Clinical Geriatrics - Original Investigations
Published: 2022-07-30

Food intake and prevalence of malnutrition in nursing homes. A multicenter observational study

IRCSS Fondazione Don Carlo Gnocchi, Milan, Bachelor in Nursing, University of Milan, Milan, Italy
Fondazione Brescia Solidale, Brescia, Italy, APRIRE Network
Istituto di Ricerche Farmacologiche Mario Negri IRCCS, Milan, Italy
Department of Biomedical and Clinical Sciences “L. Sacco”, University of Milan, Milan, Italy
Fondazione Le Rondini Città di Lumezzane Onlus, Lumezzane (BS), Italy
Freelance nurse, Gorgonzola (MI) Italy
Health Care Agency, Milan, Italy
Segesta2000, Korian Group Italy, Milan, Italy
Ippocrate Nursing Home, Korian Group Italy, Milan, Italy
Freelance psychologist, Milan, Italy
Department of Biomedical Sciences for Health, University of Milan, ASST Grande Ospedale Metropolitano Niguarda, Milan, Italy
Ente Ospedaliero Cantonale, Department of Nursing, Bellinzona, Svizzera, University of Applied Sciences and Arts of Southern Switzerland (SUPSI), Department of Business Economics, Health and Social Care, Manno, Switzerland
malnutrition nursing homes food intake prevalence observational study


Objective. The aim of this study was to investigate malnutrition prevalence and its associated factors in NHs including resident food and liquid intake.
Methods. We conducted a multicenter observational study in Italian NHs. NHs were recruited using a convenience sample from regional register. To detect malnourished residents, we used Body Mass Index (BMI) < 18.5 kg/m2 and another criterion based on unintentional weight loss combined with reduced BMI. We performed logistic regression to identify associated factors with malnutrition.
Results. We recruited 1795 residents that lived in 29 NHs. 76% were female, with mean age 85.4 ± 8.5 years. 275 out of 1787 residents (15.4, 95% CI 13.8-17.2%) were malnourished . Moreover, when we combined BMI measures (< 20 kg/m2 if residents < 70 years or < 22 kg/m2 if residents ≥ 70 years) with unintentional weight loss data in the previous 6 months (5-10% or greater than 10%) we found that 18.1% residents were malnourished. Malnourished residents drank less (median 702 ml per day) than well nourished ones (median 800 ml per day) and ate less, especially at dinner. Malnutrition was significantly associated to age, psychiatric disorders, disability in activity of daily life, feeding assistance, swallowing and chewing problems and poor fluid intake. Moreover, we found a higher number of malnourished residents amongst those who needed more feeding care during dinner, when there was a lower number of staff. This result confirms the strong relationship between the care feeding and malnutrition.
Conclusions. In this study we found that malnourished residents were older with poor functional ability, chewing problems, dependence in feeding and poor fluid intake. Interventions to prevent malnutrition in nursing homes are a priority for the practice and for further studies.


Nutrition is important for wellbeing and health at all ages, but it is a core component of health maintenance in older people 1. Malnutrition, in particular, protein-energy malnutrition, is a very common issue in older people and it worsens quality of life, increases mortality, morbidity, infection and pressure ulcers rate, and reduces wound healing 2-5. In Nursing Homes (NHs) the malnutrition risk is higher because of the characteristics of the residents, who are older and frailer; coexisting malnutrition and severe frailty predict mortality in NHs setting 6.

In NHs, the prevalence of malnutrition ranges from 1.5 to 66.5%. This variability is due to different definitions of malnutrition and screening tools 2,7.

In 2015, the European Society of Clinical Nutrition and Metabolism (ESPEN) identified, after a consensus conference, two alternative criteria to diagnose malnutrition. The main criterion was the Body Mass Index (BMI) < 18.5 kg/m2. As an alternative, malnutrition could be diagnosed combining unintentional weight loss (mandatory) and at least one of either reduced BMI or a low fat free mass index (FFMI). Unintentional weight loss must be more than 10% indefinite of time, or > 5% over the last 3 months. BMI cut-offs are < 20 kg/m2 if < 70 years of age, or < 22 kg/m2 if 70 years of age and FFMI cut offs are < 15 and 17 kg/m2 in women and men 8.

Malnutrition in older people has a multifactor etiology that includes nutritional, clinical, neuropsychiatric and social factors 2,9. The nutritional factors composed of inadequate habits, insufficient food intake, decreased appetite, drug-nutrient interaction and restrictive dietary prescriptions. Among the clinical factors, we count chewing difficulties, edentulous and ineffective management of the prosthesis, nausea, vomiting, mal-absorption, protracted diarrhea, chronic diseases and dysphagia. The neuro-psychiatric factors comprise confusion, depression, dementia and other neurological disorders. The social factors included the cultural level, poverty, institutionalization, inadequate assistance during meals 10-13.

A poor food intake is a predictor for mortality 13, therefore, it is essential to monitor food intake on people at risk to become malnourished 14. Multifactorial strategies provided to NHs residents, such as organising mealtimes in a familiar way, favouring social interactions, providing individualised care, promoting self-feeding ability during the meal, seems to increase dietary intake 13-19. Furthermore, it is fundamental to identify early malnutrition risk using systematically tools, such as the Malnutrition Universal Screening and to monitoring malnutrition with multidisciplinary approach, involving health care team 12,20.

In the recent years, several studies investigated malnutrition prevalence and food intake in NHs, especially examining disease related factors 17,21, but only a few studies were performed in Italian residential facilities 9,22,23.

Knowing the magnitude and the characteristics of malnourished people is a key element to implement preventive strategies. Therefore, this study proposed to describe in Italian NHs the prevalence of malnutrition and related risk factors, including functional factors and food intake. Specifically, we were interested in exploring the following questions:

  1. What was the prevalence of malnutrition among nursing homes’ residents?
  2. Which demographic and main clinical characteristics were associated with malnutrition?
  3. Were food and fluid intake and feeding assistance associated with malnutrition?



We carried out an observational prospective multicenter study in a network of NHs in Milan and Brescia (Northern Italy), from October to November 2016. Our research was a part of a wide multi method study that aimed to explore nutritional care in NHs.


