Factors associated with malnutrition in nursing home residents: A systematic review and meta-analysis of observational studies
Article information
Abstract
Purpose
We investigated factors associated with malnutrition in nursing home residents and estimated pooled effect sizes of those associations.
Methods
We report this study according to the PRISMA checklist 2020. After registering the protocol, we conducted a systematic search using the keywords “malnutrition,” “nursing home,” and “factor” on PubMed, EMBASE, CINAHL, and Scopus and conducted a hand search of relevant sources in November 2023. We conducted the meta-analysis using R4.3.1 software.
Results
We included 46 studies and entered 35 studies into the meta-analysis. Meta-analysis results showed that (1) demographic factors included being female: pooled odds ratio (pOR) 1.38 (95% confidence interval [95% CI], 1.29~1.47) and higher body mass index: pOR 0.74 (95% CI, 0.66~0.83); (2) health functional factors including dependence with activities of daily living: pOR 3.70 (95% CI, 2.97~4.60) and being immobile: pOR 2.50 (95% CI, 1.39~4.47); (3) eating and oral intake factors including: eating <50% of offered food portions: pOR 3.56 (95% CI, 2.05~6.19) and loss of appetite: pOR 3.60 (95% CI, 1.36~9.57); (4) disease-related factors including dementia: pOR 1.98 (95% CI, 1.59~2.46), depression: pOR 1.84 (95% CI, 1.18~2.85), diabetes: pOR 0.76 (95% CI, 0.58~0.99), and pressure ulcers: pOR 2.14 (95% CI, 1.54~2.97); and (5) medication-related factors such as taking polypharmacy pOR 1.75 (95% CI, 1.24~2.47) were found to have a significant association with malnutrition in nursing homes.
Conclusion
These results contribute to updating knowledge about factors associated with malnutrition in nursing homes. Results supported that demographic status, health function, eating and oral intake, and disease, and medication factors significantly aligned with malnutrition. Study results contribute to improving nutritional care in nursing homes and long-term-care settings.
INTRODUCTION
Nutrition plays a crucial role in human health and well-being [1]. Malnutrition, an adverse nutritional outcome, has been defined as a body state which “results from a lack of consumption of nutrition that alters body composition (decreased fat free mass) and body cell mass, thereby leading to reduced physical and mental functioning and impaired clinical outcomes due to disease” [2]. Compared to younger adults, older adults are at a higher risk of malnutrition [3].
Malnutrition is highly prevalent in older adults worldwide [4,5] and is expected to increase with the rise of the aging population. However, the prevalence of malnutrition among older adults is significantly different across settings, with a significantly higher prevalence observed in nursing homes and long-term care settings (17.5% and 29.4%, respectively) compared to the community (3.1%), outpatient (6.0%), and homecare (8.7%) settings [6].
Nursing homes for older adults are facilities that provide 24-hour care and support, particularly for individuals requiring assistance with activities of daily living and other health-related needs. These facilities provide long-term care, rehabilitation services, and, in some cases, palliative or end-of-life hospice care [7]. Terminology varies and is used interchangeably across countries, such as nursing homes, residential homes, long-term care facilities, aged care homes, skilled nursing facilities, and assisted living facilities [7]. In the present study, “nursing home” encompasses all the institutional settings that meet this definition.
Nursing home residents, who are characterized by advanced age and frailty, often experience reduced appetite, chewing difficulties, and altered body composition. Additionally, their reliance on institutional food provision may lead to dissatisfaction with the meals since institutes do not consider individual preferences. Despite controlled macronutrient consumption, these factors, along with socioeconomic disadvantages and inadequate staff expertise in gerontological nutrition, can contribute to suboptimal nutritional outcomes in this vulnerable population [8].
Malnutrition negatively impacts older adults’ health, quality of life, and mortality, which are commonly observed consequences among nursing home residents [9]. Additionally, malnutrition increases the burden of care costs [10], thereby increasing the burden on healthcare, social care, and aged care systems [11]. Therefore, it is crucial to determine the factors associated with malnutrition in nursing homes to implement appropriate interventions and prevent malnutrition and its complications.
