Malnutrition, Screening, Oral Nutritional Supplementation, Older Adult Introduction

Abstract

This study aimed to establish prevalence of malnutrition in older adult care home residents and investigate whether a nutritional screening and intervention program could improve nutritional and clinical outcomes. A community-based cohort study was conducted in five Newcastle care homes. 205 participants entered; 175 were followed up. Residents already taking oral nutritional supplements (ONS) were excluded from interventions. Those with ‘Malnutrition Universal Screening Tool’ (‘MUST’) score of 1 received dietetic advice and ≥2 received dietetic advice and were prescribed ONS (220ml, 1.5kcal/ml) twice daily for 12 weeks. Body mass index (BMI), ‘MUST’, mini nutritional assessment score®, mid upper arm muscle circumference (MAMC) and Geriatric Depression Scale (GDS) were recorded at baseline and 12 weeks. Malnutrition prevalence was 36.6%±6.6 (95% CI). A higher ‘MUST’ was associated with greater mortality (p=0.004). Type of intervention received was significantly associated with change in ‘MUST’ score (P<0.001); dietetic advice resulting in the greatest improvement. There were no significant changes in BMI (p=0.445), MAMC (p=0.256) or GDS (p=0.385) following the interventions. Dietitian advice may slow the progression of nutritional decline. In this study oral nutritional supplements over a 3-month period did not significantly improve nutritional status in malnourished care home residents.

Keywords

Malnutrition, screening, oral nutritional supplementation, older adult
Introduction

Malnutrition affects more than 3 million people in the United Kingdom, of whom approximately 93% live in the community (1). Within the community, older adult care home residents are particularly vulnerable to the effects of disease-related under nutrition. The prevalence of malnutrition in UK older adult care home residents has been reported in the range of 30% to 42% (2-4). The cause of malnutrition in older persons most commonly relates to disease associated reduction in dietary intake, including loss of appetite (5). Clinical consequences of malnutrition in older adults include falls, delayed recovery from infection and impaired wound healing as well as increased rates of hospital admission and healthcare utilization and depression (6-9)(10).

Few patients who are malnourished, or at risk of malnutrition, are identified, in the absence of a nutrition screening system (11). Screening tools such as the ‘Malnutrition Universal Screening Tool’ (MUST) have been validated across healthcare settings (12). Once identified, malnutrition can be managed by dietary advice to optimize oral intake as well as with the use of oral nutritional supplements (ONS). Significant weight increases and reduced mortality risk is reported following ONS in older people at risk of malnutrition across a variety of healthcare settings (hospital and community) (13, 14). However, ONS are expensive and adherence in older people is known to be poor (15).

Benefits associated with screening and use of ONS established in the research setting have led to the publication of guidelines on managing adult malnutrition in the community (16). However, there is little published evidence to the authors’ knowledge examining whether the results from research studies on which these guidelines are based can be successfully translated into practice in the context of clinical services. In particular, the effects of nutritional screening and intervention programs on clinical, nutritional and wellbeing outcomes in older adult community care home residents have yet to be established.

Due to the high prevalence of malnutrition evidenced above in this vulnerable group, the researchers decided to assess the extent of malnutrition in Newcastle upon Tyne care homes for older persons, to investigate whether a community based nutritional screening and intervention program could be successfully delivered and to assess the effectiveness of such a program in improving nutritional and clinical outcomes in these care home residents; specifically anthropometric measurements, mortality, hospital admission rates and depression scores.

Methods

This was a multi-centre, descriptive, cohort study with study intervention, which recruited residents from five care homes across Newcastle upon Tyne. It was conducted between January 2012 and June 2012. The study consisted of two phases: 1. Screening phase: Collection of baseline data from all eligible residents 2. Intervention phase: Those residents identified as malnourished via MUST screening in phase 1 received the study intervention and follow-up data were collected at the end of a 3 month intervention period.

