Loading...
Advancing Malnutrition Screening, Detection, and Assessment of Community-Dwelling Older Adults in Outpatient Clinics
Sullivan, Austin M.
Sullivan, Austin M.
Citations
Altmetric:
Abstract
Background. Malnutrition is a multifactorial condition characterized by poor dietary intake that increases the risk for adverse health outcomes such as hospitalization, institutionalization, morbidity, and mortality. Older adults are disproportionately affected by malnutrition, having a two-to-three-fold greater prevalence rate than the general population. Most malnutrition diagnoses occur in the hospitalized setting, where the Joint Commission requires all patients admitted to the hospital to be screened for the condition within 24 hours. However, there is no requirement to administer malnutrition screening in the outpatient setting despite an estimated 14-26% prevalence among community-dwelling older adults. Therefore, further research is needed to improve outpatient malnutrition detection in older adults.
Methods. This dissertation leverages three studies to standardize and improve malnutrition detection in older adult outpatients. Study one aimed to compare the convergent validity of the Subjective Global Assessment (SGA) and Mini Nutrition Assessment (MNA) with baseline measures of function and nutrient status in persons with dementia (PWD) and compare the predictive validity of the SGA and MNA with 1-year malnutrition-related outcomes in PWD. This secondary data analysis was performed on PWD who were ≥ 65 years. PWD were assessed for malnutrition with the MNA and SGA by two separate dietitians who were blinded to the opposite assessment. Skin carotenoids (pressure-mediated reflection spectroscopy) and hand grip strength (hand dynamometer) were collected as two objective measures hypothesized to be associated with nutrition status. Chi-squared tests of independence were completed to determine the association between MNA and SGA, and Cohen’s kappa coefficient was used to assess the agreement of the tests. Spearman’s rank-order correlations were conducted for each tool (SGA and MNA) and baseline measures of hand-grip strength and veggie meter score. Time to completion was compared through a paired samples T-test. An exact Cox proportional hazards regression was conducted to determine if nutritional status classified by either the MNA or SGA predicted time to event for falls, institutionalization, and ER visits. Study two aimed to assess the prevalence of malnutrition in outpatient clinics at a Midwestern academic medical center in patients ≥ 65 years and develop a computable phenotype using the Global Leadership Initiative on Malnutrition (GLIM) to predict malnutrition diagnosis. This cross-sectional analysis was performed using a clinical data repository. Using the GLIM criteria as a framework, a computable phenotype was created and compared to ICD-10 malnutrition codes and classification by a registered dietitian upon chart review. This study included 8,872 older adults with ≥ two encounters at outpatient clinics between October 2015 and June 2024. Sensitivity, specificity, positive predictive value, negative predictive value, accuracy, balanced accuracy, and Cohen’s kappa coefficient were used to determine performance. Predictive validity for mortality was tested using Cox proportional hazard model analysis and visualized using Kaplan-Meier survival curves. Study three aimed to assess the feasibility of administering malnutrition screening in a memory care clinic and determine barriers and facilitators to malnutrition screening. The eight-item Seniors in the Community: Risk Evaluation for Eating and Nutrition (SCREEN-8) was administered to patients presenting for their appointments through two separate screening protocols for six weeks each (12 weeks total). The Organizational Readiness to Change Assessment (ORCA) tool was administered to clinic providers before screening. A nutrition care interest survey was developed and attached to the nutrition screener to obtain patient perspectives of nutrition care and screening. Field notes were used to record clinic staff reported barriers and facilitators to screening. Descriptive statistics were used to report results for the SCREEN-8 and ORCA. A thematic analysis by two separate researchers was conducted to summarize the findings of the open-ended question included in the nutrition care interest survey.
Results. A total of 46 participants were included in the convergent validity analysis, while 44 participants were included in the predictive validity analysis in Study 1. The MNA classified 31.1% of the population at risk and 4.4% malnourished, while the SGA classified 44.4% at risk and no one as malnourished. Chi-squared analysis showed that MNA and SGA were associated (χ2= 5.940, p=0.015) and continuous scores were moderately correlated (rs =0.429, p= 0.003). There was fair agreement (κ= .294) between the SGA and MNA. The MNA was moderately correlated to hand grip strength (rs =0.569, p< 0.001) and moderately correlated to veggie meter scores (rs =0.423, p= 0.004, n=44). The SGA was moderately correlated with handgrip strength (rs= 0.486, p< 0.001) and weakly correlated with veggie meter scores (rs = 0.299, p=0.048, n=44). SGA (M=7.22 min, SD=3.17) was performed slightly faster than the MNA (M=8.78 min, SD=3.17); [t(45) = 2.985, p = .005]. There was no association between nutrition risk measured with the MNA and time to falls (HR= 2.81, CI 0.753-10.54, p=0.124), institutionalization (HR=0.66, 0.173-2.475, p= 0.533), or ER visits (HR= 1.26, CI 0.300-5.256, p= 0.756). Similarly, there was no association between nutrition risk identified with the SGA and time to falls (HR= 1.08, CI 0.288-4.029, p= 0.911), institutionalization (HR= 1.461, CI = 0.445-4.819, p= 0.53), or ER visits (HR 4.44, 0.896-22.01, p= 0.068). Study 2 found that the computable phenotype displayed good specificity (96%) but poor sensitivity (69%) compared to ICD-10 malnutrition codes. The two methods had a fair agreement (κ=0.28). When compared with a dietitian’s classification of malnutrition, the computable phenotype had good sensitivity (95%) and specificity (82%) and displayed substantial agreement (κ= 0.76) with the dietitian’s malnutrition classification. ICD-10 malnutrition codes (HR= 3.22, CI 2.15-4.82, p < 0.001) and computable phenotype identified malnutrition (HR= 2.85, CI 2.25-3.59 p < 0.001) were associated with an increased all-cause mortality risk. The computable phenotype (4.78%) identified more older adults as malnourished than ICD-10 codes (1.27%). Study 3 displayed a malnutrition screening penetration rate of 15.3%, with 233 patients of the 1,522 total patients being screened for nutrition risk, while 22 of collected screeners could not be scored. Of those screened, 58.3% (n=123) were at high nutrition risk. Lack of staffing was the significant barrier identified during the study. Clinic staff responses to the evidence domain items of the ORCA indicated moderate agreement (mean 3.7, SD= 0.26), while the context domain responses indicated strong agreement for evidence of organizational support (mean =4.12, SD = 0.38). A total of 69 patients (33.7%) indicated interest in receiving a nutrition consultation. Difficulty accessing the clinic and unwillingness to use nutrition services were the two major themes identified that would prevent patients from accessing the clinic’s nutrition care. Most patients reported that nutrition screening was easy (n= 79/211), or extremely easy (n=44).
Conclusions. Studies one and three found a high degree of nutrition risk in community dwelling older adults with and without dementia. To diagnose malnutrition in community dwelling older adults with dementia, the SGA and MNA performed comparably. A computable phenotype was a valid method to enhance malnutrition identification. Malnutrition screening is feasible in an outpatient memory care clinic, but staff time and other commitments make sustainability challenging. Future studies should continue to improve malnutrition identification and further validate nutrition assessment methods in older adults with and without dementia.
Description
Date
2025-01-01
Journal Title
Journal ISSN
Volume Title
Publisher
University of Kansas
Collections
Archive Status
This item contains archived web content.
Files
Sullivan_ku_0099D_19990.pdf
Adobe PDF, 1.46 MB
- Embargoed until 2176-05-31
Research Projects
Organizational Units
Journal Issue
Keywords
Nutrition, Assessment, Community-Dwelling, Informatics, Malnutrition, Screening