Many of the behavioral deficits and excesses of nursing home residents with dementia may result from an environment promoting dependent behavior rather than a disease process (Baltes & Zerbe, 1976a; Beck, Heacock, Rapp & Mercer, 1993; Dawson, Kline, Wiancko, & Wells, 1986; Osborn & Marshall, 1993). In the gerontological literature, this phenomenon is termed "excess disability* (Brody, Kleban, Lawton, & Silverman, 1971). For example, daily living skills of nursing facility residents may deteriorate because, as Baltes and Zerbe (1976a) have said, "the personnel take over for the elderly patient" (p. 24).
One area where this issue appears particularly troublesome is at mealtime (e.g., Geiger & Johnson, 1974; Osborn & Marshall, 1993; Sandman, Norberg, & Adolfsson, 1988). In fact, Baltes and Zerbe (1976a) have noted independent eating is often "the first to go in a chain of self-maintenance skills" (p. 25). In addition, after residents lose independent eating skills, they are at increased risk of health problems such as malnutrition (Hall, 1994).
A number of researchers have examined ways to help elderly individuals re-acquire independent eating skills, taking the approach that eating is an operant behavior, and is therefore a function of its antecedents and consequences. These studies have typically focused on the use of prompts and praise to increase independent eating (Baltes & Zerbe, 1976a, 1976b; Coyne, 1988; Geiger & Johnson, 1974, Van Ort & Phillips, 1995) and improve dietary habits (Stock & Milan, 1993). Similar studies have made changes in the mealtime environment to increase behaviors such as caloric intake (Elmstahl, Blabolil, Fex, Kuller, & Steen, 1987), communication (Davies & Snaith, 1980), and resident-to-resident assistance (Sandman, et al., 1988). In addition, the work of Kayser-Jones suggests staffing, staff training, and mealtime atmosphere have an important impact on mealtime outcomes for the residents (Kayser-Jones, 1996, 1997; Kayser-Jones & Schell, 1997a, 1997b). Collectively, these studies underscore the importance of the environment in setting the occasion for, and maintaining, important mealtime behaviors.
While procedures to reacquire independent eating are clearly important, techniques are needed to prevent its loss in the first place - especially because research indicates repeated practice of skills by persons with dementia enhances retention (Dick et al., 1996). One procedure likely to contribute to the loss of mealtime independence is the common institutional practice of serving meals on prepared plates. This practice removes opportunities to select foods, choose portion sizes, serve oneself, pass serving bowls to others, take seconds, and engage in the social interactions revolving around the serving and passing of food.
Researchers have not directly studied the relationship between use of prepared plates and dependent behavior at mealtimes. However, studies have shown the use of serving dishes, in place of prepared plates, results in improvements in communication and eating skills (Melin & Gotestam, 1981), nutritional intake (Elmstahl, et al., 1987), and mealtime atmosphere (Sherwood, 1973). In addition, the use of serving bowls normalizes the mealtime atmosphere, giving it a more home-like feel. As Kayser-Jones and Schell (1997b) have shown, "creating a pleasant mealtime context can positively affect behavior" (p. 38). They also said, "Nursing homes should strive to provide meals as attractively as possible, reflecting mealtime as it would be in one's home" (p. 38).
The purpose of this study was to examine the impact of using serving dishes (i.e., eating "family-style") versus prepared plates on participation in mealtime tasks by residents with dementia who still possess independent eating skills. The authors' hypothesis was that the use of familystyle meals would result in an increase in independent resident behavior. A secondary hypothesis was that the serving and passing of food during family-style meals would result in an increase in appropriate resident communication, as shown in research with individuals with developmental disabilities (VanBiervliet, Spangler, & Marshall, 1981).
The authors viewed increased communication as an important goal because social interaction among persons with dementia is typically rare (VanHaitsma, Lawton, Kleban, Klapper, & Corn, 1997). Inappropriate communication also was measured to insure an increase in communication was not simply an increase in ineffective vocalizations, as one study found when participation in nursing home activities increased (Carstensen & Erickson, 1989).