The literature review revealed two review studies on the factors associated with malnutrition in nursing homes [9,12]. They utilized narrative synthesis and reported demographic factors, health-related factors, eating/oral consumption, disease-related factors, medication-related factors, and nursing home-related factors such as low financial resources for food, social isolation, and small-sized facilities. The most recent systematic review on malnutrition among nursing home residents was conducted over a decade ago [12]. Since then, several updates and newly identified factors have emerged, thereby expanding the scope of research across multiple domains, particularly in the nutritional domain. With advancements in economic development, living standards, and high-quality healthcare, nutritional care has evolved beyond meeting calorie intake requirements. Current approaches emphasize nutrient balance [13] and are increasingly focusing on micronutrient intake, including minerals and vitamins [14]. Furthermore, as the proportion of globally aging population continues to increase, the demand for nursing home care is also increasing, thereby leading to a greater attention on institutional factors influencing nutrition. These include structural support during mealtimes, staff awareness of nutrition and malnutrition, and the administration of food and nutritional care [15]. Therefore, a systematic review is required to update data.
Moreover, despite a growing body of research, the associations between malnutrition risk factors remain inconsistent across studies, countries [16-18], and sample sizes [19,20]. Therefore, a meta-analysis is necessary to calculate the effect estimate of each factor to obtain a comprehensive view in a large sample. Thus, the present study identified additional risk factors and quantified the effect estimates of the associations, where possible, through a systematic review and meta-analysis.
METHODS
This study was approved by the Institutional Review Board (IRB) of Keimyung University (IRB No 40525-202310-HR-04401). Informed consent was obtained from the participants.
1. Study Design
The present study employed a systematic review and meta-analysis design. The study protocol was registered in PROSPERO (ID: CRD42023425403). The study followed the guidelines outlined in the JBI Manual for Evidence Synthesis for Systematic Reviews of Etiology and Risk [21], and was reported according to the PRISMA guidelines and statement 2020 [22].
2. Data Sources and Study Selection
The selection criteria for this study was defined using PEO (population, exposure of interest, and outcome) based on the JBI Manual for Evidence Synthesis for Systematic Reviews of Etiology and Risk. Therefore, population (P) involved older adults residing in nursing homes, exposure of interest (E) included demographic factors, health-related factors, nutritional factors, eating and oral consumption factors, disease, medication, nursing home facility, staff, and caring process, and outcome (O) focused on malnutrition.
The reviewer (TTTN) searched four databases: PubMed, EMBASE, CINAHL, and Scopus. The search strategy was determined according to the Medical Subject Headings term and manual words. The keywords included “malnutrition,” “nursing home,” and “factor” using Boolean connector words OR and AND (Supplementary Material 1). The reviewer (TTTN) manually searched references and citations from the included studies, academic nutritional journals, and Google Scholar databases to identify more relevant studies. After the systematic search, duplicate studies were removed using EndNote21 (Clarivate). Two reviewers (HP and TTTN) independently screened the titles and abstracts. Subsequently, the reviewers (HP and TTTN) downloaded and independently screened the full texts of the remaining studies. Studies were included according to the inclusion criteria. The reviewers (HP and TTTN) discussed their decisions and resolved any inconsistencies.
The inclusion criteria were: (1) the study was conducted in a nursing home or multiple settings in which the results of the nursing home were reported separately; (2) outcome was malnutrition, which was measured using validated assessment tools, such as Subjective Global Assessment [23] and Mini Nutritional Assessment full form [24] and short form [25], or included malnutrition criteria according to the nutritional academic society, such as European Society for Clinical Nutrition and Metabolism [26]; (3) the study reported data regarding the association between malnutrition and another variable as odds ratio (OR), hazard ratio, prevalence ratio, and relative risk; (4) the study design was quantitative observational study, cohort study, cross-sectional study, and/or case-control study. Other inclusion criteria were original academic article written in English language and published in a peer-reviewed journal between month date, year and November 1st, 2023 (search day).
