- Top of page
- Conflict of interest
Behavioral disturbances that accompany dementia as it progresses represent particularly great challenges to both clinicians and informal caregivers (Wooltorton, 2002). Research shows that almost all nursing home residents present at least one behavioral problem and half showed four or more behavior problems (Tariot et al., 1996). Wandering, one of the most common behavioral disturbances among people with dementia (PWD), has been associated with patient morbidity (Rheaume et al., 1988, Evans and Strumpf, 1989) and mortality (Moritz et al., 1997) owing to safety risks including elopements, falls, and injuries. It has also been reported to require costlier care of PWD (Schnaider Beeri et al., 2002). There is an urgent need, therefore, to isolate potentially modifiable factors that may contribute to wandering behavior among nursing home residents.
The need-driven dementia-compromised behavior (NDB) model explains how wandering results from the interplay of background and proximal factors (Algase et al., 1996). Background factors include neurocognitive factors, general health, personal characteristics, and socio-demographics, which are generally not subject to change in the short-term; proximal factors include both physical and social environments, physiological needs, and psychological needs (Algase et al., 1996), which are more dynamic and amenable to intervention. Emotional expression, which can be reliably measured using observational techniques (Larsen and Fredrickson, 1999), even among those with moderately severe dementia (Lawton et al., 1996; Vogelpohl and Beck, 1997; Kolanowski et al., 2002), may reflect psychological needs that fluctuate in response to environmental or internal stimuli. In contrast to more persistent affective states such as mood, which are relatively stable, emotional expression may represent a signal of underlying unmet need state that prompts behaviors such as wandering, and therefore serves as a potential intervention target for prevention or treatment of behaviors among those with dementia. As dementia progresses, behaviors of many PWD become more stimulus bound, owing to memory deficits and degeneration of frontal lobe function (Mesulam, 2008), and thus, a great deal of the lives of PWD are carried out in the “here and now” (Kolanowski et al., 2002; Mesulam, 2008). Relative to the other changes in cognitive function that accompany dementia, including language, memory, and higher order reasoning, the ability to express basic emotions nonverbally is relatively well preserved in dementia (Magai et al., 1996, Kolanowski et al., 2002).
Clinical observations support the inference that certain emotions and behaviors in dementia are associated (Bartol, 1979, Mayhew, 2005). Nursing care that targets emotional response, such as the use of music, touch, rocking, and massage, reduces behavioral disturbances (Finnema et al., 2000, Cohen-Mansfield, 2001, Ayalon et al., 2006). Several studies have examined the impact of mood states (e.g., depression and anxiety) on problematic behaviors including wandering (Lyketsos et al., 1997, Teri et al., 1999, Lai and Arthur, 2003). Specifically, Alzheimer's disease patients who had depression showed significantly higher rates of wandering than did Alzheimer's disease patients without depression (Lyketsos et al., 1997); likewise, anxiety symptoms were positively correlated with wandering (Teri et al., 1999, Lai and Arthur, 2003). A study found that high ambiance scores of the environment (i.e., positive environmental valence of an environment) may reduce a wanderer's walking (Yao and Algase, 2006).
Although several studies used emotion-focused care approaches or mood states as a measure of emotion to examine the relationships between emotion and behavioral disturbances, the emotional expression of PWD has received only limited attention as a potential mediator of behavioral disturbance (Finnema et al., 2000; Kolanowski and Whall, 2000). Thus, this study explored the relationship between emotional expression and wandering. Research questions were as follows:
- How is observable emotional expression related to wandering in PWD?
- Does observable emotional expression predict wandering in PWD after controlling for cognitive impairment, participant characteristics, and time of day?
- Top of page
- Conflict of interest
Wandering behavior is a common dementia-related behavior of PWD that is receiving increased attention because of its association with morbidity, mortality, and cost. This study shows that wandering occurs frequently among ambulatory PWD in nursing homes—in this study, an average of three episodes per hour. Although many factors have been shown to influence wandering behavior, this is the first study to focus on the relationship between wandering and directly observed emotional expression. Both positive and negative emotional expressions were significantly associated with wandering rates. On the basis of propositions from the NDB model, we hypothesized that emotion would represent a proximal factor influencing dementia-compromised behaviors (Algase et al., 1996). Proximal factors represent factors that are potentially modifiable, such as physiological need states (e.g., pain, hunger, and thirst) and psychological needs (e.g., emotion). This study supported the NDB model by suggesting that current emotional expression varies over the course of the day and is associated with wandering.
Although several studies reported the relationship between mood states (e.g., depression and anxiety) and wandering rates (Lyketsos et al., 1997, Teri et al., 1999, Lai and Arthur, 2003), a unique feature of this study was that it explored the relationship between wandering and observable emotional expressions. In contrast to mood states, which last days to months (Kolanowski et al., 2002), emotional expressions occur over minutes to hours. Because PWD tend to focus on the present owing to their cognitive impairment (Kolanowski et al., 2002), observable emotional expressions represent a promising approach for capturing psychological need states of PWD.
