.29%), urinary tract infection (2.17%), surgical site infection (1.75%), and bloodstream infection (0.87%) (Fig. 3).
Table 3.
Clinical outcomes by nutritional status
| Outcome | Well-nourished | Malnourished | P value |
|---|---|---|---|
| (n = 2,004) | (n = 852) | ||
| Hospital-acquired infection | |||
| Any HAI, n (%) | 106 (5.29) | 125 (14.67) | < 0.001 |
| Pneumonia, n (%) | 42 (2.10) | 52 (6.10) | < 0.001 |
| Urinary tract infection, n (%) | 28 (1.40) | 34 (3.99) | < 0.001 |
| Surgical site infection, n (%) | 24 (1.20) | 26 (3.05) | < 0.001 |
| Bloodstream infection, n (%) | 12 (0.60) | 13 (1.53) | 0.014 |
| 30-day readmission | |||
| 30-day readmission among discharged alive, n/N (%) | 219/1,958 (11.18) | 163/782 (20.84) | < 0.001 |
| In-hospital mortality | |||
| Deaths, n (%) | 46 (2.30) | 70 (8.22) | < 0.001 |
| Length of stay | |||
| Median (IQR), days | 6.0 (4.0–9.0) | 10.0 (6.0–16.0) | < 0.001 |
| LOS > 7 days, n (%) | 642 (32.04) | 512 (60.09) | < 0.001 |
HAI, hospital-acquired infection; IQR, interquartile range; LOS, length of stay
Fig. 3.
Forest plot of adjusted odds ratios for clinical outcomes associated with malnutrition. Readmission models used the discharged-alive analytic set. Estimates were adjusted for age, sex, CCI, albumin, and department
30-Day Readmission. Among 2,740 patients discharged alive, the 30-day readmission rate was 13.94% (382/2,740). Malnourished patients had higher readmission rates than well-nourished patients (163/782, 20.84% vs. 219/1,958, 11.18%; P < 0.001). In the discharged-alive multivariable model, malnutrition was associated with 30-day readmission (aOR = 1.81; 95% CI: 1.38–2.37; P < 0.001).
In-Hospital Mortality. The overall in-hospital mortality rate was 4.06% (116/2,856). Mortality was significantly higher in malnourished patients (8.22% vs. 2.30%; P < 0.001). Among malnourished patients, severe malnutrition was associated with markedly higher mortality (14.17%) compared with moderate malnutrition (5.69%; P < 0.001).
Length of Stay. Median LOS was significantly longer in malnourished patients (10.0 days; IQR: 6.0–16.0) compared with well-nourished patients (6.0 days; IQR: 4.0–9.0; P < 0.001). The proportion of prolonged hospitalization (LOS > 7 days) was 60.09% in malnourished versus 32.04% in well-nourished patients (P < 0.001).
Multivariable logistic regression analysis
After adjustment for age, sex, CCI, albumin, and department, malnutrition was associated with HAI (aOR = 2.19; 95% CI: 1.57–3.07), 30-day readmission among discharged-alive patients (aOR = 1.81; 95% CI: 1.38–2.37), and in-hospital mortality (aOR = 2.54; 95% CI: 1.59–4.04). Table 4 presents the full multivariable models for the three outcomes.
Table 4.