Setting and Participants
This study took place in the locked dementia care unit of an assisted living facility in a midwestern town with a population of 80,000. This unit contained private rooms and baths for up to seven residents along with a small living room, dining room, kitchen, and enclosed courtyard. The unit was staffed by one certified nursing assistant (CNA) per 8-hour shift.
A convenience sample was used, which included the six residents who were living in the dementia care unit at the beginning of the study. One resident died before the study was completed. Therefore, her data are not included in this report. The remaining five residents were women who ranged in age from 76 to 87 (M = 80).
All residents had a physician's diagnosis of dementia or Alzheimer's disease. Residents' levels of functioning ranged from moderate to severe dementia, as suggested by scores ranging from 3 to 16 (M = 8) on the MiniMental State Examination (Folstein, Folstein, & McHugh, 1975)- a commonly used cognitive screening instrument. All residents were ambulatory and none required skilled nursing care. Residents ate their meals independently or with minimal assistance from the CNA. An evaluation of residents' eating skill level by the CNA revealed a mean score of 4.6 (range, 3 to 5) on a 5point scale, with 1 being "cannot do it at all" and 5 being "not at all difficult."
One CNA also served as a participant. This CNA was a 24-year-old woman with a high school education who had worked in the facility for 3 years. She was selected as a participant because she worked full time on the day shift in the dementia care unit. She had no specialized training in dementia care.
Definitions and Measurement
Two resident behaviors (participation and communication) and one CNA behavior (offering praise) were observed during lunchtime in the locked dementia care unit. Observations occurred approximately three times per week at lunchtime and lasted for a mean of 87 minutes. The first and second authors served as observers. Observers stood at the sides of the dining room where they had an unobstructed view of the participants, but were not in the way of mealtime activities. The observers had been conducting observations in this setting for more than a year as a pan of a series of research studies, so staff and residents were accustomed to their presence.
Participation. Resident participation in mealtime tasks was observed using a checklist of tasks. This checklist included four categories: preparation tasks, serving and passing food, taking seconds, and clean-up tasks. Each task had an associated behavioral definition. Preparation tasks involved putting place mats, napkins, silverware, plates, and cups at the appropriate spot the table, and sitting down at one's place. Serving food and taking seconds involved putting a serving spoon in a serving bowl and placing a portion of the bowl's contents on one's plate. Passing food involved picking up a serving bowl and handing it to one's neighbor. Clean-up tasks included taking one's silverware, plate, and cup to the cart next to the dining table, along with throwing away one's paper napkin and place mat.
Twenty-two fixed tasks (e.g., sitting down at the table) were observed at each meal, along with as many as 26 additional tasks (e.g., passing the butter), depending on the day's menu and the condition of the experiment (i.e., family-style meal vs. prepared plates). Eating was not included on the task checklist because residents were relatively independent in their eating skills.
For each task on the checklist, observers scored the most intrusive level of assistance (independent, verbal, gestural, modeling, physical or complete assistance) received by residents from the CNA to complete the task. This scoring system was based on the system of least prompts (Doyle, Wolery, AuIt, & Gast, 1988). Participation was calculated by counting the number of tasks residents completed independently or with a verbal, gestural, or modeling prompt. Tasks where residents received physical prompts or complete assistance were not counted as participation.
Communication. Resident communication was scored as either appropriate or inappropriate. Appropriate communication included any intelligible vocalizations, contextually suitable and non-aggressive. Inappropriate communication included any unintelligible vocalizations, out of context, aggressive, or abnormally repetitive (perseveration). Residents' communication was observed throughout the meal using an interval recording procedure. Each resident's behavior was observed sequentially for 30 seconds (25 seconds to observe and 5 seconds to score). If both appropriate and inappropriate communication occurred during the same interval, only inappropriate communication was scored.
Praise. Observers tallied the number of praise statements made by the CNA during the lunchtime observation. Praise was defined as a statement of approval, encouragement, or affirmation directed at residents (e.g., "good job passing the peas," "the table looks great").