We involved regional NHs that included two main typologies of units based on residents’ characteristics: residents with prevalent chronic clinical problems and functional impairments (NHs units) and residents with severe dementia and Behavioural and Psychological Symptoms of Dementia (special care units for dementia).

NHs were recruited using a convenience sample. First, we enlisted the NHs by the regional register, considering their ownership (public, private), unit typology (NHs units and special care units for dementia) and facility size: small (40-90 residents), medium (91-160 residents) and large (more than 160 residents). Further information on NH selection is reported in Supplementary File.

We invited 32 NHs, representing various facility sizes, 20 NHs of which also had special care units for dementia. Out of 32 NHs, 29 accepted to participate in the study. In every nursing home, we selected randomly some units and special care wards for dementia. We included all residents living in randomized units at the time of the study, excepting people admitted in hospital for acute treatment.


To measure the prevalence and factors associated to malnutrition, data regarding:

  1. socio-demographic characteristics data of residents (gender, age, length of stay);
  2. clinical characteristics: autonomy in Activities of Daily Living (ADL) assessed with the Modified Barthel Index 24,25; cognitive function, using the Mini Mental State Examination (MMSE) 26,27; comorbidity by the Cumulative Illness Rating Scale (CIRS), which includes psychiatric/mental disorders 28,29, previous falls and pressure ulcers;
  3. nutritional conditions: chewing and swallowing problems and body mass index (BMI);
  4. NH characteristics such as ownership, facility size and unit type;
  5. residents’ intake, which included fluid (ml/day) and food intake (in a day) and need of assistance in feeding during the meals.

All healthcare personnel received a training to collect research data in two steps. In the first step, researcher AC, LB and MM, for every NH, trained a contact nurse, all registered nurses and physicians involved in the study. In the second stage, each nurse, with the support of researchers, trained and supervised the assistant nurses before and during the study.

Demographic and clinical data were collected from residents’ record with the support of nurses and physicians. All clinical data pertained to the previous three months; while MMSE and previous falls referred to six months prior to the study.

For all the residents that expressed positive consent, food and fluid intake were collected for two consecutive days including weekend days to detect any variability, through a specific questionnaire.

Intake observation tool, used in a previous pilot study 23, was reviewed and re-tested in one NH, before the study. Residents were observed continuously during the main meals (breakfast, lunch, dinner and refreshments).

After each meal, on the daily intake tool, the staff (nurses and assistant nurses) took note about: a) the type and amount of food and/or liquid consumed by residents b) time to consume the meal c) medications and feeding assistance during meal. The staff estimated the amount of food, in terms of portions eaten using a picture of a plate and a glass: full, three-quarters, half, a quarter, or nothing (observation tool are available in Supplementary File).

To guarantee reliable data, we planned maximum intake observations of two residents per day.


Primary outcome was nutritional status, codified as malnourished (BMI < 18.5) and well-nourished resident (BMI ≥ 18.5) 8. In addition, we explored the secondary diagnostic criterion based on unintentional weight loss > 10% indefinite of time, combined with reduced BMI: < 20 or < 22 kg/m2 in subjects younger or older than 70 years, respectively 8. We chose above-mentioned criterion, according to a preliminary feasibility analysis study.

According to criterion BMI < 18.5 to diagnose malnutrition, suggested by ESPEN 8, we expected prevalence of 30% malnourished residents.

The sample size was calculated on the average prevalence of malnutrition in NHs, as reported in the literature. Expecting a 30% prevalence of malnutrition, a sample of at least 323 people was estimated at 95% level of confidence, with a power of 80%. Calculating for possible dropouts of people unable to give the consent, we calculated a sample of approximately 700 residents in almost 23 NHs.

Regarding the intake and feeding assistance, we expected residents receiving more feeding assistance could be less malnourished.

With regard to nutritional conditions, we analysed the following variables: daily fluid intake (ml/day; calculated as the mean of intake in two consecutive days) and daily food intake at lunch and at dinner (categorised in 4 classes: 0-25%, 26-50%, 51-75%, 76-100%), feeding assistance during lunch and dinner (no assistance, minimal, supervision, active and total assistance). Details are shown in Supplementary File.

Categorical variables were reported using absolute frequencies and percentages.

Numerical variables were reported using mean and standard deviation or median, quartiles (Q1, Q3) and interquartile range (IRQ) for the variables with evident asymmetric distribution.

The prevalence of malnutrition and the association with the risk factors were assessed through univariate and multivariate logistic regression models, with nutritional status as response variable.

In the univariate analysis, the association was assessed including each factor in a logistic model as an explanatory variable.

In the multivariate analysis, we determined a priori the number of variables to put in the model, according to the 10-events-per-variable rule 30, to avoid overfitting 31. According to this rule, the variables considered were: gender, age, type of unit, Barthel score, psychiatric/mental disorders, pressure ulcers, chewing difficulty, swallowing difficulty, feeding assistance, food intake at lunch, food intake at dinner, fluid intake. Age and gender were considered as potential confounding factors. Using a backward variable selection procedure we excluded variables with not significant contribution to the model, except the confounding factors that were always kept in the model.

The results from the univariate and multivariate analysis were reported in terms of estimated Odd Ratio (OR) of malnutrition with respective 95% confidence intervals, and tests of association. The overall association was assessed by the Likelihood Ratio (LR) test (chi-square distribution).

For each test, the statistical significance was considered if the p-value was lower than 0.05. All the analyses were performed with the R software version 3.4.1 for Windows 32.

The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guideline was used to report the findings. Adherence to the STROBE checklist and further information on statistical analysis are documented in Supplementary File.



We recruited 1795 residents who lived in 29 NHs located in two provinces in Northern Italy, and we collected the data, from October to December 2016.

Primarily, we detected clinical and nutritional data, including all 1795 recruited residents. Out of 1795 residents, 8 had missing data on nutritional status. Residents were average 85.4 (± 8.5) years old; the 75.8% were female.

Overall, we found that out of 1787 residents, 275 were malnourished (15.4, 95% C.I. 13.8-17.2%) (Tab. I). The women, older residents, with pressure ulcer, with chawing and swallowing problems were more malnourished; these differences were statistically significant.