‘MUST’ is a five step screening tool to identify adults who are malnourished or at risk of malnutrition (available from BAPEN Office, Redditch and online at It was chosen to use in this study as it is validated in community care home residents and is in widespread use in the UK including Newcastle upon Tyne. It incorporates height and weight measurements to calculate body mass index, unplanned weight loss and an acute disease effect score where a patient is acutely ill and there is likely to be no nutritional intake for more than five days. A BMI>20 and <5% weight loss in past 3-6 months is scored as ‘MUST’ 0; BMI 18.5-20 or unplanned weight loss 5-10% in past 3-6 months scored ‘MUST’ 1; BMI <18.5 or unplanned weight loss >10% in past 3-6 months is scored ‘MUST’ 2. BMI 18.5-20 and unplanned weight loss >10% past 3-6 months or BMI <18.5 and unplanned weight loss 5-10% past 3-6 months is scored ‘MUST’ 3; and BMI<18.5 and unplanned weight loss >10% past 3-6 months is scored ‘MUST’ 4.

The study was considered and approved as a service evaluation by the Chair of the National Health Service (NHS) National Research Ethics Service Newcastle and North Tyneside 1 Research and Ethics Committee.

Five care homes were invited to take part in the study. Inclusion criterion was that the home was located within the Newcastle upon Tyne Primary Care Trust (PCT) geographical area. They were selected at random from five different wards within the city by the PCT Newcastle Care Homes Project team. Selection from five wards was performed to minimize potential confounding by socioeconomic status. There were no exclusion criteria to care homes taking part in the study. All five care homes approached consented to take part in the study.

Inclusion criteria

Subjects within care homes were eligible for entry into the screening phase of the study if they were permanent residents in residential, nursing or Elderly Mentally Infirm (EMI) care, able to eat and drink and not receiving end of life care. Subsequently, subjects were eligible for inclusion to the intervention component to the study if they had either a body mass index (BMI) < 20 or BMI ≥ 20 but < 25 with weight loss > 5% in the 6 months prior to the study period (i.e. a ‘MUST’ score of 1 or more).

Exclusion criteria

Subjects were excluded from the screening phase of the study if they were i) receiving end of life care and with life expectancy less than 2 months, ii) short-term respite residents with stays less than 3 months, iii) unavailable due to admission to hospital at the time of recruitment or iv) declined to be weighed or have anthropometric measurements carried out on them. In addition to the above criteria, residents were excluded from the intervention phase of the study if they were already prescribed ONS, intolerant to any of the ingredients in the study supplements or required total parenteral nutrition or enteral feeds as their sole source of nutrition.

Study intervention

Residents were allocated to study interventions based on their baseline ‘MUST’ score. Those residents with a MUST score of 0, considered not to be at risk of malnutrition, received no intervention. Residents with a MUST score of 1 were allocated to receive advice from a community dietitian on strategies to increase oral intake including meal size, ways to increase protein and energy intake with normal food and prompting. A laminated dietary advice sheet summarizing these recommendations was left with each resident in their care plan to act as a future point of reference for staff. Residents with a MUST score of 2 or more at baseline were allocated to receive the same dietary advice as those residents with score of 1, but in addition to this they were prescribed ONS drinks (Abbott Ensure Plus® 220ml, 1.5 kcal/ml, complete, balanced ONS drinks twice per day for a period of 12 weeks). This ONS was used as it provided a nutritionally complete ready to drink supplement and was in use within the hospital setting locally at the time of the study. ONS were prescribed for consumption between mealtimes. ONS drinks were funded and delivered to the care homes by the project team. Residents receiving this intervention were given a starter pack of a variety of flavors of milkshake or juice style drinks for the first 2 weeks of the study. Thereafter they were given their preferred flavor and style of drink for the remainder of the study.

Measurements

Baseline data included residents’ age, gender and level of care (residential, nursing and elderly mentally infirm). Existing ONS use was noted. Anthropometric measurements taken included current weight, weight 3 months prior, body mass index, mid upper arm circumference (MUAC), triceps skinfold thickness (TSF) and calf circumference. MUST screening score and the full mini nutritional assessment score (Full MNA score) were calculated.

