Effect of family style mealtimes on quality of life, physical performance, and body weight of nursing home residents: cluster randomised controlled trial

Introduction

Residents of nursing homes face not only physical deterioration but also loss of independence, privacy, their spouse, and a familiar environment. These factors lead to a high prevalence of loneliness and depression and a low perceived quality of life.1

Mealtimes in nursing homes provide an opportunity to integrate and implement physical care with measures to improve quality of life. Food and nutrition are essential components of “the good life.”2 Mealtimes also enable residents to socialise with staff and other residents, to make choices according to personal preferences, and to relax. A convivial and social environment at mealtimes might add a sense of security, meaning, order, and structure to the day and improve satisfaction with life.23

In a pilot study we showed that improving the social and physical ambience during mealtimes in Dutch nursing homes counteracted a decline in reported quality of life of the residents.4 Residents with dementia were less irritated, anxious, agitated, and depressed and increased their participation and communication during mealtimes when food was served in family style or when music was played.5^–7 Studies in this specialty are characterised by small sample sizes (six to 29 participants),4^–7 so the results need to be interpreted cautiously and should not be generalised on a larger scale.

In most nursing homes, meals are individually served on trays in a non-stimulating social environment. Such meals provide task oriented care rather than resident oriented care.8 In our study we offered residents family style mealtimes (see table 1[t1]). We hypothesised that, during a six month study period, quality of life, physical performance, and body weight would remain stable among residents offered family style mealtimes but would decline in those receiving the usual pre-plated service.

Methods

In most Dutch nursing homes two types of care are available: psychogeriatric care for residents with dementia or chronic somatic care for patients with conditions such as stroke or Parkinson's disease. These two groups of residents live in separate wards, with an average of 30 residents per ward. A traditional ward has three or four single rooms, four to six double rooms, and four dormitories for four people each. Many Dutch nursing homes have planned major reorganisations to their infrastructure to offer residents their own room and better care services. These services may include family style mealtimes.

Results

Overall, 250 of 282 invited residents took part in the study. We excluded five residents: one was partially comatose, two were terminally ill, and two were switched to parenteral nutrition. During the study period 34 residents died, 11 moved to another institution or ward, 19 were discharged, and three withdrew consent. Overall, 178 residents completed the study. The intervention group consisted of 95 residents and the control group 83 residents (figure). The residents who did not complete the study had similar personal and baseline data to those who completed the study (table 2).

Flow of residents through trial

• Download figure
• Open in new tab
• Download powerpoint

The difference in changes in quality of life between both groups was significant (6.1 units, 95% confidence interval 2.1 to 10.3 units; table 3). The intervention group remained stable (0.4, − 1.8 to 2.5) whereas the control group declined (− 5.0, − 9.4 to − 0.6). This difference in change was also seen in physical functioning (5.1, 0.8 to 9.4), psychosocial functioning (7.3, 1.6 to 13.1), and perceived safety (16.6, 4.3 to 28.9). The changes within the groups were not different from each other for sensory functioning (3.6, − 1.7 to 8.9) and perceived autonomy (2.1, − 1.4 to 5.7).

Scores for physical performance were stable in the intervention group (0.2 units, − 2.3 to 2.7 units) but declined significantly in the control group (− 2.2, − 4.1 to − 0.4). The difference in change was mainly in the subscale fine motor function (1.8, 0.6 to 3.0), where the scores for controls declined significantly (− 2.1, − 2.6 to − 1.5). No statistically significant changes were found in scores for gross motor function.

Mean body weight remained relatively stable in the intervention group (0.5 kg, − 0.3 to 1.2 kg) but decreased significantly in the control group (− 1.1, − 1.9 to − 0.2). Changes in body weight between control and intervention groups were significantly different (1.5, 0.6 to 2.4).

Mean energy intake increased significantly in the intervention group (481 kJ, 84 to 878 kJ) but decreased significantly in the control group (− 420, − 713 to − 127). Changes in energy intake between control and intervention group were significantly different (991, 504 to 1479).

Discussion

Family style mealtimes prevent a decline in the quality of life, physical performance, and body weight of nursing home residents without dementia.

In our study, 89% of the invited residents agreed to participate. A total of 28% of the residents did not complete the study for various reasons (see figure). The personal characteristics of these residents were similar to those who completed the study. Moreover, considering the low non-response (11%) and our inclusion criteria we conclude that the study population was representative of residents in Dutch nursing homes with a chronic somatic disorder.

Earlier research showed that residents with dementia benefit from these kinds of interventions.5^–7 Although we excluded this important group, we think that our principal conclusion may be extended to all nursing home residents.

Although many people support the idea that a warm and social ambience during mealtimes can contribute to the wellbeing of nursing home residents, only a small number of studies have measured its effects in a systematic way.4^–7 19 20 Our pilot study produced similar results to the present study, but the statistical power was too low to make strong inferences.4 Most of the studies concerned residents with cognitive impairment and with outcome variables that reflected changes in behaviour instead of changes in quality of life. The sample sizes in these studies were small (< 30 participants).4^–7 Two other studies optimised ambience during mealtimes by changing the food delivery system and by focusing on outcome measures such as food intake and body weight.19 20 In both studies there were no statistically significant changes in body weight, which was probably due to the short intervention periods (three months and 10 weeks).

Considering our simple method for optimising the ambience during mealtimes, the already low quality of life of the residents, and the limited prospect of revalidation or discharge, the observed difference of 6 points (relative change = 10%) for quality of life is important. Earlier research with the same questionnaire showed differences of 15 units (26%) of quality of life in nursing home residents between admission to a stroke rehabilitation programme and discharge.18 If we recalculated our estimates as proportions, then studies with the more drastic intervention of coronary artery bypass graft surgery had higher differences in physical functioning (25% v 13%) and social functioning (16% v 13%).21 However, in an activity stimulating programme for older community dwelling people and lay health mentoring in older people with ischaemic heart diseases the differences for physical functioning were similar (6.1 and 5.4) and were substantially lower for psychosocial functioning related aspects (0.4 and 4.4 v 7.4) than in our study.22 23

The meals that were offered to the residents in the control and intervention groups were of similar weight and nutrient content. In the intervention group, nursing staff who helped residents with feeding were trained to order the same amounts of food as before the intervention, and the kitchen staff ensured that both wards got identical food. The only difference was the timing of choice. This implies that the effects of the current intervention cannot be attributed to differences in availability of food.

The protocol we used has to be considered as one package as we cannot say which part of the intervention protocol had most effect on the residents. This protocol was chosen on the basis of the Dutch situation, but there are other models to improve ambience during mealtimes, such as a restaurant, meals prepared by the residents themselves and waiter service.19 20 24

As with most countries, Dutch nursing homes are limited by staff and budgets. Therefore we had to design an intervention protocol that did not necessitate extra staff and that would not increase the workload of the staff or the costs of the meal. The costs were limited to buying the required materials, such as tablecloths and crockery. With motivated staff this programme is easy to achieve on a low budget. The enthusiasm of the residents and nursing staff convinced the management teams and nursing home boards to proceed with the protocols and to implement the programme in other wards.