The exclusion criteria were: (1) the study was conducted in a child-care nursing home or short-term aged care setting such as acute care clinical setting, free-living community, or daycare center; (2) the study only utilized body mass index (BMI) cut-off points without another criteria, such as weight loss or arm or calf circumferences, which can result in inaccurate weight and height measurements of older adults due to potential age-related changes in their body composition [27]; (3) the study considered serum albumin as a marker of malnutrition since it is highly influenced by inflammation in older adults [11]; (4) lack of clarity regarding the method used to assess malnutrition; (5) duplicate data or in case there are more than two data from the same population, the largest sample size’s data will be selected. Other exclusion criteria were unavailability of the full text despite contacting the corresponding author and if the study was a qualitative study, intervention study, psychometric evaluation study, review study, case report, letter, editorial, or conference abstract.
3. Study Quality Assessment
The quality of the selected studies were analyzed using the Joanna Briggs Institute Critical Appraisal Checklist version 2020 [21]. The analytical cross-sectional study checklist and cohort study checklist included eight and eleven statements, respectively. Each statement was judged by “Yes,” “No,” “Unclear,” or “Not applicable” response according to the checklist guide. The response “Yes” is scored as 1, “No” and “Unclear” are scored as 0, and “Not applicable” is not scored [21]. The percentage is calculated as the Number of Yes/(Number of Yes+Number of No+Number of Unclear)×100%. A percentage of >75% is identified as high-quality, 50%~74% as medium quality, and <50% as poor quality [28]. Two reviewers independently assessed the study quality and resolved any uncertainty in scoring through discussion.
4. Data Extraction
Data were extracted using a pre-designed extraction form, including first author’s name, publication year, study location, type of study design, sample size, average age or age range, proportion of females, assessment instrument/tool/criteria, prevalence or proportion of malnutrition, associated factors, and ORs with confidence intervals (CIs). Data were extracted by reviewer TTTN. Subsequently, two reviewers (HP and TTTN) discussed the extracted results to reach consensus.
5. Data Analysis
ORs and their CI values were used in the meta-analysis to estimate the pooled effect size when there were at least two data points in one factor. Heterogeneity was assessed using Cochran’s Q statistic and I2 index. In the present study, a Q statistic p-value <0.05 and/or an I2 value >50% was considered as significant heterogeneity across studies [29]. A random-effects model was applied if the data showed heterogeneity and a fixed-effects model was used if there was no significant heterogeneity. Publication bias is evaluated when the analysis includes 10 and more studies. An asymmetrical funnel plot or a p-value <0.05, as determined by Egger’s test, was considered evidence of significant publication bias. Statistical analyses were performed using R4.3.1 software.
RESULTS
1. Description of Included Studies
The database search yielded 9,665 studies. After removing duplicates and screening titles and abstracts, 479 studies remained. The full-text review excluded 428 articles owing to criteria such as setting, design, and outcome. Of the 9,665 studies, 40 studies were found to be eligible, and six studies were added via manual searches. Finally, 46 studies were selected, of which 35 were used for the meta-analysis (Figure 1).
Flow diagram of study screening and selection.
The 46 included studies [13-20,30-67], published between 2001~2023, comprised 45 cross-sectional and three cohort components (two studies contained both designs) [58,63]. The cross-sectional studies (94.92%) were distributed across Europe (21 studies), Asia (15 studies), the United States (five studies), and Australia (four studies). The average age of nursing home residents ranged from 72.2±8.8 to 86.8±7.8 years. Additionally, the proportion of female residents was more than male residents in all studies and ranged between 50.4% and 85.7%. The sample sizes ranged from 52 to 19,876 residents. Malnutrition was predominantly assessed using the Mini Nutritional Assessment (MNA, n=20), followed by anthropometrics (n=12), Mini Nutritional Assessment short form MNA-SF (n=9), and other tools, including the Subjective Global Assessment (SGA) (n=3) and Malnutrition Universal Screening Tool (MUST) (n=1). In the three cohort studies (2022~2023), the average age of residents was 83.6±7.0 years, ranging between 65 and 107 years. Additionally, the proportion of female residents ranged between 68.1% and 71.3%, and anthropometric measures were utilized for assessing malnutrition. The sample sizes ranged from 3,836 to 11,923, with follow-up durations varying from six months to approximately one year.