Multivariable logistic regression analysis for clinical outcomes
| Outcome | Variable | aOR (95% CI) | P value |
|---|---|---|---|
| Hospital-acquired infection | Reference categories | Well-nourished; Internal medicine | - |
| Malnourished | 2.19 (1.57–3.07) | < 0.001 | |
| Age > = 65 years | 1.11 (0.83–1.47) | 0.480 | |
| Male sex | 0.94 (0.71–1.24) | 0.658 | |
| CCI > = 2 | 1.46 (1.10–1.95) | 0.010 | |
| Albumin, per g/L | 0.98 (0.96–1.00) | 0.105 | |
| General surgery | 1.31 (0.88–1.93) | 0.181 | |
| Oncology | 1.71 (1.17–2.48) | 0.005 | |
| ICU | 1.28 (0.76–2.15) | 0.352 | |
| Orthopedics | 0.81 (0.47–1.40) | 0.444 | |
| Others | 0.73 (0.38–1.42) | 0.358 | |
| 30-day readmission among discharged-alive patients | Reference categories | Well-nourished; Internal medicine | - |
| Malnourished | 1.81 (1.38–2.37) | < 0.001 | |
| Age > = 65 years | 1.04 (0.82–1.31) | 0.753 | |
| Male sex | 1.11 (0.89–1.38) | 0.350 | |
| CCI > = 2 | 1.04 (0.83–1.30) | 0.741 | |
| Albumin, per g/L | 0.99 (0.97–1.01) | 0.164 | |
| General surgery | 0.89 (0.65–1.22) | 0.469 | |
| Oncology | 1.11 (0.81–1.52) | 0.516 | |
| ICU | 1.60 (1.08–2.36) | 0.019 | |
| Orthopedics | 0.95 (0.65–1.39) | 0.787 | |
| Others | 1.20 (0.79–1.83) | 0.396 | |
| In-hospital mortality | Reference categories | Well-nourished; Internal medicine | - |
| Malnourished | 2.54 (1.59–4.04) | < 0.001 | |
| Age > = 65 years | 1.24 (0.84–1.83) | 0.282 | |
| Male sex | 0.84 (0.57–1.22) | 0.361 | |
| CCI > = 2 | 1.42 (0.95–2.12) | 0.084 | |
| Albumin, per g/L | 0.96 (0.93–1.00) | 0.032 | |
| General surgery | 1.04 (0.61–1.76) | 0.888 | |
| Oncology | 0.90 (0.52–1.54) | 0.700 | |
| ICU | 1.43 (0.76–2.69) | 0.265 | |
| Orthopedics | 0.62 (0.28–1.35) | 0.227 | |
| Others | 0.93 (0.42–2.05) | 0.855 |
aOR, adjusted odds ratio; CCI, Charlson Comorbidity Index; CI, confidence interval; HAI, hospital-acquired infection; ICU, intensive care unit. Reference categories were well-nourished and internal medicine. Models adjusted for age > = 65 years, sex, CCI > = 2, albumin, and department. The readmission model included discharged-alive patients only
Subgroup analysis by malnutrition severity
When stratified by malnutrition severity, a dose-response relationship was observed for all outcomes (Table 5). Compared with well-nourished patients, severely malnourished patients had the highest risk of HAI (aOR = 3.86; 95% CI: 2.62–5.68), 30-day readmission (aOR = 2.42; 95% CI: 1.76–3.33), and in-hospital mortality (aOR = 4.68; 95% CI: 2.86–7.66).
Table 5.
Clinical outcomes by malnutrition severity
| Outcome | Well-nourished (n = 2,004) | Moderate (n = 598) | Severe (n = 254) | P for trend |
|---|---|---|---|---|
| HAI, n (%) | 106 (5.29) | 68 (11.37) | 57 (22.44) | < 0.001 |
| 30-day readmission among discharged alive, n/N (%) | 219/1,958 (11.18) | 106/564 (18.79) | 57/218 (26.15) | < 0.001 |
| In-hospital mortality, n (%) | 46 (2.30) | 34 (5.69) | 36 (14.17) | < 0.001 |
| LOS, median (IQR), days | 6.0 (4.0–9.0) | 9.0 (6.0–14.0) | 13.0 (8.0–21.0) | < 0.001 |
HAI, hospital-acquired infection; IQR, interquartile range; LOS, length of stay
In sensitivity analyses restricted to NRS2002-positive patients, malnutrition remained associated with HAI (aOR = 1.89; 95% CI: 1.18–3.03), 30-day readmission (aOR = 1.94; 95% CI: 1.31–2.86), and in-hospital mortality (aOR = 6.42; 95% CI: 2.28–18.11). Among NRS-positive patients discharged alive, malnutrition remained associated with 30-day readmission (aOR = 1.99; 95% CI: 1.33–2.96). HAI incidence density was 11.55 per 1,000 patient-days in malnourished patients and 7.46 per 1,000 patient-days in well-nourished patients.