I nter~ob server Agreement. Reliability observations were conducted by the two observers, working independently, during at least 20% of the observation sessions per condition for participation, communication, and praise. To calculate inter-observer agreement for participation, observers' records were compared item by item. The total number of agreements was divided by agreements plus disagreements. Mean inter-observer agreement on participation was 93% and ranged from 83% to 100%.
For communication, observers' records were compared interval by interval, and inter-observer agreement was calculated by dividing the total number of agreements by the sum of agreements plus disagreements. Mean inter-observer agreement for communication was 97% and ranged from 94% to 99%. Mean inter-observer agreement for praise, which was calculated by dividing the smaller tally by the larger tally, was 93% and ranged from 71% to 100%.
Experimental Design and Treatment Procedures
An ABA reversal design was used to assess the effects of family-style meals on resident participation and communication at lunchtime. A subsequent B' condition was added to evaluate additional treatment procedures described below.
Prepared Plates (A). During the baseline condition, the facility's normal operating conditions were in place. Meals were fixed in the facility's main kitchen by a cook who also prepared each resident's plate in advance of the meal. The cook chose the types and amounts of food the residents received on their plates. Plates were covered with plastic wrap and delivered to the locked dementia care unit on a cart. Residents had access to the cart, which was left in the dementia care unit for the duration of the meal.
Family-style Meals (B). During the family-style meal condition, the cook put the lunchtime food into communal serving dishes instead of preparing individual resident plates. Serving dishes, serving spoons, and empty plates were delivered to the locked dementia care unit on a cart. All other procedures remained the same.
Figure 1. Mean percentage of tasks with resident participation during lunchtime observations across the four experimental conditions.
Prepared PUtes (A). During the reversal condition, the serving dishes were removed and the cook delivered prepared plates to the residents on a cart as in the baseline condition.
Family-style Meals Piws CNA Training (B'). In an attempt to strengthen the family-style meal intervention, the CNA participated in a 45minute in-service training session on prompting and praising appropriate resident behavior. She received instruction on three topics:
* Following a system oí graduated prompts, moving from least to most intrusive (i.e., verbal prompt, gestural prompt, modeling prompt, physical prompt, complete assistance).
* Offering frequent praise to the residents for appropriate behaviors.
* Providing increased opportunities for resident involvement in routine tasks.
This training session was followed by two lunchtime sessions of on-thejob modeling and coaching. During these sessions, the CNA received feedback on her performance of prompting, praising, and providing opportunities for resident involvement. After the CNA completed this training, the family-style meal procedure was reinstated as before.
At the beginning and end of the study, the CNA was asked to rate Ket satisfaction with residents' levels of participation and communication during lunchtime and her overall satisfaction with lunchtime in the locked dementia unit. Ratings were made on a 5-point Likert-type scale ranging from 1 (very dissatisfied) to 5 (very satisfied). At the end of the study, the CNA was asked several additional questions relating specifically to her use of, and satisfaction with, the family-style meal procedure.
Figure 1 shows the mean percentage of resident participation in mealtime tasks. During baseline, when prepared plates were used, residents participated in a mean of 10% of mealtime tasks. When prepared plates were replaced by family-style meals, the mean percent of tasks in which residents participated increased to 24%, and dropped to 6% when prepared plates were reintroduced. After the CNA received training and family-style meals were reintroduced, resident participation increased to a mean of 65% of tasks. All five participants showed behavior mirroring the mean group pattern.
The percentage of intervals during which residents engaged in appropriate communication showed a similar pattern to that of resident participation. During the baseline condition, when residents received prepared plates, the mean percentage of intervals with appropriate communication was 5.5%. This mean increased to 10.6% when family-style meals were introduced and dropped to 3.8% when prepared plates were reinstated. After the CNA received training and the family-style meal procedure was reinstated, the mean percentage of intervals with appropriate communication increased to 17.9%.
Inappropriate communication was relatively low and stable throughout the study, occurring during a mean of 3.9% baseline (prepared plates) intervals, 4.1% of family-style meals intervals, 5% of reversal intervals, and 1.4% of intervals after the CNA received training and family-style meals were reinstated.