Moreover, according to second diagnostic criterion by ESPEN, when we considered the combination of BMI (20 kg/m2 if residents < 70 years old or < 22 kg/m2 if residents ≥ 70 years old) and unintentional weight loss (between 5-10% and greater than 10%) we found that 7.6% of residents (128 out of 1675) were malnourished (see also Supplementary file).

In the second stage of the study, we observed food and liquid intake, excluding residents on tube feeding or parenteral nutrition. Therefore, we included 1300 residents who met eligibility criteria and consented to meals observation (Tab. II).

Tables I-II show the main characteristics of participants comparing malnourished to those well nourished residents.

Malnourished residents drank less (median 702 ml per day) than well nourished (median 800 ml per day). Also they ate less, especially at the dinner; indeed the residents who ate nothing to half of the meal were 13.6% among the malnourished and 9% among well nourished people. Moreover malnourished residents needed more feeding care than the well-nourished one: 40.8 vs 18.7% residents received total assistance during the meals (Tab. II).


Results of the logistic regression models are presented in Tables III, IV and V.

In univariate analysis, malnutrition was significantly associated with female gender, older age, severe mental disorders, poor autonomy in daily activities (Barthel Index), and pressure ulcers in the last three months. In addition, residents were more malnourished when needed total assistance in feeding, had chewing and swallowing problems (Tab. III).

Overall, we found that more malnourished residents had a daily-reduced fluid and dinner food intake and had higher need of assistance during meals, especially during dinner (total assistance OR 3.92, 95% CI 2.60-5.93) (See Table IV). Finally, we found a lower prevalence of malnutrition in residents who lived in larger NHs (OR 0.62, 95% CI 0.42-0.90). No significant association emerged for the remaining variables.

In the multivariate analysis, we found significant association with nutritional status for: age, pressure ulcers, chewing difficulty, fluid intake, and feeding assistance (Tab. V).


This study aimed to measure the prevalence of malnutrition and to explore the variables associated with malnutrition, including food intake in Northern Italian NHs. Based on our knowledge, this is one of the few studies conducted in Italian facilities 9,22,23, that involved a large sample of NH residents. In addition, it is one of the first studies that used the ESPEN criteria 8, to intercept malnourishment within NHs, although recently new criteria to diagnosis malnutrition have been developed by the Global Leadership Initiative on Malnutrition (GLIM) malnutrition 33. The new GLIM criteria define malnutrition considering phenotypic (weight loss, low body mass index (BMI) and reduced muscle mass) and etiologic (reduced food intake or assimilation and inflammation) variables 33.

According the cut off of BMI < 18.5 Kg/m2, prevalence of malnutrition in our study was 15%, similarly to other studies 2. Interestingly, when used an alternative diagnostic criterion, based on the combination of unintentional weight loss, and low BMI we found a higher rate of 18% 8.

Older age, female, functional ability, psychiatric/mental disorders, pressure ulcers, chewing problems, swallowing problems, lower daily liquid intake and lower food intake were associated with a higher risk of being malnourished in univariate analysis. The results about older age and female are in line with findings from Keller et al. 34, but they are inconsistent with Van Zwienen et al. 21 Moreover, a strong association between malnutrition and pressure ulcers has also been found in other studies 35-37.

The level of assistance received during mealtimes was significantly associated with a greater number of malnourished residents, similarly to other studies 2,9-12,34. This is a crucial issue for healthcare professionals that should pay closer attention to nutrition problems and food intake among NHs’ residents 15.

Differently to Pezzana et al. 38, we found a higher prevalence of malnutrition in small sized NHs. This difference could be due to diverse local health systems, although both studies were conducted in Italy. Moreover, our results could be explained because bigger NHs have a more standardized approach to deal with nutrition services and the quality of nutritional care could be better. They have probably also more technical and human resources. Many factors such as the dining room environment, quality of the food, eating behaviours and preferences, clinical conditions, and level of autonomy in the activity of daily life, can influence food intake among NHs residents 34.

The multivariate logistic regression confirmed the positive association among malnutrition, swallowing and chewing problems, as other studies 39,40. Nurses play a strategic role in systematic assessment of residents’ oral health, since they could intercept at early stages a dysfunction in chewing and swallowing or other issues and adapt the food consistency and flavour consequently 41.

Multivariate analysis didn’t confirm the association between malnutrition and female gender, differently from a study conducted in hospitals within the same geographic area 42. However, as in the mentioned study 42, older age and pressure ulcers were associated with malnutrition also in multivariate analysis. This could represent a trend in the older population in Italy, as well as worldwide in developed countries.

Feeding assistance at dinner was also confirmed in the multivariate analysis, as a factor significantly associated with malnutrition. This is consistent with new GLIM criteria in which one among the etiological variable is food intake 33. This result confirms the strong relationship between feeding care and malnutrition. In fact, to provide a correct meal intake, it is important to offer residents an individualized nutritional care, guaranteeing them a sufficient time to eat and adequate staff during mealtimes 14-19,42. To overcome staff shortage issues in healthcare facilities, some strategies that could improve food intake in older people include the involvement of trained relatives or voluntaries during mealtimes 16,43.

This study confirmed the association between some variables and malnutrition. Therefore, NHs healthcare professionals should consider such variables in order to adopt prompt preventive strategies. Malnutrition can be considered a disease in the disease 42 and if not adequately prevented, it can worsen the clinical outcome in an already extremely frail population. Malnutrition among NHs residents should be treated by a multidisciplinary team, that include all healthcare professionals, dieticians, NHs food services’ and canteens’ personnel 20,44. Moreover, in order to ensure adequate nutritional intake, multimodal interventions should take into account individualised and comprehensive nutritional, hydration, oral care and mealtime assistance 44. For future studies, the new GLIM 33criteria should be used for the diagnosis of malnutrition.


This study has some limitations. Firstly, a different person in each NH performed data collection. Although the personnel involved in each facility received specific training and support by researchers, we cannot exclude some inaccuracies in data collection, due to the complexity of data.