Discussion
This large-scale retrospective cohort study evaluated nutritional risk and pathway-defined malnutrition in hospitalized patients using an integrated screening-assessment system. The findings indicate that approximately 30% of patients were diagnosed with malnutrition through this stepwise pathway, and that malnutrition was associated with higher HAI, 30-day readmission, in-hospital mortality, and LOS.
Prevalence of nutritional risk and malnutrition
Our NRS2002 nutritional risk prevalence of 41.56% is consistent with Chinese multicenter data reporting substantial nutritional risk among hospitalized patients [6]. The subsequent malnutrition proportion of 29.83% should be interpreted as pathway-defined malnutrition after NRS2002 screening and diagnostic assessment, not as a diagnostic prevalence obtained by applying all diagnostic tools to the entire cohort (Table 6).
Table 6.
Nutritional risk and malnutrition prevalence by department
| Department | Total | NRS2002 ≥ 3 n | Assessment tool | Malnourished n |
|---|---|---|---|---|
| (n) | (%) | (%) | ||
| Internal medicine | 914 | 346 (37.86) | SGA | 242 (26.48) |
| General surgery | 628 | 221 (35.19) | SGA | 152 (24.20) |
| Oncology | 514 | 312 (60.70) | PG-SGA | 234 (45.53) |
| Orthopedics | 343 | 102 (29.74) | SGA/GLIM | 78 (22.74) |
| ICU | 229 | 138 (60.26) | SGA/GLIM | 91 (39.74) |
| Others | 228 | 68 (29.82) | SGA | 55 (24.12) |
| Total | 2,856 | 1,187 (41.56) | - | 852 (29.83) |
GLIM, Global Leadership Initiative on Malnutrition; ICU, intensive care unit; NRS2002, Nutritional Risk Screening 2002; PG-SGA, Patient-Generated Subjective Global Assessment; SGA, Subjective Global Assessment
The variation in pathway-defined malnutrition across departments is clinically important but should be interpreted cautiously because assessment tool selection was linked to patient group. Oncology patients were assessed with PG-SGA, low-BMI non-oncology patients with GLIM, and other at-risk patients with SGA. Consequently, department and tool comparisons are confounded by clinical case mix and should not be read as evidence that one tool is intrinsically more sensitive than another.
Association between malnutrition and clinical outcomes
Our finding that malnutrition was associated with increased HAI risk is consistent with prior evidence linking malnutrition with nosocomial infection [25–27]. However, malnourished patients also had longer LOS, creating greater time at risk for HAI. We therefore calculated HAI incidence density as a descriptive sensitivity measure (11.55 vs. 7.46 per 1,000 patient-days), but a formal time-to-event model was not available in the present dataset.
The association between malnutrition and 30-day readmission has important care transition implications. To avoid including patients who died in hospital in the readmission risk set, the revised main analysis was restricted to discharged-alive patients. Readmission was also captured within the same hospital only, which may underestimate total readmissions if patients sought care elsewhere.
The mortality association (aOR = 2.54) observed in our study is consistent with international evidence. Higher NRS2002 scores were associated with an increased risk of in-hospital mortality in both unadjusted and adjusted models (OR = 1.602–1.736, P < 0.001). The dose-response relationship observed—with severe malnutrition conferring nearly five-fold increased mortality risk—emphasizes the clinical imperative for early identification and intervention.
Integrated nutrition software system
A notable feature of this study is the use of a clinical nutrition software system interfaced with the HIS for structured screening, diagnostic assessment, and outcome data capture. These functions may improve workflow standardization and documentation completeness. However, because the study did not include a pre-implementation or non-integrated comparator group, it cannot determine whether the software itself improved clinical outcomes.