Somewhat surprisingly, PEE was associated positively with wandering rate, whereas NEE had an inverse relationship to wandering rate. One possible explanation for these results is that PWD who are sad or angry may respond by sitting alone or staying in their rooms for periods, rather than walking around. On the other hand, PWD who are happy or experiencing pleasant emotions may respond to this state by engaging in physical activity such as walking around. Owing to deficits in executive function in PWDs, this physical activity may take on the form of wandering.
An alternative explanation for these findings is that wandering behavior cannot be explained by emotional expressions alone because wandering behavior reflects an interaction between background factors and proximal factors. For example, although not included in this study, the combination effect of personality traits (i.e., one of the background factors) and emotional expression (i.e., one of the proximal factors) on wandering among PWD may influence the behaviors we observed. It may be that when more introverted people experience NEE, their behavior responses are likely to be more private and less public. They are likely to spend time alone with NEE. Because the participants were a convenience sample from each nursing home, it is possible that one personality type may have been overrepresented in this sample.
Yet, another explanation is related to the nursing home environment. Study institutions are likely to represent typical nursing homes in Michigan or Pennsylvania because random cluster sampling was used to select the study sites. In typical nursing homes, PWD may not have enough opportunities to engage in structured physical activities owing to inadequate nurse-staffing levels, limited programming, or constrained spaces (Bates-Jensen et al., 2004, Meeks and Looney, 2011). Even if opportunities for structured physical activity exist, they may not take into account patient preference for when the activity occurs. In environments where opportunities for physical activity are tailored to individualized needs, the association between PEE and wandering may decrease.
This study also showed wandering to be dependent on both cognition and emotion. According to the NDB model, wandering is an interrelated response of both emotional reaction and cognition (Yao and Algase, 2008); emotional reaction occurs in response to a variety of cognition changes or refines perception of stimulus, emotional reaction, and action impulse, ideally, resulting in an optimal behavior response (Yao and Algase, 2008). PWD with more severe cognitive dysfunction lack the ability to modulate the action impulse, sometimes resulting in disorganized or unpredictable behavior. Consistent with NDB, and with other studies, this study confirmed that higher MMSE scores (i.e., better cognitive functioning) were related to less frequent wandering (Buchner and Larson, 1987, Burns et al., 1990, Algase et al., 2001, Song and Algase, 2008). This study also found that younger age had a significant positive relationship with wandering rates. This finding was consistent with previous studies that showed younger age to be significantly related to wandering (Schreiner et al., 2000), which may be reflecting better health or stamina among those from this age group.
Limitations of this study include the use of a convenience sample of nursing home residents, the cross-sectional design used to obtain data over a brief period, the availability of only limited measures of cognition (MMSE) and other potentially important background factors such as personality, and the lack of measures of the facility context. In addition, emotional expression of PWD can result in pathological changes associated with dementia. However, these patients are in the minority (Starkstein et al., 1995), and we attempted to minimize the likelihood that psychiatric symptoms might account for emotional displays by excluding PWD who had acute psychiatric or medical illness. Strengths of the study include the selection of nursing homes using a randomized cluster design, which increases the likelihood that study participants were exposed to customary nursing home care environments, and the quantification of emotional expressions and wandering using repeated-measures observation data, which provided a more precise measure of frequency for each variable.
Additional studies are needed to better understand the temporal relationship between emotional expression and wandering. Specifically, we do not know whether PEE precedes wandering or is in response to wandering. Likewise, exploring the effect of the interaction between emotional expression and cognition with wandering behavior would help to clarify whether level of cognitive function moderates the effect of emotional expression on wandering.
In this exploratory study, wandering behavior was related not only to cognition but also to emotional expression. Health care workers should therefore consider both the cognitive function and emotional status of PWD to develop individualized interventions for care to reduce wandering behavior. Nursing staff need to be aware of the finding that increased frequency of PEE over baseline may be related to an increased risk of wandering, and anticipate the need to intervene in response to this change. Examples of strategies that could be employed include engaging volunteer staff or safety assistants to increase surveillance when changes in emotional expressions are observed. If a temporal patterning of changes in emotional expression can be discerned, then specific activity programming might be a strategy to direct the increased physical activity that may accompany increases in PEE in a safe manner. Future studies should examine whether tailored interventions to maintain physical activity for people with moderate or severe dementia influence emotional expression and related wandering.
- Lower cognitive status was related to higher rates of wandering.
- Positive emotional expression was positively related to wandering rate, whereas negative emotional expression was negatively related to wandering rate.
- Both emotional expression and cognitive status should be considered when developing interventions to improve wandering behaviors of people with dementia.