The prevalence of malnutrition ranged between 2% and 53%, which was measured using nutritional assessment tools and anthropometric measurements. Demographic factors showing significant associations with malnutrition were reported in 29/48 studies (60.4%). Health-function-related factors were reported in 31/48 studies (64.6%). Eating and oral consumption-related factors were reported in 19/48 studies (39.6%). Nutritional factors were reported in 3/48 studies (6.3%). Disease-related factors were reported in 20/48 studies (41.7%). Medication-related factors were reported in 4/48 studies (8.3%). Nursing home facilities, staff, and caring process factors were reported in 9/48 studies (18.8%) (Supplementary Material 2, 3).
Compare to previous review studies, the present study identified several novel factors influencing malnutrition among nursing home residents. Demographic factors included educational level, number of children, race/ethnicity, and employment status before nursing home admission [35,43-45]. Medicines with significant influence on outcome included angiotensin II receptor blockers, calcium channel blockers, atypical antipsychotics, anti-Parkinson drugs, antihypertensives, lipid-lowering agents, laxatives, and analgesics [65]. Other newly summarized factors included tongue pressure, peak expiratory flow, and lip-closure ability, whereas oral dryness, nausea, tongue strain, and plaque index were not significant [53]. Nutritional factors such as food consumption, vitamin and mineral supplementation, and fruit/vegetable consumption were significantly associated with malnutrition [14,52], whereas structural mealtime support, staff awareness, and food administration processes were not significant (Supplementary Material 2, 3) [15].
2. Quality Appraisal Results
Of the 45 cross-sectional studies, 20 (44.4%), 24 (53.3%), and 1 (2.3%) studies were of high, medium, and low quality, respectively. All the three cohort studies (100%) were of medium quality. Low quality scores were accorded due to unclear inclusion criteria, inadequate description of study participants and settings, and unreliable measurements.
3. Meta-Analysis
The meta-analysis results revealed that higher BMI (pooled OR [pOR]=0.74; 95% CI, 0.66~0.83) and diabetes (pOR=0.76; 95% CI, 0.58~0.99) were associated with decreasing odds of malnutrition. All the other remaining significant factors were associated with higher odds of malnutrition among nursing home residents. These included older age (pOR=1.02; 95% CI, 1.01~1.03), being female (pOR=1.38; 95% CI, 1.29~1.47), dependence in activities of daily living (pOR=3.70; 95% CI, 2.97~4.60), supported eating (pOR=1.71; 95% CI, 1.43~2.04), and immobility (pOR=2.50; 95% CI, 1.39~4.47). In eating and oral consumption domain, eating half or less than half of the food offered (pOR=3.56; 95% CI, 2.05~6.19), loss of appetite (pOR=3.60; 95% CI, 1.36~9.57), chewing problems (pOR=1.42; 95% CI, 1.15~1.76), and swallowing problems (pOR=1.71; 95% CI, 1.21~2.42) were associated with higher odds of malnutrition. Considering diseases, presence of digestive tract disease (pOR=1.31; 95% CI, 1.19~1.43), constipation (pOR=1.81; 95% CI, 1.36~2.41), respiratory disease (pOR=1.18; 95% CI, 1.07~1.30), dementia (pOR=1.98; 95% CI, 1.59~2.47), depression (pOR=1.84; 95% CI, 1.18~2.85), and cancer (pOR=1.23; 95% CI, 1.09~1.38) were associated with higher odds of malnutrition. Additionally, the presence of pressure ulcers was a risk factor of malnutrition (pOR=2.14; 95% CI, 1.54~2.97). Other factors included multiple diseases (pOR=1.03; 95% CI, 1.02~1.05) and polypharmacy (pOR=1.75; 95% CI, 1.24~2.47) (Table 1, Supplementary Material 4).