Figure 2. Number of praise statements made by the CNA during lunchtime observations across the four experimental conditions.
Figure 2 shows the frequency of praise by the CNA across mealtime observations. During baseline (prepared plates), the CNA praised the residents a mean of 0.2 times per meal. The number of praise statements increased to a mean of 7.2 when family-style meals were introduced and dropped to zero when family-style meals were withdrawn. After the CNA received training and family-style meals were reinstated, the CNA praised residents a mean of 14.2 times per meal.
Ratings revealed the CNA's satisfaction with the level of resident participation during lunchtime increased from a score of 2 (somewhat dissatisfied) at the start of the study to a score of 5 (very satisfied) at the end of the study. The CNA provided a score of 4 (somewhat satisfied) for her level of satisfaction with resident socialization during lunch at the both the beginning and end of the study. Her overall satisfaction with lunchtime increased from a score of 4 (somewhat satisfied) at the beginning of the study to a score of 5 (very satisfied) at the end of the study.
The additional CNA ratings gathered at the end of the study revealed she "definitely will" continue using family-style meals during lunchtime. She was "very satisfied* with the amount of work she had to do during family-style meals, and using the family-style meal procedure was "much less work" than using prepared plates. When asked if she would recommend the family-style meals to others working in similar situations, she gave a rating of 4 on a scale from 1 (do not recommend at all) to 5 (recommend very much). The CNA was asked two additional questions regarding which meal procedure (prepared plates or family-style) resulted in the most independent behavior and the most social interaction. In both cases, she selected family-style meals.
Resident participation and appropriate communication increased with the introduction of family-style meals and decreased when they were withdrawn, suggesting the family-style meal procedure had a positive impact on resident behavior. Inappropriate communication did not increase when family-style meals were introduced, indicating the residents did not find the procedure aversive. However, only modest increases in behavior occurred during the first introduction of the family-style meal procedure and the data suggest staff training may be necessary to realize large behavior changes.
Indeed, it was only after the CNA was trained in prompting, praising, and providing opportunities, that resident participation and communication showed substantial gains. The CNA training procedure, however, was not experimentally analyzed because staff and resident turnover necessitated discontinuance of the study after the 25th observation.
The data on praise suggests the family-style meal procedure alone, without any staff training on the topic, set the occasion for the CNA to praise the residents. This increase in praise may explain, in part, the increase in resident participation and communication during the first family-style meal condition. Unfortunately, comparable data were not collected on the rate of prompting because observers only scored the most intrusive prompt delivered by the CNA. Future studies might examine the rate of CNA prompting to gain a better understanding of the salient features of the familystyle meal procedure.
It was not possible to collect comparable follow-up data on the maintenance of family-style meals because of the turnover of staff and residents. However, an observation conducted 2 months after the end of data collection showed the facility continued to use the family-style meals at lunchtime and had, on its own, adopted most of the family-style meal procedures during breakfast.
One may be concerned that persons with dementia would not make appropriate food choices and would have a higher likelihood of catching communicable diseases when using common serving dishes during family-style meals. While it would be wise for facilities to monitor these issues, resident bodyweights and illnesses did not appear to be a problem in this study. A review of residents' charts indicated no instances of viral or bacterial infections during this study.
Three of the five residents gained weight during the study (5, 8 and 10 lbs). Two of these residents were seriously underweight and the weight gain was welcome. The two remaining residents lost weight (5 and 6 lbs). One was overweight and the weight loss was welcome, while the other had a serious, non-communicable illness contributing to her weight loss. The facility's registered nurse indicated her weight loss had nothing to do with the family-style meal procedure.
Anecdotally, the observers reported residents typically served themselves from each serving dish and, as a result, received balanced meals. This is not unexpected, as one study showed the use of serving dishes, in comparison to prepared plates, resulted in improvements in nutritional intake (Elmstahl et al., 1987). Future studies, however, would be advised to track health outcomes to insure family-style meals did not have an adverse affect on resident health.