The automatic selection of assessment tools based on patient characteristics (oncology patients receiving PG-SGA, low BMI patients receiving GLIM evaluation) ensures appropriate tool application while reducing clinician burden. The nutrition and dietetics profession has a key role in the interdisciplinary team in implementing the GLIM framework in clinical practice. Such integrated systems support this role by facilitating efficient workflows and data capture for quality improvement.
Clinical implications
Our findings have several clinical implications. First, admission nutritional risk screening can identify a large group requiring further evaluation, but screening positivity should not be conflated with malnutrition diagnosis. Second, patients identified as at risk require timely diagnostic assessment using population-appropriate tools. Third, the association between malnutrition severity and adverse outcomes supports prioritizing nutrition intervention planning for patients with confirmed moderate or severe malnutrition.
Evidence from randomized trials supports the effectiveness of nutritional interventions. The EFFORT trial demonstrated that individualized nutritional support in patients at nutritional risk reduced adverse outcomes including mortality [28–30]. Implementation of systematic screening programs has also shown cost-effectiveness through reduced complications and shorter hospital stays [31–33].
For ICU patients, screening and diagnosis should translate into individualized nutrition prescription. Recent quality-improvement evidence in critical care showed substantial discrepancies between indirect calorimetry and predictive equations, with measurement frequently changing nutrition therapy plans [34]. Where indirect calorimetry is unavailable, clinicians should recognize the limitations of predictive equations and reassess nutrition targets dynamically.
For patients requiring enteral nutrition, recent review evidence emphasizes indication-based formula selection, standard polymeric formula as usual initial practice, and early recognition and management of enteral feeding complications [35]. Thus, the clinical value of screening lies not only in identifying malnutrition but also in connecting diagnosis to bedside nutrition delivery.
Strengths and Limitations
This study has several strengths, including a large sample size, 42-month data collection period, use of validated screening and assessment tools, and integration with electronic records enabling outcome ascertainment in routine practice. The real-world setting enhances relevance for similar Chinese healthcare institutions.
Several limitations should be acknowledged. First, the retrospective design precludes causal inference. Second, the single-center design may limit generalizability to other settings with different patient populations or screening policies. Third, diagnostic tools were selected by clinical algorithm rather than applied uniformly to all patients, preventing direct comparison of SGA, PG-SGA, and GLIM performance and confounding department-level comparisons with case mix. Fourth, the GLIM subgroup was preselected by low BMI and therefore is not comparable with broader SGA or PG-SGA groups. Fifth, nutritional intervention and prescription data were not analyzed, limiting conclusions about treatment effectiveness. Sixth, longer LOS in malnourished patients may introduce time-at-risk bias for HAI; HAI incidence density was provided as a descriptive sensitivity analysis, but time-to-event modeling was not possible. Seventh, 30-day readmission was captured only within the same hospital and was analyzed among discharged-alive patients; readmissions to other institutions could have been missed. Eighth, illness severity scores, frailty, functional status, and inflammatory markers were unavailable across the full cohort, so residual confounding remains possible.
Conclusions
This retrospective cohort study found that malnutrition identified through a stepwise NRS2002-based screening-assessment pathway affected approximately 30% of hospitalized patients and was associated with higher HAI, readmission, mortality, and LOS. The findings support careful distinction between nutritional risk screening and malnutrition diagnosis, and suggest that HIS-integrated nutrition documentation can support standardized workflow. Further multicenter prospective studies are needed to evaluate implementation barriers, nutrition interventions, and causal effects on clinical outcomes.
Acknowledgements
Not applicable.