Another possible weakness is selection bias, since we involved voluntary NHs. This could affect the reliability of the association, because of confounding variables, even if the sample size might have partially reduced this risk. To detect the intake, we used similar tools adopted by some facilities. We reviewed and re-tested each tool for the estimation of the amount of food and liquid, even if we did not test the interrater reliability (staff agreement).

Moreover, based on the feasibility analysis, in order to detect malnutrition it was not possible using a standard malnutrition risk tool and the fat free mass index (FFMI), since many NHs did not use regularly both and preferred not to implement them during that time.

Finally, even if we involved a large sample, our findings could not be representative of all Italian NHs.


Older people living in residential facilities are frailer, and therefore they are at a higher risk of complications such as malnutrition. This study provided a comprehensive description of malnutrition prevalence in NHs’ residents.

In our study malnutrition was associated with older age, pressure ulcers, chewing problems, fluid intake and feeding assistance. Therefore, healthcare personnel should systematically monitor resident’s nutritional status, including oral health, and prevent and treat malnutrition and dehydration as soon as possible. In order to favorite a good nutrition, it is important to offer residents an individualized nutritional care, guaranteeing them a sufficient time to eat and adequate staff during mealtimes.

The strong point of this study is that we involved a large number of NHs, with different size, in a vast area. Therefore, our findings could be considered a good picture of recent situation about malnutrition in the Italian NHs.


We thank Don Pierangelo Ondei that agreed to promote this study. We thank all the Nursing home directors and health personnel for their collaboration and contribution in data collection. We thank Silvia Revere and Gaetano De Angelis for their support for data management.

We applaud the nursing homes that agreed to participate in this study and made a great effort to encourage the involvement of consented residents. It was a challenge, considering that obtaining informed consent from nursing home residents, is a mainly obstacle, especially form resident with cognitive impairment. Moreover, we believe that these studies could promote a cultural change among all people – administrators, personnel, residents, relatives – involved for any level of the research. Finally, we thank Amalia Morelli for the valuable support on the English translation.


All the Authors declare no conflicts of interest.


For this study we did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

The nursing home “Piccolo Cottolengo don Orione”, in Milan, where worked in that time the principal investigator Anna Castaldo, promoted and supported the study with own funding. The nursing home “Korian group”, as research team, supported the certificated training of health personnel.


AC, EZ and LB: study concept and design; MZ, MM, MCG and GA: acquisition of data; AC, LB, EZ, GM, AN, AI and ML: Analysis and interpretation of data; AC and LB: drafting of the manuscript; EZ, AN and ML: critical revision of the manuscript for important intellectual content.


The ethical committee of each province (Number protocol: P1_267_2016; P2_2444_2016) and the Institutional Review Boards (IRB) of every participating NHs approved the study. All residents or their legal guardians (e.g. next of kin) were invited to give written informed consent, before participating in the study. All data were anonymised and collected in accordance with Italian regulations and the 1964 Helsinki Declaration and its latest amendments.