Ulnar length, mid upper arm circumference (MUAC) and calf circumference were measured on the non-dominant side over the largest bulk of the gastrocnemius muscle, 9 inches superior to the participants heel, using an anatomical tape measure (graduation 1mm, range 150cm) by the lead investigator. Weight was measured using SECA 955 chair scales (Class III medical device, capacity 300kg, graduation weight 20 g < 50 kg > 50 g < 150 kg > 100 g). The same set of scales was used for all subjects in the study and was calibrated prior to study commencement. Triceps skinfold thickness (TSF) was measured vertically along the tricep midline on the non-dominant side at the midpoint between the superior aspect of the acromial process (shoulder) and the inferior aspect of the olecranon process of the ulna (elbow). The lead investigator used the Harpenden skinfold callipers (Class 1A medical device, graduation 0.2mm, range 0-80mm), calibrated prior to study commencement. Mid upper arm muscle circumference (MAMC) was derived from MUAC and TSF measurements using the formula: MAMC = MUAC – π x TSF. Height was derived from ulnar measurement, using the conversion charts within the MUST toolkit. Height was calculated in this way because in this population group, where many individuals have mobility difficulties or postural changes. The lead investigator facilitated residents’ completion of the short-form geriatric depression scale questionnaire (17). It was recorded where cognitive impairment prevented study participants’ completion of the form.

Three months after commencement of the study intervention, the baseline measures were repeated. Where subjects failed to complete the intervention period this was recorded, along with the reason.

During the intervention, a datasheet was given to care home staff to record consumption of ONS during two separate weeks, week 3 and week 10. This represented a reasonable attempt to gauge compliance with the intervention, whilst not risking disengagement with the project by care home staff due to excessive documentation requirements. Hospital admission data were collected from Newcastle Upon Tyne Hospitals NHS Foundation Trust health records for the homes involved during the study period and comparison data were also collected for the same months’ admission data for the preceding year. Members of the project team completed timesheets, providing crude data for cost analysis.

Statistical analysis

Statistical analysis was carried out using the software package SPSS (Version 19, SPSS Inc. Chicago, Illinois, United States of America). Analysis was on an intention to treat basis (last value carried forward). Normality of the data was tested using the one-sample Kolmogorov-Smirnov test. Descriptive statistics were reported as mean scores with standard deviation for normally distributed data.

Student’s paired t-tests were performed to assess changes within each intervention group after the three-month study intervention. Differences between two groups were examined with the student’s independent t test for parametric data. One-way analysis of variance (parametric data) or Kruskal Wallis (non parametric data) was used when there were more than two groups. Chi-square analysis was used for comparing two categorical variables. Statistical significance was established at p <0.05.

Results

Screening

Two hundred and twenty five residents were identified living in 5 care homes, 205 met the inclusion criteria for the screening phase of the study. Figure 1 summarizes study recruitment and allocation to intervention groups.

Baseline characteristics

Mean age of resident included in the study was 84.2 years  8.5. Male to female ratio was 1:2.1. The proportion of each category of resident care setting were as follows: 37.6% residential, 28.3% nursing and 34.1% EMI. There were no statistically significant differences in age (P=0.365), sex (P=0.841) or care setting (P=0.059) between those residents excluded at baseline (n=20) compared to those entered into the screening phase of the study (n=205).

Results of baseline measures of nutritional status of residents included in the screening phase of the study are summarized in table 1. Mean MAMC measures for males and females were very similar to a reference population of mean age and sex matched individuals (study: males 232mm  33, females 206mm  30; reference: males 230mm, females 202mm) (18).

Using ‘MUST’ the prevalence of those at high risk of malnutrition within the study group was 36.6% (95% CI 30.0 to 43.2).

Twentynine (14%) residents included in the study were noted to have ONS already prescribed to them at the time of the baseline visit. Nine of these residents with ‘MUST’ scores of 0 had no ongoing indication for their use at that point in time. In those residents already prescribed ONS at baseline, only 18 out of 29 had been assessed by a dietitian. None of the other residents appeared to have been assessed by a dietitian.