The authors realize the ability to generalize the findings of this pilot study to other settings, and populations are limited because of the small sample size. Indeed, small sample size is an inherent weakness of single-subject design research. However, as Rubin and Babbie (1997) said, "a study must be generalizable to some real-world settings and it must represent that which it intends to represent. It does not have to represent every conceivable population or setting" (p. 304).
In this study, there was an attempt to represent small, dementia care units that are part of a larger facility. The setting used in the present study was located in a facility that is part of a large, regional chain of facilities using similar floor plans and staffing procedures, indicating the setting is not unique and likely shares characteristics with other dementia care units. Nevertheless, the external validity of the present study would be strengthened through replication.
One attribute of this study that likely improves its external validity is the fact that the study was performed in a real-world setting using the CNA as the change agent rather than the researcher. The researchers in this study had no formal role in the setting other than to serve as observers. The staff of the facility, rather than the researchers, were the ones delivering the family-style meal intervention. As a result, the intervention is not dependent on idiosyncratic practices of the researchers for its success.
Residential care facilities may be willing to adopt a procedure as simple as family-style meals, especially given its enthusiastic acceptance and continuation by nursing and kitchen staff members in this study. Indeed, ratings showed the CNA appeared very satisfied with die family-style meal procedure, both in its impact on improving resident behavior and on easing her workload. In addition, the cook reported being very satisfied with familystyle meals, noting it was much easier to put food into serving bowls than it was to prepare individual resident plates.
The authors suspect family-style meals may set the occasion for a more positive caregiving experience for CNAs. For example, watching residents exhibit a greater level of independence and an increased rate of conversation may be very satisfying for staff. In addition, if residents are doing more for themselves, the CNAs may be free to attend to other important tasks that might improve the mealtime environment even further.
The results of this study suggest the training of nursing staff in the effective use of praise and prompts is worthwhile. However, even if facilities do not provide the staff training in using praise and prompts, the family-style meal procedure, by itself, is more likely to maintain resident independence than the common practice of using prepared plates.
Given that many, if not most, institutional practices foster dependent behavior, anything having potential to set the occasion for independence of residents with dementia is worth promoting. As Beck (1981) concluded in an article on the dining experiences of institutionalized elderly, "a healthy appetite and enjoyment of eating are not consistent with...dependency" (p. 106). In addition, anything normalizing the mealtime environment and makes the mealtime atmosphere more pleasant and homelike is worth promoting. As Kayser-Jones (1996) said, "it is tragic that in the last years of their lives, many nursing home residents are denied one of life's greatest pleasures - to sit down to a nice meal of one's liking in pleasant surroundings" (p. 31).
- Baltes, M.M., & Zerbe, M.B. (1976a). Reestablishing self-feeding in a nursing home resident. Nursing Research, 25(1), 24-26.
- Baltes, M.M., & Zerbe, M.B. (1976b). Independence training in nursing home residents. The Gerontologist, 16, 428-432.
- Beck, C. (1981). Dining experiences of the institutionalized aged. Journal of Gerontological Nursing, 7(2), 104-107.
- Beck, C, Heacock, P., Rapp, CG., & Mercer, CO. (1993). Assisting cognitively impaired elders with activities of daily living. The American Journal of Alzheimer's Care and Related Disorders and Research, 8(6), 1 1-20.
- Brody, E.M., Kleban, M.H., Lawton, M.P., & Silverman, H.A. (1971). Excess disabilities of mentally impaired aged: Impact of individualized treatment. The Gerontologist, 11, 124-133.
- Carstensen, L.L., & Erickson, RJ. (1989). Enhancing the social environments of elderly nursing home residents: Are high rates of interaction enough? Journal of Applied Behavior Analysis, 19(4), 349-355.
- Coyne, M.L. (1988). The effect of directed verbal prompts and positive reinforcement on die level of eating independence of elderly nursing home clients with dementia. Dissertation Abstracts International, 50/08, 3396B.