Figures and tables

Malnourished (n = 275) Well nourished (n = 1512) Overall (n = 1787) P-value
Female * n (%) 236 (85.8) 1119 (74.0) 1360 (75.8) < .001
Age (years) < .001
Mean (SD) 87.5 (8.4) 85.0 (8.4) 85.4 (8.5)
Range 51.2-105.5 49.0-105.2 49.0-105.5
CIRS comorbidity Index §
Mean (SD) 5.55 (2.07) 5.41 (2.11) 5.43 (2.11) .32
CIRS severity index ||
Mean (SD) 2.04 (0.39) 2.02 (0.40) 2.02 (0.40) .23
Range 1.00-3.46 1.00-3.38 1.00-3.46
Psychiatric/mental state
(including dementia (14^ category of CIRS)* .006
No problem 22 (8.0) 131 (8.7) 153 (8.6)
Mild 15 (5.5) 89 (5.9) 104 (5.8)
Moderate 66 (24.0) 519 (34.3) 585 (32.7)
Severe 132 (48.0) 633 (41.9) 765 (42.8)
Very severe 39 (14.2) 138 (9.1) 177 (9.9)
Missing 1 (0.4) 2 (0.1) 3 (0.2)
Barthel score **
Median; Q1-Q3 8; 0-23 24; 6-57 21; 5-52 < .001
Range 0-95 0-100 0-100
Missing* 3 (1.1) 1 (0.1) 4 (0.2)
Chewing problem *
No 133 (48.4) 1037 (68.5) 1170 (65.5) < .001
Yes 140 (50.9) 468 (31.0) 608 (34.0)
Missing 2 (0.7) 7 (0.5) 9 (0.5)
Swallowing problem*
No 183 (66.5) 1252 (82.8) 1435 (80.3) < .001
Yes 90 (32.7) 254 (16.8) 344 (19.3)
Missing 2 (0.7) 6 (0.4) 8 (0.4)
Pressure ulcers*
None 227 (82.5) 1371 (90.6) 1598 (89.4) < .001
One or more 48 (17.5) 140 (9.3) 187 (10.6)
Missing 0 1 (0.1) 1 (0.1)
Accidental falls*
No falls 232 (84.5) 1263 (83.5) 1495 (83.7) .72
One or more falls 41 (4.9) 242 (16.0) 283 (15.8)
Missing 2 (0.4) 7 (0.5) 9 (0.5)
Length of stay (years)
Median; Q1-Q3 2.3; 0.8-5.2 2.0; 0.8-4.4 2.1; 0.8-4.5 .20
Range 0.0-66.9 0.0-76.6 0.0-76.6
Missing* 3 (1.1) 8 (0.5) 11 (0.6)
Table I.Demographic and clinical characteristics of NHs residents.
Malnourished (n = 169) Well nourished (n = 1131) Overall (n = 1300) P-value
Fluid intake (ml)
Median; Q1-Q3 702; 612-881 800; 672-854 796; 650-975 < .001
Range 435-1550 310-1800 310-1800
Missing* 14 (8.2) 104 (9.2) 118 (9.1)
Food intake at lunch (portion)*
0-25% 7 (4.1) 28 (2.5) 35 (2.7) .17
26-50% 12 (7.1) 70 (6.2) 82 (6.3)
51-75% 37 (21.9) 184 (16.3) 221 (17.0)
76-100% 112 (66.3) 848 (75.0) 960 (73.8)
Missing 1 (0.6) 1 (0.1) 2 (0.2
Food intake at dinner (portion)*
0-25% 3 (1.8) 28 (2.5) 31 (2.4) .013
26-50% 20 (11.8) 73 (6.5) 93 (7.2)
51-75% 40 (23.7) 196 (17.3) 236 (18.2)
76-100% 106 (62.7) 829 (73.3) 935 (71.9)
Missing 0 (0.0) 5 (0.4) 5 (0.4)
Feeding lunch, day 1* < .001
No assistance 41 (24.3) 494 (43.7) 535 (41.2)
Minimal assistance 28 (16.6) 250 (22.1) 278 (21.4)
Supervision 24 (14.2) 106 (9.4) 130 (10.0)
Active assistance 9 (5.3) 44 (3.9) 53 (4.1)
otal assistance 63 (37.3) 223 (19.7) 286 (22.0)
Did not eat that day 4 (2.4) 10 (0.9) 14 (1.1)
Missing 0 (0.0) 4 (0.4) 4 (0.3)
Feeding dinner, day 1 * < .001
No assistance 44 (26.0) 530 (46.9) 574 (44.2)
Minimal assistance 26 (15.4) 222 (29.6) 248 (19.1)
Supervision 19 (11.2) 98 (8.7) 117 (9.0)
Active assistance 8 (4.7) 32 (2.8) 40 (3.1)
Total assistance 69 (40.8) 212 (18.7) 281 (21.6)
Did not eat that day 2 (1.2) 3 (2.7) 33 (2.5)
Missing 1 (0.6) 6 (0.5) 7 (0.5)
Table II.Food and fluid intake characteristics of NHs residents.
Strength of association: OR (95% C.I.) LR test: Chi-square (df) P-value
Gender 19.512(1) < .001
Male 1
Female 2.13 (1.49-3.04)
Age * 1.48 (1.25-1.75) 22.128(1) < .001
length of stay 3.659(3) .30
0-1 year 1
1.1-3 years 0.95 (0.67-1.33)
3.1-10 years 1.27 (0.92-1.74)
> 10 years 1.11 (0.60-2.07)
Type of structure 2.796(1) .09
Public 1
Private 1.31 (0.95-1.80)
Facility size 17.344(2) < .001
40-90 residents 1
91-160 residents 1.08 (0.74-1.58)
> 160 residents 0.62 (0.42-0.90)
Type of unit 0.195(1) .65
Nursing home units 1
Special care units for dementia 0.93 (0.69-1.23)
Unit size 2.57(1) .10
Small 1
Medium/high 1.33 (0.95-1.87)
Pressure ulcers 14.605(1) < .001
None 1
One or more 2.07 (1.45-2.96)
Accidental falls 0.197(1) .65
None 1
One or more 0.92 (0.64-1.32)
CIRS severity index 1.21 (0.88-1.67) 1.353(1) .24
CIRS comorbidity index 1.03 (0.97-1.10) 0.961(1) .32
Psychiatric/mental disorders (by CIRS) 15.822(4) .003
1 No problem 1
2 Mild 1.00 (0.49-2.04)
3 Moderate 0.76 (0.45-1.27)
4 Severe 1.24 (0.76-2.02)
5 Very severe 1.68 (0.95-2.99)
AdL - Barthel score 1.27 (1.20-1.35) 80.770(1) < .001
Chewing difficulty 40.023(1) < .001
No 1
Yes 2.33 (1.80-3.03)
Swallowing difficulty 33.384(1) < .001
No 1
Yes 2.42 (1.82-3.23)
Feeding assistance (Barthel Index) 58.630(4) < .001
None 1
Minimal 1.19 (0.72-1.97)
Supervision 2.70 (1.73-4.22)
Active 2.33 (1.41-3.85)
Total 3.94 (2.52-6.16)
Table III.Association between malnutrition and demographic and clinical variables: univariate analysis.
Strength of association: OR (95% C.I.) Global test: Chi-square (df) P-value
Fluid intake* (ml): 1.15 (1.07-1.24) 14.481(1) < .001
Food intake at lunch 4.978(3) .17
0-25 1
26-50 0.77 (0.26-2.26)
51-75 0.90 (0.35-2.35)
76-100 0.59 (0.24-1.47)
Food intake at dinner 10.753(3) .013
0-25 1
26-50 2.47 (0.68-8.97)
51-75 1.84 (0.53-6.35)
76-100 1.15 (0.34-3.86)
Feeding assistance, lunch, day 1 39.085(3) < .001
None 1
Minimal 1.35 (0.82-2.23)
Supervision 2.73 (1.58-4.71)
Active 2.46 (1.12-5.40)
Total 3.40 (2.23-5.20)
Feeding assistance, dinner, day 1 47.919(3) < .001
None 1
Minimal 1.41 (0.85-2.35)
Supervision 2.34 (1.31-4.17)
Active 3.01 (1.31-6.93)
Total 3.92 (2.60-5.93)
Table IV.Association between malnutrition and food intake: univariate analysis.
Strength of association: OR (95% C.I.) LR test: Chi-square (df) P-value
Gender 3.83(1) 0.05
Male 1
Female 1.63 (1.00-2.65)
Age 6.81(1) .009
< 80 years 1
≥ 80 years 1.99 (1.19-3.34)
Pressure ulcers 5.49 (0) .019
No 1
Yes 1.89 (1.11-3.22)
Chewing difficulty 10.58(1) .001
No 1
Yes 1.97 (1.31-2.96)
Fluid intake* (ml) 1.14 (1.05-1.24) 10.26(1) .001
Feeding assistance, dinner 8.42(1) .003
Low 1
High 1.84 (1.22-2.77)
Table V.Association between malnutrition and demographic, clinical and food intake: multivariate analysis.



Before proceeding with recruitment, the NHs contractualized with the Regional Health Care System were classified on the basis of their size defined on number of beds and/or the presence of specialised unit for dementia. Expecting a 30% prevalence of malnutrition, we estimated involving 1000 residents, recruiting 23 NHs, that represent 10% of the overall contractualized NHs. We invited NHs in order to obtain a sample of residents equally representative of different NH size (see Table 1).