Meta-Analysis Results of Factors Associated with Malnutrition
4. Publication Bias
The results of asymmetrical funnel plot and Egger’s test (p>0.05) revealed an absence of publication bias in factors such as age and sex (Supplementary Material 5).
DISCUSSION
This systematic review investigated the factors associated with malnutrition among nursing home residents. Malnutrition can be measured using assessment tools and anthropometric measurements. The assessment tools included the MNA, MNA-SF, MUST, and SGA, which were originally developed for older adults. However, there are concerns about the appropriateness of using these tools in nursing home settings and the lack of evidence regarding their validity concerning the true nutritional status of nursing home residents [27]. Anthropometric measurements included weight, height, BMI, weight loss, and arm or calf circumference. However, this method has limitations related to changes in body composition and discrepancies in weight and height measurements among older adults [68,69]. Therefore, it is necessary to have a “gold standard” assessment tool for measuring malnutrition in nursing home settings.
This systematic review identified a broader range of novel factors that influence malnutrition among nursing home residents than previous reviews. The increasing emphasis on resident-centered care has resulted in a more comprehensive examination of demographic characteristics [70]. Factors such as educational level, number of children, race or ethnicity, employment status before admission to a nursing home, and financial status were significantly associated with the risk of malnutrition [39,47-49]. Additionally, a wider variety of medications were investigated, with significant associations found for angiotensin II receptor blockers, calcium channel blockers, atypical antipsychotics (e.g., risperidone), anti-Parkinson drugs, antihypertensive drugs, lipid-lowering agents, laxatives, and analgesics [65]. Therefore, these findings indicate the need for personalized nutritional strategies that consider the diverse backgrounds of nursing home residents.
In the eating and oral consumption domain, the present review identified new factors that have not been extensively discussed in previous reviews. Factors such as tongue pressure, peak expiratory flow, and lip-closure ability were significantly associated with malnutrition, whereas oral dryness, nausea, tongue strain, and the plaque index showed no significant association [53]. Additionally, self-reported hunger, dietary intake relative to recommended allowances, vitamin and mineral supplementation, and fruit and vegetable consumption were associated with malnutrition [14,52]. Although structural support during mealtime, staff awareness, and food administration were not significantly associated with malnutrition [15], their inclusion underscores the complexity of nutritional care in nursing homes. A multidisciplinary approach involving residents, caregivers, nutritionists, kitchen staff, and healthcare providers is essential to improve nutritional outcomes among older adults.
The meta-analysis revealed that older age and female sex were significant demographic factors. These results are consistent with results of previous meta-analyses among community-dwelling older adults [71]. Appropriate nutritional care for different age groups should be considered owing to the increasing numbers of older adults admitted to nursing homes [72]. In the eating and oral consumption domain, loss of appetite had the strongest association with higher odds of malnutrition among nursing home residents. This result is consistent with previous evidence among community-dwelling older adults [73]. Since nursing homes residents are not involved in the nutrition process, they may consider the food as unappetizing or the food service is inadequate. Poor communication among residents and between residents and staff due to lack of time and resources may result in unfinished tasks [74]. Additionally, the high prevalence of mental health concerns among residents may lead to reduced appetite [75].
Regarding health-related problems, dependency in activities of daily living and eating dependency or the need for assistance while eating were associated with malnutrition. These results are consistent with previous evidence of a significant association between low nutritional status and decreased activities of daily living among older adults [76]. Furthermore, immobility was associated with malnutrition in nursing homes. This result is supported by previous evidence that loss of muscle mass and weight due to swallowing problems, constipation, pressure ulcers, and an increased incidence of new diseases [77,78] can indirectly contribute to malnutrition.