- Davies, A.D.M., & Snaith, P.A. (1980). The social behaviour of geriatric patíents at mealtimes: An observational and an intervention study. Age and Ageing, 9(2), 93-99.
- Dawson, P., Kline, K., Wiancko, D.C., & Wells, D. (1986). Preventing excess disability in patients with Alzheimer's Disease. Geriatric Nursing, 7(6), 298-301.
- Dick, M.B., Shankle, R.W., Beth, R.E., DickMuehlke, C, Cotman, C.W., & Kean, M.L. (1996). Acquisition and long-term retention of a gross motor skill in Alzheimer's disease patients under constant and varied practice conditions. Journal of Gerontology: Psychological Sciences, SlB, 103-111.
- Doyle, P.M., Wolery, M., AuIt, M.J., & Gast, D.L. (1988). System of least prompts: A literature review of procedural parameters. Journal of the Association for the Severely Handicapped, 13(1), 28-40.
- Elmstahl, S., Blabolil V., Fex, G., Kuller, R., & Steen, B. (1987). Hospital nutrition in geriatric long-term care medicine. Comprehensive Gerontology, 1(1), 29-33.
- Folstein, M.E, Folstein, S.E., & McHugh, P. R. (1975). "Mini-mental state": A practical method for grading the cognitive state of patients for the clinician. Journal of Psychiatric Research, 12(3), 189-198.
- Geiger, O.G., & Johnson, L.A. (1974). Positive education for elderly persons: Correa eating through reinforcement The Gerontologist, 14, 432-436.
- Hall, GR. (1994). Chronic dementia: Challenges in feeding a patient. Journal of Gerontological Nursing, 20(4), 21-30.
- Kayser-Jones, J. (1996). Mealtime in nursing homes: The importance of individualized care. Journal of Gerontological Nursing, 22(3), 26-31.
- Kayser-Jones, J. (1997). Inadequate staffing at mealtime: Implications for nursing and health policy. Journal of Gerontological Nursing, 23(9), 14-21.
- Kayser-Jones, J., & Schell, E. (1997a). The effect of staffing on the quality of care at mealtime. Nursing Outlook, 45(2), 64-72.
- Kayser-Jones, J., & Schell, E. (1997b). The mealtime experience of a cognitively impaired elder: Ineffective and effective strategies. Journal of Gerontological Nursing, 23(7), 33-39.
- Melin, L., & Gotestam, K.G. (1981). The effects of rearranging ward routines on communication and eating behavior of psychogeriatric patients. Journal of Applied Behavior Analysis, 14(1), 47-51.
- Osborn, CL., & Marshall, M.J. (1993). Selffeeding performance in nursing home residents. Journal of Gerontological Nursing, 19(3), 7-14.
- Rubin, A., & Babbie, E. (1997). Research methods for social work (3rd ed.). Pacific Grove, CA: Brooks/Cole.
- Sandman, P.O., Norberg, A., & Adolfsson, R. (1988). Verbal communication and behavior during meals in five institutionalized patients with Alzheimer-type dementia. Journal of Advanced Nursing, 13(b), 571-578.
- Sherwood, S. (1973). Sociology of food and eating: Implications for action for the elderly. American Journal of Clinical Nutrition, 26, 1108-1110.
- Stock, L.Z., & Milan, M.A. (1993). Improving dietary practices of elderly individuals: The power of prompting, feedback, and social reinforcement. Journal of Applied Behavior Analysis, 26, 379-387.
- VanBiervliet, A., Spangler, RE, & Marshall, A.M. (1981). An ecobehavioral examination of a simple strategy for increasing mealtime language in residential facilities. Journal of Applied Behavior Analysis, 14, 295-305.
- VanHaitsma, K., Lawton, M.P., Kleban, M.H., Klapper, J., & Corn, J. (1997). Methodological aspects of the study of streams of behavior in elders with dementing illness. Alzheimer Disease and Associated Disorders, 11(4), 228-238.
- Van Ort, S., & Phillips, L.R. (1995). Nursing interventions to promote functional feeding. Journal of Gerontological Nursing, 21(10), 6-14.