The main goal of the analysis was to assess the prevalence of malnutrition and its association with the variables previously illustrated. Categorical variables were reported using absolute frequencies and percentages. Numerical variables were reported using mean and standard deviation or median quartiles (Q1 Q3) for the variables with evident asymmetric distribution. The prevalence of malnutrition and the association with the considered factors were assessed through logistic regression models with nutritional status as response variable. The strength of association was estimated as the ratio of the odds (OR) of malnutrition for each category of the explanatory variable considering one category as reference. The variables measured on continuous scale (i.e.: age, length of stay, severity and comorbidity indices, Barthel score and liquids intake) were included in the regression models as linear terms in their original measurement scale.

First, the association was assessed separately for each factor (univariate analysis), including each factor in a logistic model as explanatory variable. In multivariate analysis we followed the guidelines described in Harrell’s textbook (Harrell FE. Regression modelling strategies With Applications to Linear Models Logistic and Ordinal Regression and Survival Analysis. 2nd ed. Springer editor. 2015. 582 p.) to avoid the possibility of unreliable conclusions caused by overfitting the model. The data considered for the multivariable model consisted of 1116 records of subjects with complete information of the variables listed below. The number of variables to be considered in the analysis was determined a priori, according to the 10-event-per-variable rule (Peduzzi P, Concato J, Kemper E, et al. A simulation study of the number of events per variable in logistic regression analysis. J Clin Epidemiol 1996;49:1373-1379). Within the 1116 available records 97 events (i.e. primary outcomes, malnourished subjects) were observed, thus the maximum number of coefficients to be included in the model was 97/10 = approximately 10. The choice of variables was based on clinical judgement:

  1. Gender (M/F);
  2. Age (codified as categorical variables with two modalities: ≤ 80 years; > 80 years;
  3. NH ownership (public, private);
  4. NH size (40-90 residents; 91-160 residents; > 160 residents);
  5. Barthel Index score (0-90; 91-100);
  6. Mental state by CIRS (1-3; 4-5);
  7. Pressure ulcers (None, present);
  8. chewing difficulty (No/YES);
  9. swallowing difficulty (No/YES);
  10. assisted nutrition by Barthel Index (low = minimal or no assistance; high = partial or constant or total assistance);
  11. daily total liquids intake;
  12. daily food intake at lunch (0-50; 51-100);
  13. daily food intake at dinner (0-50; 51-100);
  14. assisted nutrition at lunch at day 1 (low = minimal or no assistance; high = partial or constant or total assistance);
  15. assisted nutrition at dinner at day 1 (low = minimal or no assistance; high = continuous or substantial or total assistance).

The model building procedure was as follows: first, a model with all the variables above as independent variables was fitted. Then (a backward variable selection technique was used for excluding variables with not significant multivariate association with nutritional status. The procedure was “constrained” to keep age and gender in the model at every step, due to their role of potential confounders.

The results from all the models (univariate and multivariate) were reported in terms of estimated OR of malnutrition with respective 95% confidence intervals and tests of association. Concerning the latter ones the overall association was assessed by Likelihood Ratio (LR) test (chi-square distribution). Moreover the Wald test (student t distribution) was used in the following circumstances:

  1. in univariate analysis, for categorical variables with more than two modalities: when the global association resulted statistically significant, the test was used for multiple comparisons between pre-specified modalities;
  2. in multivariate analysis, to assess the association between nutritional status and each of the variables included in the model.
  3. For each test, statistical significance was deemed when the p-value was lower than 0.05. All the analyses were performed with the R software version 3.4.1 (R Core Team, 2017).
40-90 residents N (%) 91-160 residents N (%) > 160 residents N (%) Total N (%)
MI -Tot residents contractualized NHs 3926(57) 7004(60) 4872(15) 15802(131)
MI - number minimum residents for sample size* 118(6) 210(6) 146(2) 474(14)
MI - Tot residents in NHs invited 303(4) 959(11) 1319(7) 2581(22)
MI - Selected residents in NHs 63(2) 504(11) 903(7) 1470(20)
BS - Tot residents in contractualized NHs 3389(55) 2759(25) 512(3) 6660(83)
BS - N minimum residents for sample size* 102(5) 83(3) 15(1) 200(9)
BS - Tot residents in NHs invited 391(5) 480(4) 175(1) 1046(10)
BS - selected residents in NHs 115(5) 162(3) 65(1) 342(9)
Tot invited residents in both provinces 694(9) 1439(15) 1494(8) 3527(32)
Tot selected residents in both provinces 178(7) 666(14) 968(8) 2396(29)
Tot effective residents in both provinces 250 639 906 1795
Table 1.Definition of sample size and involved residents’ NHs.
Malnourished (n = 275) N (%) Well nourished (n = 1512) N (%) Overall (n = 1787) N (%) P-value
MMSE (cognitive status) < .001
> 24 (intact) 23 (8.4) 247 (16.3) 270 (15.0)
20-24 (borderline) 16 (5.8) 223 (14.7) 239 (13.4)
17-19 (mild impairment) 13 (4.7) 131 (8.7) 144 (8.1)
10-16 (moderate impairment) 38 (13.8) 255 (16.9) 293 (16.4)
0-9 (severe impairment) 83 (30.2) 258 (17.1) 341 (19.1)
Missing 2 (0.7) 24 (1.6) 26 (1.5)
Not applicable 100 (36.4) 374 (24.7) 474 (26.6)
MUST (malnutrition risk) < .001
Low risk (0) - 1054 (69.7) 1054 (58.9)
Medium risk (1) - 259 (17.1) 259 (14.5)
High risk (2 or more) 254 (92.4) 106 (7.0) 360 (20.1)
Missing 21 (7.6) 93 (6.2) 114 (6.4)
Feeding by Barthel Index < .001
Total assistance 86 (31.5) 253 (16.7) 339 (19.0)
Active assistance 41 (14.9) 204 (13.5) 245 (13.7)
Supervision 78 (28.4) 335 (22.2) 413 (23.1)
Minimal assistance 78 (28.4) 370 (24.5) 408 (22.8)
No assistance 30 (10.9) 348 (23.0) 378 (21.2)
Missing 2 (0.7) 2 (0.1) 4 (0.2)
Brescia 60 (21.8) 256 (16.9) 316 (17.7) 0.06
Milano 215 (78.2) 1256 (83.1) 1471 (82.3)
NH ownership type
Public 54 (19.6) 366 (24.2) 420 (23.5) 0.11
Private 221 (80.4) 1146 (75.8) 1367 (76.5)
Facility size < .001
40-90 residents 45 (16.4) 204 (13.5) 249 (13.9)
91-160 residents 123 (44.7) 515 (34.1) 638 (35.7)
> 160 residents 107 (38.9) 793 (52.4) 900 (50.4)
Type of ward
Nursing home 212 (77.1) 1147 (75.9) 1359 (76.0) 0.71
Special care units for dementia 63 (22.9) 365 (24.1) 428 (24.0)
Unit size:
Small 226 (82.2) 1300 (86.0) 1526 (85.4) 0.12
Medium/high 49 (17.8) 212 (14.0) 261 (14.6)
Table 2.Further clinical characteristics of NHs residents and organizational characteristics.
Unintentional weight loss
No or < 5% 5-10% > 10% Total
Underweight (BMI) N (%) N (%) N (%)
No 866 (60.0) 87 (50.9) 17 (27.9) 970 (57.9)
Yes* 577 (40.0) 84 (49.1) 44 (72.1) 705 (42.0)
Total 1443 (100.0) 171 (100.0) 61 (100.0) 1675§
Table 3.Malnourished residents according to combined unintentional weight loss with BMI.


