Dining Environments in Long Term Care: Prevalence of Features and Construct Validity of Two Measures

Abstract

Background: Previous research suggests that the physical and psychosocial environments can improve outcomes for residents living in long term care (LTC). However, research has inconsistently implemented interventions that target these environments resulting in the inability to compare results across studies. These inconsistencies are due to a lack of standardized measures that quantify these environments reliably and validly. Thus, instruments that address this gap would result in dependable claims of mealtime experience summarizations and consistent evaluations. Additionally, prior to this study, there has not been an examination of the differences in prevalence of physical and psychosocial dining environments across Canada. At this point it is unclear whether consistency exists nationally with regard to environmental characteristics and delivery of care and it is unknown what areas require further improvements to meet industry standards.

Purposes: 1) Assess the construct validity of the dining environment audit protocol (DEAP), 2) assess the construct validity of the Mealtime Scan (MTS), 3) examine the construct validity of the mealtime relational care checklist (M-RCC) and 4) demonstrate the prevalence of key features of these instruments and differences where they exist, among provinces included in this data set.

Methods and Findings: This thesis is a secondary data analysis of the Making Most of Mealtimes (M3) study, which is a cross sectional Canadian study conducted in Alberta, Manitoba, New Brunswick and Ontario. This study collected data on the multilevel determinants of food intake in 32 LTC homes, which included 639 residents and 82 dining rooms. Resident energy and protein intake estimated from three weighed and estimated food intake records were proxies for intake. Energy intake per kilogram body weight and protein intake per kilogram body weight variables were created as outcomes for these analyses, and gender, age and cognitive performance score (CPS) were used as covariates in cluster regression stratified by dementia care and general care units. Other constructs compared to measures included: nutritional status, CPS, and dining room level constructs and staff perceptions of person centered care. Features and characteristics of instruments are described and analyzed to determine their association with key summary scales of instruments. Finally, comparisons across provinces were made to determine differences in prevalence of instrument variables. The key methods and findings for analysis of each measure will be discussed.

1) Cluster regression analysis determined that the needs of residents in dementia care and general care units differed. Further, energy and protein intake was minimally influenced by the physical characteristics of the dining room as assessed by the DEAP. Through regression analysis (p<0.05) it was found that the DEAP homelikeness summary scale was positively associated with a view of the garden, clock and posted menu. Functionality summary scale was positively associated with number of chairs and lighting, while negatively associated with furniture with rounded edges and clutter. The construct validity of the homelikeness and functionality scales of the DEAP was determined through correlations (p<0.05). The functionality scale was positively associated (p<0.05) with the MTS physical scale, the dining room M-RCC, the resident M-RCC and the Mini Nutritional Assessment- Short Form (MNA-SF). Homelikeness was positively associated (p<0.05) with the staff person directed care (PDC) score and the Cognitive Performance Scale (CPS), while negatively associated with energy and protein intake. Further, the homelikeness and functionality scales were associated with one another. These associations determined that the DEAP summary scales are construct valid. Few physical characteristics of the dining room as assessed by the DEAP differed (p<0.01) across Alberta, Manitoba. New Brunswick and Ontario.

2) Energy and protein intake was minimally influenced by the physical and psychosocial characteristics of the dining room as assessed by the MTS. Regression analysis revealed that the MTS physical summary scale was positively associated with music availability and the dining room M-RCC ratio, while negatively associated with number of staff passing food and number of residents. The social environment scale was positively associated with social noise, number of residents requiring assistance and the M-RCC ratio. The person centered care (PCC) summary scale was positively associated with adequate lighting, excess noise and the dining room M-RCC ratio. Construct validity of the scales was examined using correlations (p<0.05). The three MTS summary scales were positively associated. The physical scale was also positively associated (p<0.05) with the DEAP functionality scale, the resident and dining room M-RCC and the MNA-SF. The social scale was positively associated (p<0.05) with the dining room M-RCC, the MNA-SF and CPS score. The PCC scale was positively associated (p<0.05) with the dining room and resident M-RCC, the MNA-SF and CPS score. These associations determined that the MTS summary scales are construct valid. Physical and psychosocial environments as assessed by the MTS minimally differed (p<0.01) across Alberta, Manitoba, New Brunswick and Ontario.

3) Correlations were computed to determine the construct validity of the resident level M-RCC ratio. The resident M-RCC was positively associated (p<0.05) with the DEAP functionality scale, the dining room M-RCC, the MTS PCC summary scale, and the MNA-SF score and negatively associated (p<0.05) with protein intake and CPS score. These associations determined that the resident M-RCC is construct valid. RCC and PCC practices as assessed by the resident M-RCC differed (p<0.01) across the provinces of Alberta, Manitoba, New Brunswick and Ontario.

Conclusion: In conclusion, the physical and psychosocial environments as assessed by the MTS and DEAP minimally explained the variance of energy and protein intake in both dementia care and general care units, but summary scales were associated with nutritional status. The DEAP, MTS and M-RCC exhibit validity through the significant associations between the summary scales and the individual variables of each instrument. Additionally, the construct validity of these instruments was supported through the significant correlations with other instruments collected in M3. The physical and psychosocial components of the dining environment can be improved in Alberta, Manitoba, New Brunswick and Ontario to promote consistency on a national level. This secondary analysis of the M3 dataset suggests that the DEAP, MTS and M-RCC are construct valid standardized instruments that may be used to quantify the physical and psychosocial environments. Prior to this study, construct valid instruments did not exist, thus this analysis offers a basis for future research. Prevalence estimates identify areas where practices can be improved further to promote the physical and psychosocial environments.