Abbreviations
- NRS2002
Nutritional risk screening 2002
- SGA
Subjective Global Assessment
- PG-SGA
Patient-generated subjective global assessment
- BMI
Body mass index
- GLIM
Global leadership initiative on malnutrition
- HAI
Hospital-acquired infection
- ESPEN
European society for clinical nutrition and metabolism
- HIS
Hospital information systems
- CCI
Charlson Comorbidity Index
- LOS
Length of stay
- IQR
Interquartile range
- ANOVA
Analysis of variance
- aOR
Adjusted odds ratios
- CI
Confidence intervals
- ICU
Intensive care unit
Author contributions
Qunfang Hua: Conceptualization; Data curation; Formal analysis; Investigation; Writing-original draft; Writing-review and editing. Yang Yu: Data curation; Formal analysis; Investigation; Methodology; Writing-review and editing. Yun Feng: Conceptualization; Investigation; Methodology; Writing-original draft; Writing-review and editing.
Funding
This study received no external funding.
Data availability
The data supporting the findings of this study are available from the corresponding author upon reasonable request.
Declarations
Ethics approval and consent to participate
The study protocol was approved by the Institutional Review Board of Xishan People’s Hospital (Approval No.: xs2024js048). All methods were carried out in accordance with Declaration of Helsinki. Due to the retrospective nature of the study and the use of de-identified data, the requirement for informed consent was waived.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Cederholm T, Jensen GL, Correia M, Gonzalez MC, Fukushima R, Higashiguchi T, Baptista G, Barazzoni R, Blaauw R, Coats A, et al. GLIM criteria for the diagnosis of malnutrition - A consensus report from the global clinical nutrition community. Clin Nutr. 2019;38:1–9. [DOI] [PubMed] [Google Scholar]
- 2.Barker LA, Gout BS, Crowe TC. Hospital malnutrition: prevalence, identification and impact on patients and the healthcare system. Int J Environ Res Public Health. 2011;8:514–27. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Stratton RJ, Hackston A, Longmore D, Dixon R, Price S, Stroud M, et al. Malnutrition in hospital outpatients and inpatients: prevalence, concurrent validity and ease of use of the “malnutrition universal screening tool” ('MUST’) for adults. Br J Nutr. 2004;92:799–808. [DOI] [PubMed]
- 4.Zhang H, Wang Y, Mu SY, Jiang ZM, Shi D, Yu K. [Prevalence of nutritional risks, malnutrition and application of nutritional support rates at one Chongqing teaching hospital]. Zhonghua Yi Xue Za Zhi. 2012;92:3417–9. [PubMed] [Google Scholar]
- 5.Chen X, Liu X, Ji W, Zhao Y, He Y, Liu Y, et al. The PG-SGA outperforms the NRS 2002 for nutritional risk screening in cancer patients: a retrospective study from China. Front Nutr. 2002;10:1272420, 1272420. [DOI] [PMC free article] [PubMed]
- 6.Fang S, Long J, Tan R, Mai H, Lu W, Yan F, Peng J. A multicentre assessment of malnutrition, nutritional risk, and application of nutritional support among hospitalized patients in Guangzhou hospitals. Asia Pac J Clin Nutr. 2013;22:54–9. [DOI] [PubMed] [Google Scholar]
- 7.Lim SL, Ong KC, Chan YH, Loke WC, Ferguson M, Daniels LM. Malnutrition and its impact on cost of hospitalization, length of stay, readmission and 3-year mortality. Clin Nutr. 2012;31:345–50. [DOI] [PubMed] [Google Scholar]
- 8.Norman K, Pichard C, Lochs H, Pirlich MPiod-rm. Prognostic impact of disease-related malnutrition. Clin Nutr. 2008;27:5–15. [DOI] [PubMed] [Google Scholar]