  1. Chen CCH, Schilling LS, Lyder CH. A concept analysis of malnutrition in the elderly. J Adv Nurs. 2001; 36:131-142. DOI
  2. Bell CL, Lee ASW, Tamura BK. Malnutrition in the nursing home. Curr Opin Clin Nutr Metab Care. 2015; 18:17-23. DOI
  3. Beck AM. Weight loss, mortality and associated potentially modifiable nutritional risk factors among nursing home residents – a Danish follow-up study. J Nutr Health Aging. 2015; 19:96-101. DOI
  4. Chen LY, Liu LK, Hwang AC. Impact of malnutrition on physical, cognitive function and mortality among older men living in veteran homes by minimum data set: a prospective cohort study in Taiwan. J Nutr Heal Aging. 2016; 20:41-47. DOI
  5. Saarela RKT, Muurinen S, Suominen MH. Changes in malnutrition and quality of nutritional care among aged residents in all nursing homes and assisted living facilities in Helsinki 2003-2011. Arch Gerontol Geriatr. 2017; 72:169-173. DOI
  6. Kamo T, Takayama K, Ishii H. Coexisting severe frailty and malnutrition predict mortality among the oldest old in nursing homes: a 1-year prospective study. Arch Gerontol Geriatr. 2017; 70:99-104. DOI
  7. Van Bokhorst-de van der Schueren MAE, Guaitoli PR, Jansma EP. Nutrition screening tools: does one size fit all? A systematic review of screening tools for the hospital setting. Clin Nutr. 2014; 33:39-58. DOI
  8. Cederholm T, Bosaeus I, Barazzoni R. Diagnostic criteria for malnutrition – an ESPEN Consensus Statement. Clin Nutr. 2015; 34:335-340. DOI
  9. Papparotto C, Bidoli E, Palese A. Risk factors associated with malnutrition in older adults living in Italian nursing homes: a cross-sectional study. Res Gerontol Nurs. 2013; 6:187-197. DOI
  10. Tamura BK, Bell CL, Masaki KH. Factors associated with weight loss, low BMI, and malnutrition among nursing home patients: a systematic review of the literature. J Am Med Dir Assoc. 2013; 14:649-655. DOI
  11. Palm R, Reuther S, Bartholomeyczik S. Associated factors of different nutrition indicators in German nursing home residents Comparative results of a multicenter cross-sectional study. Z Gerontol Geriatr. 2012; 45:658-664. DOI
  12. Willis H. Causes, assessment and treatment of malnutrition in older people. Nurs Older People. 2017; 29:20-25. DOI
  13. Streicher M, Themessl-Huber M, Schindler K. NutritionDay in nursing homes – the Association of Nutritional Intake and Nutritional Interventions with 6-month mortality in malnourished residents. J Am Med Dir Assoc. 2017; 18:162-168. DOI
  14. Baldwin C, Kimber KL, Gibbs M. Supportive interventions for enhancing dietary intake in malnourished or nutritionally at-risk adults. Cochrane Database Syst Rev. 2016;12. DOI
  15. Keller H, Beck AM, Namasivayam A. Improving food and fluid intake for older adults living in long-term care: a research agenda. J Am Med Dir Assoc. 2015; 16:93-100. DOI
  16. Edwards D, Carrier J, Hopkinson J. Assistance at mealtimes in hospital settings and rehabilitation units for patients (&gt; 65 years) from the perspective of patients, families and healthcare professionals: a mixed methods systematic review. Int J Nurs Stud. 2017; 69:100-118. DOI
  17. Van Nie-Visser NC, Meijers J, Schols J. Which characteristics of nursing home residents influence differences in malnutrition prevalence? An international comparison of The Netherlands, Germany and Austria. Br J Nutr. 2014; 111:1129-1136. DOI
  18. Hedman S, Nydahl M, Faxén-Irving G. Individually prescribed diet is fundamental to optimize nutritional treatment in geriatric patients. Clin Nutr. 2016; 35:692-698. DOI
  19. Lee LC, Tsai AC, Wang JY. Need-based nutritional intervention is effective in improving handgrip strength and Barthel Index scores of older people living in a nursing home: a randomized controlled trial. Int J Nurs Stud. 2015; 52:904-912. DOI
  20. Lusignani M, Mari D. Evaluation of the nutritional state of the elderly. Eur J Intern Med. 2013; 24:e11-e12. DOI
  21. van Zwienen-Pot JI, Visser M, Kuijpers M. Undernutrition in nursing home rehabilitation patients. Clin Nutr. 2017; 36:755-759. DOI
  22. Donini LM, Neri B, De Chiara S. Nutritional care in a nursing home in Italy. PLoS One. 2013; 8:e55804.
  23. Di Stefano C, Castaldo A, Pancheri L. L’assistenza nutrizionale agli anziani: uno studio osservazionale presso una struttura residenziale per anziani a Milano. I Luoghi della Cura. 2015; 4:16-21.
  24. Shah S, Vanclay F, Cooper B. Improving the sensitivity of the Barthel Index for stroke rehabilitation. J Clin Epidemiol. 1989; 42:703-709. DOI
  25. Zanetti E. Carrocci &amp; Faber: Rome, Italy; 2003.
  26. Folstein MF, Folstein S, McHugh PR. “Mini-Mental State” a pratical method for grading the cognitive state of patients for the clinician. J Psychiatr Res. 1975; 12:189-198. DOI
  27. Frisoni GB, Rozzini R, Bianchetti A. Principal lifetime occupation and MMSE score in elderly persons. J Gerontol. 1993; 48:S310-S314.
  28. Parmelee PA, Thuras PD, Katz IR. Validation of the Cumulative Illness Rating scale in a geriatric residential population. J Am Geriatr Soc. 1995; 43:130-137. DOI
  29. Salvi F, Miller MD, Towers AL. Manuale di linee guida per la valutazione della scala di comorbidità “Modified Cumulative Illness Rating Scale (CIRS)” Manuale CIRS, tradotto dall’ANQ, versione 10 /2013 Indice. 2013; 5610:1-30.
  30. Peduzzi P, Concato J, Kemper E. A simulation study of the number of events per variable in logistic regression analysis. J Clin Epidemiol. 1996; 49:1373-1379.
  31. Harrell FE. Springer: New York, NY; 2015.
  32. R Core Team. R: a language and environment for statistical computing. 2017.
  33. Cederholm T, Jensen GL, Correia MITD. GLIM criteria for the diagnosis of malnutrition – a consensus report from the global clinical nutrition community. 2019; 1:207-217.
  34. Keller HH, Carrier N, Slaughter S. Making the Most of Mealtimes (M3): protocol of a multi-centre cross-sectional study of food intake and its determinants in older adults living in long term care homes. BMC Geriatr. 2017; 17:1-15. DOI
  35. Iizaka S, Okuwa M, Sugama J. The impact of malnutrition and nutrition-related factors on the development and severity of pressure ulcers in older patients receiving home care. Clin Nutr. 2010; 29:47-53. DOI
  36. Little MO. Nutrition and skin ulcers. Curr Opin Clin Nutr Metab Care. 2013; 16:39-49. DOI
  37. Shahin ESM, Meijers JMM, Schols JMGA. The relationship between malnutrition parameters and pressure ulcers in hospitals and nursing homes. Nutrition. 2010; 26:886-889. DOI
  38. Pezzana A, Cereda E, Avagnina P. Nutritional care needs in elderly residents of long-term care institutions: Potential implications for policies. J Nutr Health Aging. 2015; 19:947-954. DOI
  39. Mann T, Heuberger R, Wong H. The association between chewing and swallowing difficulties and nutritional status in older adults. Aust Dent J. 2013; 58:200-206. DOI
  40. Kossioni AE. The Association of poor oral health parameters with malnutrition in older adults: a review considering the potential implications for cognitive impairment. Nutrients. 2018;10. DOI
  41. Aro T, Laitala M, Syrjälä AM. Perceptions of older people’s oral health care among nurses working in geriatric home care. Acta Odontol Scand. 2018; 76:427-432. DOI
  42. Bonetti L, Terzoni S, Lusignani M. Prevalence of malnutrition among older people in medical and surgical wards in hospital and quality of nutritional care: a multicenter, cross-sectional study. J Clin Nurs. 2017; 26:5082-5092. DOI
  43. Vucea V, Keller HH, Ducak K. Interventions for improving mealtime experiences in long-term care. J Nutr Gerontol Geriatr. 2014; 33:249-324. DOI
  44. Volkert D, Beck AM, Cederholm T. ESPEN guideline on clinical nutrition and hydration in geriatrics. Clin Nutr. 2019; 38:10-47. DOI