Moreover, dementia was associated with higher odds of malnutrition in nursing home residents in the present study. This result is consistent with that of a previous study among community-dwelling older adults [79]. The pooled dementia prevalence is also high in nursing homes. More than half (53%) of the residents suffer from dementia, which is significantly higher than that of older adults living in their own homes [80]. More attention should be paid to nutritional care and dementia management among residents with dementia. Conversely, diabetes significantly reduced the odds of malnutrition in nursing homes. Evidence has shown that regular care provided by nursing homes to diabetic residents positively affects diabetes management and residents’ health outcomes, including nutrition [81]. Among medication-related factors, polypharmacy was associated with malnutrition in nursing homes, which is consistent with a previous study among older adults [82]. An increasing number of residents take polypharmacy [72]; therefore, a well-designed nutritional care plan through collaboration between pharmacists and nutritionists should be considered.
This study had several limitations. First, the observational nature of included studies restricted the identification of causal relationships. Since most studies were analytical cross-sectional, there is a need for more cohort studies, which is the preferred design for assessing risk factors. Most of the included studies had moderate quality, while one had low quality due to unclear inclusion criteria, inadequate description of study participants and settings, and unreliable measurements. Future studies should address these issues to provide high-quality evidence. Moreover, although previous studies published in reliable journal have used the JBI appraisal checklist, there is no available appraisal tool that comprehensively addresses all potential biases in cross-sectional studies. The development of a comprehensive tool to assess the risk of bias in observational studies is required. Moreover, some factors resulted in high heterogeneity across studies. This was addressed by employing the random-effects model to ensure appropriate results. Some important nursing home-related factors, such as daily food budget, presence of nutritional experts, or guidelines related to malnutrition in nursing homes, were not included in the meta-analysis because of the number of included studies. Fourth, we only included studies written in English; therefore, relevant studies published in other languages were excluded. In the present study, we did not include PsycINFO as a data source. Although we manually searched for relevant sources, the omission of PsycINFO might have led to a loss of studies on psychological factors influencing malnutrition among nursing home residents. Future studies should include this data source to obtain a more comprehensive understanding of the psychological and cognitive aspects of malnutrition. Lastly, the assessment of malnutrition differed across studies. Therefore, a “gold standard” assessment tool is required to measure malnutrition, especially among nursing home residents.
The results of this study will contribute to nutritional care in nursing and inform effective interventions to prevent malnutrition and its consequences in vulnerable residents and minimize the consumption of nursing home sources for future studies. The association of advanced age and female sex with malnutrition among nursing home residents emphasizes the need for enhanced nutritional care in the target groups. Promoting residents’ involvement in food selection and fostering social interactions during meals may enhance oral consumption and eating behaviors, and reduce the risk of malnutrition. Regular nutritional screening is crucial for residents requiring assistance with their daily activities or eating. Individuals with immobility or dementia require specialized nursing and dietary interventions to maintain their health. Furthermore, the increasing prevalence of polypharmacy among older adults necessitates nursing home policies that mitigate its impact on their nutritional status and overall well-being.
CONCLUSION
This systematic review and meta-analytic study examined factors associated with malnutrition among older adults in nursing homes. It determined significant factors across demographic, functional, dietary, disease-related, and medication domains. Significant factors associated with malnutrition included dependency in activities of daily living, immobility, loss of appetite, consumption of less than half of the food offered, dementia, depression, and pressure ulcers. Thus, these results provide crucial evidence for enhancing nutritional care strategies in nursing homes and long-term care settings and developing targeted interventions to reduce the risk of malnutrition among residents.
Notes
Authors' contribution
Study concept and design - TTTN and HP; Acquisition of data - TTTN and HP; Data analysis and interpretation - TTTN and HP; Manuscript preparation - TTTN and HP.
Conflict of interest
No existing or potential conflict of interest relevant to this article was reported.
Funding
None.
Data availability
Please contact authors in reasonable data sharing requests.
Acknowledgements
This study is modified from the Nguyen Thi Thu Thao’s master thesis of Keimyung University.
Supplementary materials
Supplementary Material 1.
Search Strategy
Supplementary Material 2.
General Characteristics of Included Cross-sectional Studies
Supplementary Material 3.
General Characteristics of Included Cohort Studies
Supplementary Material 4.
Forest Plots of Meta-analysis Results
Supplementary Material 5.
Funnel Plots