Food intake and prevalence of malnutrition in nursing homes. A multicenter 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 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:
- What was the prevalence of malnutrition among nursing homes’ residents?
- Which demographic and main clinical characteristics were associated with malnutrition?
- 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.
SETTING AND PARTICIPANTS
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.
DATA COLLECTION AND MEASURES
To measure the prevalence and factors associated to malnutrition, data regarding:
- socio-demographic characteristics data of residents (gender, age, length of stay);
- 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;
- nutritional conditions: chewing and swallowing problems and body mass index (BMI);
- NH characteristics such as ownership, facility size and unit type;
- 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.
OUTCOME, SAMPLE SIZE AND DATA ANALYSIS
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.
RESIDENTS CHARACTERISTICS, MALNUTRITION PREVALENCE AND FOOD INTAKE
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).
FACTORS ASSOCIATED WITH MALNUTRITION
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.
CONFLICT OF INTEREST STATEMENT
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)|
|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|
|(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|
|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)|
|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)|
|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)|
|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|
|Missing*||3 (1.1)||8 (0.5)||11 (0.6)|
|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|
|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)|
|Strength of association: OR (95% C.I.)||LR test: Chi-square (df)||P-value|
|Age *||1.48 (1.25-1.75)||22.128(1)||< .001|
|length of stay||3.659(3)||.30|
|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|
|Facility size||17.344(2)||< .001|
|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)|
|Pressure ulcers||14.605(1)||< .001|
|One or more||2.07 (1.45-2.96)|
|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|
|Swallowing difficulty||33.384(1)||< .001|
|Feeding assistance (Barthel Index)||58.630(4)||< .001|
|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|
|Food intake at dinner||10.753(3)||.013|
|Feeding assistance, lunch, day 1||39.085(3)||< .001|
|Feeding assistance, dinner, day 1||47.919(3)||< .001|
|Strength of association: OR (95% C.I.)||LR test: Chi-square (df)||P-value|
|< 80 years||1|
|≥ 80 years||1.99 (1.19-3.34)|
|Pressure ulcers||5.49 (0)||.019|
|Fluid intake* (ml)||1.14 (1.05-1.24)||10.26(1)||.001|
|Feeding assistance, dinner||8.42(1)||.003|
SUPPLEMENTARY FILE “PREVALENCE OF MALNUTRITION AND FOOD INTAKE IN NURSING HOMES. A MULTI-CENTER OBSERVATIONAL STUDY”
DEFINITION OF SAMPLE SIZE AND NHS RESIDENTS’ INVOLVEMENT
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).
ADDITIONAL INFORMATION ON STATISTICAL ANALYSIS PLAN
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:
- Gender (M/F);
- Age (codified as categorical variables with two modalities: ≤ 80 years; > 80 years;
- NH ownership (public, private);
- NH size (40-90 residents; 91-160 residents; > 160 residents);
- Barthel Index score (0-90; 91-100);
- Mental state by CIRS (1-3; 4-5);
- Pressure ulcers (None, present);
- chewing difficulty (No/YES);
- swallowing difficulty (No/YES);
- assisted nutrition by Barthel Index (low = minimal or no assistance; high = partial or constant or total assistance);
- daily total liquids intake;
- daily food intake at lunch (0-50; 51-100);
- daily food intake at dinner (0-50; 51-100);
- assisted nutrition at lunch at day 1 (low = minimal or no assistance; high = partial or constant or total assistance);
- 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:
- 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;
- in multivariate analysis, to assess the association between nutritional status and each of the variables included in the model.
- 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|
|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)|
|Small||226 (82.2)||1300 (86.0)||1526 (85.4)||0.12|
|Medium/high||49 (17.8)||212 (14.0)||261 (14.6)|
|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§|
- Chen CCH, Schilling LS, Lyder CH. A concept analysis of malnutrition in the elderly. J Adv Nurs. 2001; 36:131-142. DOI
- Bell CL, Lee ASW, Tamura BK. Malnutrition in the nursing home. Curr Opin Clin Nutr Metab Care. 2015; 18:17-23. DOI
- 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
- 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
- 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
- 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
- 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
- Cederholm T, Bosaeus I, Barazzoni R. Diagnostic criteria for malnutrition – an ESPEN Consensus Statement. Clin Nutr. 2015; 34:335-340. DOI
- 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
- 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
- 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
- Willis H. Causes, assessment and treatment of malnutrition in older people. Nurs Older People. 2017; 29:20-25. DOI
- 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
- 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
- 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
- Edwards D, Carrier J, Hopkinson J. Assistance at mealtimes in hospital settings and rehabilitation units for patients (> 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
- 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
- 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
- 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
- Lusignani M, Mari D. Evaluation of the nutritional state of the elderly. Eur J Intern Med. 2013; 24:e11-e12. DOI
- van Zwienen-Pot JI, Visser M, Kuijpers M. Undernutrition in nursing home rehabilitation patients. Clin Nutr. 2017; 36:755-759. DOI
- Donini LM, Neri B, De Chiara S. Nutritional care in a nursing home in Italy. PLoS One. 2013; 8:e55804.
- 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.
- Shah S, Vanclay F, Cooper B. Improving the sensitivity of the Barthel Index for stroke rehabilitation. J Clin Epidemiol. 1989; 42:703-709. DOI
- Zanetti E. Carrocci & Faber: Rome, Italy; 2003.
- 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
- Frisoni GB, Rozzini R, Bianchetti A. Principal lifetime occupation and MMSE score in elderly persons. J Gerontol. 1993; 48:S310-S314.
- 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
- 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.
- 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.
- Harrell FE. Springer: New York, NY; 2015.
- R Core Team. R: a language and environment for statistical computing. 2017.
- 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.
- 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
- 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
- Little MO. Nutrition and skin ulcers. Curr Opin Clin Nutr Metab Care. 2013; 16:39-49. DOI
- 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
- 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
- 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
- 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
- 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
- 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
- Vucea V, Keller HH, Ducak K. Interventions for improving mealtime experiences in long-term care. J Nutr Gerontol Geriatr. 2014; 33:249-324. DOI
- Volkert D, Beck AM, Cederholm T. ESPEN guideline on clinical nutrition and hydration in geriatrics. Clin Nutr. 2019; 38:10-47. DOI
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
© JOURNAL OF GERONTOLOGY AND GERIATRICS , 2022
How to Cite
- Abstract viewed - 289 times
- PDF downloaded - 46 times