- 9.Agarwal E, Ferguson M, Banks M, Batterham M, Bauer J, Capra S, Isenring E. Malnutrition and poor food intake are associated with prolonged hospital stay, frequent readmissions, and greater in-hospital mortality: results from the Nutrition Care Day Survey 2010. Clin Nutr. 2013;32:737–45. [DOI] [PubMed] [Google Scholar]
- 10.Kondrup J, Rasmussen HH, Hamberg O, Stanga Z, Ad Hoc EWG. Nutritional risk screening (NRS 2002): a new method based on an analysis of controlled clinical trials. Clin Nutr. 2003;22:321–36. [DOI] [PubMed] [Google Scholar]
- 11.Kondrup J, Allison SP, Elia M, Vellas B, Plauth M, Educational. Clinical Practice Committee ESoP, Enteral N: ESPEN guidelines for nutrition screening 2002. Clin Nutr. 2003;22:415–21. [DOI] [PubMed] [Google Scholar]
- 12.Zhou Y, Chen Y, Zhang X, Zhao B, Gao F, Yuan X, Zhu Y, Liu D. Nutritional risk and a high NRS2002 score are closely related to disease progression and poor prognosis in patients with COVID-19. Front Nutr. 2023;10:1089972. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Detsky AS, McLaughlin JR, Baker JP, Johnston N, Whittaker S, Mendelson RA, Jeejeebhoy KN. What is subjective global assessment of nutritional status? JPEN J Parenter Enter Nutr. 1987;11:8–13. [DOI] [PubMed] [Google Scholar]
- 14.Ottery FD. Definition of standardized nutritional assessment and interventional pathways in oncology. Nutrition. 1996;12:S15–19. [DOI] [PubMed] [Google Scholar]
- 15.Bauer J, Capra S, Ferguson M. Use of the scored Patient-Generated Subjective Global Assessment (PG-SGA) as a nutrition assessment tool in patients with cancer. Eur J Clin Nutr. 2002;56:779–85. [DOI] [PubMed] [Google Scholar]
- 16.Zhang Z, Wan Z, Zhu Y, Zhang L, Zhang L, Wan H. Prevalence of malnutrition comparing NRS2002, MUST, and PG-SGA with the GLIM criteria in adults with cancer: A multi-center study. Nutrition. 2021;83:111072. [DOI] [PubMed] [Google Scholar]
- 17.Jensen GL, Cederholm T, Correia M, Gonzalez MC, Fukushima R, Higashiguchi T, de Baptista GA, Barazzoni R, Blaauw R, Coats AJS, et al. GLIM Criteria for the Diagnosis of Malnutrition: A Consensus Report From the Global Clinical Nutrition Community. JPEN J Parenter Enter Nutr. 2019;43:32–40. [DOI] [PubMed] [Google Scholar]
- 18.Jensen GL, Cederholm T, Correia M, Gonzalez MC, Fukushima R, Pisprasert V, et al. GLIM consensus approach to diagnosis of malnutrition: A 5-year update. JPEN J Parenter Enteral Nutr. 2025;49:414–27. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Jazinaki MS, Safarian M, Arabi SM, Jamali J, Norouzy. Validation of GLIM criteria for hospital malnutrition diagnosis by comparison of three different anthropometric approaches to evaluate reduced muscle mass: a prospective cohort study. Front Nutr. 2024;11:1438158, 1438158. [DOI] [PMC free article] [PubMed]
- 20.Trollebo MA, Skeie E, Revheim I, Stangeland H, Erstein MH, Gronning MK, Tangvik RJ, Morken MH, Nygard O, Eagan TML, et al. Comparison of nutritional risk screening with NRS2002 and the GLIM diagnostic criteria for malnutrition in hospitalized patients. Sci Rep. 2022;12:19743. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Jeejeebhoy KN, Detsky AS, Baker JAons. Assessment of nutritional status. JPEN J Parenter Enteral Nutr. 1990;14:193S-196S. [DOI] [PubMed] [Google Scholar]
- 22.Marshall S, Young A, Bauer J, Isenring EMiGRP. Malnutrition in Geriatric Rehabilitation: Prevalence, Patient Outcomes, and Criterion Validity of the Scored Patient-Generated Subjective Global Assessment and the Mini Nutritional Assessment. J Acad Nutr Diet. 2016;116:785–94. [DOI] [PubMed] [Google Scholar]