A significant association between class of resident and ‘MUST’ score was noted (chi squared 15.76, df 4, P=0.003). The proportion of residents classed as ‘residential’ in relation to their care status significantly reduced as ‘MUST’ score increased. Conversely, the proportion of ‘nursing’ care residents significantly increased as ‘MUST’ score increased. No clear pattern emerged in relation to nutritional status and ‘EMI’ care status.

Mean age, BMI, MAMC, calf circumference, ‘MUST’ scores, Full MNA scoresand GDS scores were comparable between homes with no significant differences established. There was no significant difference in age of resident between ‘MUST’ score categories (P=0.173) or association between sex of resident and ‘MUST’ score (P=0.220). There was also no significant difference in GDS between ‘MUST’ categories (P=0.895).

Results following 12 week intervention

One hundred and seventy three residents took part in the intervention phase of the study. One hundred and fifty two of these (88%) provided follow-up data (figure 2). Despite 32 residents not being eligible for the intervention itself, 12-week follow-up data were collected for 23 of them.

Anthropometry

Table 2 summarizes changes in anthropometry. The only statistically significant changes over the study period were an increase in calf circumference following the 12-week intervention in the ‘no intervention’ group and ‘dietitian advice’ and a statistically significant reduction in BMI in the dietitian and ONS group.

‘MUST’ score

Table 3 summarizes the changes in ‘MUST’ scores that occurred in residents by intervention group. No change in ‘MUST’ score was the most common outcome across all intervention groups. A highly significant association was present between change in ‘MUST’ score and intervention received (chi squared 44.85, df 6, P0.001). Dietitian advice resulted in the greatest improvement in ‘MUST’ score. Dietitian and ONS intervention was associated with the largest proportion of residents experiencing deterioration in ‘MUST’ score over the study period.

Full MNA

There were no significant changes in Full MNA score following any of the interventions or between intervention groups (P=0.235).

Geriatric depression scale

There were no significant changes in GDS with any intervention at 12 weeks and no significant difference in change in GDS score between intervention groups (P=0.385).

Residents excluded from the study intervention

Follow-up data were collected on 23 out of 32 residents excluded from the intervention (22 existing ONS prescription at baseline and 1 declined dietitian intervention) (figure 2). There were no significant changes in anthropometric measures or GDS in these residents compared to baseline.

Compliance

Sixty five percent of ONS prescribed were offered and 83% of these were consumed. Fifty four percent of ONS prescribed were therefore consumed. Unpleasant taste, large volume and day-to-day fluctuations in cognitive functioning were cited in several instances as reason for reduced compliance with ONS.

Nutritional outcomes in residents compliant with oral nutritional supplements

Five out of 22 residents taking part in the dietitian advice and ONS group demonstrated compliance (consumption ≥75% of prescribed ONS). There were no statistically significant changes in nutritional status from baseline to the end of the study intervention in these residents.

Hospital admission rates

Hospital admission rates were studied in all residents completing baseline measurements, regardless of intervention. Most hospital admissions (70 out of 88) were of residents with a ‘MUST’ score of 0. Comparing the study period in 2012 with the same period in the preceding year, there were no statistically significant differences in mean number of admissions per care home (19.66.2 in 2011, 17.610.8 in 2012, P=0.537) or mean length of hospital admission (84.7 days in 2011, 11.44.5 days in 2012, P=0.205). Hospital lengths of stay and in-patient mortality were similar between the 5 study care homes and other care homes in the city for the periods analyzed.

Mortality

Mortality was investigated in all residents completing baseline measurements. There was a statistically significant difference in BMI between those who died, 21.1 kg/m2 5.5, and those who survived, 24.3 kg/m26.6, (p=0.022). There was also a significant association between baseline MUST and 12-week mortality during the study (p=0.004); Those with MUST of 0 or 1 had a mortality of 7.9% rising to 22.5% in those with a MUST of 2, 33.3% in those with a MUST of 3 and peaking at 50% in those with a MUST of 4.