Anna Castaldo

IRCSS Fondazione Don Carlo Gnocchi, Milan, Bachelor in Nursing, University of Milan, Milan, Italy

Ermellina Silvia Zanetti

Fondazione Brescia Solidale, Brescia, Italy, APRIRE Network

Alessandro Nobili

Istituto di Ricerche Farmacologiche Mario Negri IRCCS, Milan, Italy

Giuseppe Marano

Department of Biomedical and Clinical Sciences “L. Sacco”, University of Milan, Milan, Italy

Michele Zani

Fondazione Le Rondini Città di Lumezzane Onlus, Lumezzane (BS), Italy

Miriam Magri

Freelance nurse, Gorgonzola (MI) Italy

Annamaria Antonia Verardi

Health Care Agency, Milan, Italy

Aladar Ianes

Segesta2000, Korian Group Italy, Milan, Italy

Giuseppe Ardoino

Ippocrate Nursing Home, Korian Group Italy, Milan, Italy

Maria Chiara Gugiari

Freelance psychologist, Milan, Italy

Maura Lusignani

Department of Biomedical Sciences for Health, University of Milan, ASST Grande Ospedale Metropolitano Niguarda, Milan, Italy

Loris Bonetti

Ente Ospedaliero Cantonale, Department of Nursing, Bellinzona, Svizzera, University of Applied Sciences and Arts of Southern Switzerland (SUPSI), Department of Business Economics, Health and Social Care, Manno, Switzerland



How to Cite

Castaldo, A., Zanetti , E.S., Nobili, A., Marano, G., Zani, M., Magri, M., Verardi, A.A., Ianes, A., Ardoino, G., Gugiari, M.C., Lusignani, M. and Bonetti, L. 2022. Food intake and prevalence of malnutrition in nursing homes. A multicenter observational study. JOURNAL OF GERONTOLOGY AND GERIATRICS. (Jul. 2022), 1-14. DOI:
  • Abstract viewed - 96 times
  • PDF downloaded - 19 times