- 23.Bian W, Li Y, Wang Y, Chang L, Deng L, Li Y, et al. Prevalence of malnutrition based on global leadership initiative in malnutrition criteria for completeness of diagnosis and future risk of malnutrition based on current malnutrition diagnosis: systematic review and meta-analysis. Front Nutr. 2023;10:1174945, 1174945. [DOI] [PMC free article] [PubMed]
- 24.Horan TC, Andrus M, Dudeck MA. CDC/NHSN surveillance definition of health care-associated infection and criteria for specific types of infections in the acute care setting. Am J Infect Control. 2008;36:309–32. [DOI] [PubMed] [Google Scholar]
- 25.Chandra RK. Nutrition and the immune system: an introduction. Am J Clin Nutr. 1997;66:S460–3. [DOI] [PubMed] [Google Scholar]
- 26.Li W, Hou Z, Chen J, Pi R, Wang H, Dai M, Ouyang Y, Li S. Association Between Malnutrition and Multi-Drug Resistant Bacterial Infections in Neurosurgical Intensive Care Unit Patients. Infect Drug Resist. 2025;18:4485–98. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Schneider SM, Veyres P, Pivot X, Soummer AM, Jambou P, Filippi J, van Obberghen E, Hebuterne X. Malnutrition is an independent factor associated with nosocomial infections. Br J Nutr. 2004;92:105–11. [DOI] [PubMed] [Google Scholar]
- 28.Schuetz P, Fehr R, Baechli V, Geiser M, Deiss M, Gomes F, et al. Individualised nutritional support in medical inpatients at nutritional risk: a randomised clinical trial. Lancet. 2019;393:2312–21. [DOI] [PubMed] [Google Scholar]
- 29.Schuetz P, Sulo S, Walzer S, Vollmer L, Stanga Z, Gomes F, Rueda R, Mueller B, Partridge J. collaborators Et: Economic evaluation of individualized nutritional support in medical inpatients: Secondary analysis of the EFFORT trial. Clin Nutr. 2020;39:3361–8. [DOI] [PubMed] [Google Scholar]
- 30.Boaz M, Kaufman-Shriqui V. Systematic Review and Meta-Analysis: Malnutrition and In-Hospital Death in Adults Hospitalized with COVID-19. Nutrients 2023, 15. [DOI] [PMC free article] [PubMed]
- 31.Cortes-Aguilar R, Malih N, Abbate M, Fresneda S, Yanez A, Bennasar-Veny. Validity of nutrition screening tools for risk of malnutrition among hospitalized adult patients: a systematic review and meta-analysis. Clin Nutr. 2024;43:1094–116. [DOI] [PubMed]
- 32.Wong A, Huang Y, Banks MD, Sowa PM, Bauer JD. A Cost-Consequence Analysis of Nutritional Interventions Used in Hospital Settings for Older Adults with or at Risk of Malnutrition. Healthc (Basel) 2024, 12. [DOI] [PMC free article] [PubMed]
- 33.Liu JQ, He MJ, Zhang XQ, Zeng FH, Mo H, Shen JH. The association between nutrition risk status assessment and hospital mortality in Chinese older inpatients: a retrospective study. J Health Popul Nutr. 2024;43:229, 229. [DOI] [PMC free article] [PubMed]
- 34.Morand A, Graber K, Mohamed Elfadil O, Patel Y, Patel S, Loersch C, Wiggins I, Santoro A, Johnson N, Eckert K et al. Expanding the use of indirect calorimetry in critical care settings to determine the impact on nutrition interventions and compare to predictive equations: A quality improvement initiative. Clin Nutr ESPEN : 2025, 70:562–567. 10.1016/j.clnesp.2025.10.034. PMID: 41192554. [DOI] [PubMed]
- 35.Mohamed Elfadil O, Velapati SR, Patel J, Hurt RT, Mundi MS. Enteral Nutrition Therapy: Historical Perspective, Utilization, and Complications. Curr Gastroenterol Rep 2024, 26:200–210. 10.1007/s11894-024-00934-8. PMID: 38787510. [DOI] [PubMed]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The data supporting the findings of this study are available from the corresponding author upon reasonable request.