BACKGROUND/RATIONALE:
Wandering is a syndrome of dementia-related locomotion, characterized by frequent and repetitive ambulation, resulting in morbidity and mortality. Approximately 60% of persons with dementia who live in the community will wander and become lost at least once per year. Once lost, mortality rates are high for those not discovered within 24 hours, with death attributed to exposure (68%), drowning (23%), or other accidents (9%). Research published to date has focused almost exclusively on nursing home settings, disproportionately weighting samples with persons in the more advanced stages of dementia. Consequently, very little is known about the early onset of wandering, including its emergence and earliest behavioral manifestations, associated risks, or the evolution of these behaviors with the progression of dementia. This study examined the early trajectory of wandering by focusing Veterans who are newly diagnosed with mild dementia residing in the community.

OBJECTIVE(S):
The objectives of this study were to: (1) Characterize the early trajectory of wandering and (2) Test a new conceptual model of hazardous wandering, addressing factors contributing to the onset of wandering, as well as the adverse events associated with types and intensity of wandering behaviors over time. Primary Research Questions include: (Q1) What is the incidence and prevalence of specific types of wandering in community-dwelling veterans with mild dementia? (Q2) How do the types and intensity of wandering change over time? (Q3) To what extent do contributing factors predict types and intensity of wandering?

METHODS:
This was a 3-year prospective longitudinal study of community-dwelling veterans with mild dementia and their caregivers. Outcome measures included types and intensity of wandering, and potential adverse outcomes as measured by the Revised Algase Wandering Scale-Community Version (RAWS-CV) sub-scales. Data were collected on a series of variables that likely impact wandering behavior (contributing factors). Data were collected from the caregiver and Veteran through, interviews, self-report questionnaires and activity meters. Inclusion criteria: (1) Veterans over age 60; (2) newly diagnosed with mild dementia; (3) living in a home-type setting within 60 miles of the hospital; (4) with a caregiver, who is willing and able to serve as a study partner; and (5) both Veteran and caregiver must speak English. RAWS-CV sub-scales target types of wandering (persistent walking, spatial disorientation) intensity of wandering (repetitive walking), and immediate adverse outcomes (elopement behavior, mealtime impulsivity and negative outcomes). Mean scores were calculated for each sub-scale based on five point response options (Never/Unable = 1, Seldom = 2, Sometime = 3, Usually = 4 and Always = 5). Contributing factors measured included demographics, physical functioning, neuro-cognitive measures, etc. A description of types and intensity of wandering and immediate adverse outcomes at baseline was conducted. The association between contributing factors and wandering type, intensity and immediate adverse outcomes at baseline was calculated using correlation coefficients (adjusting for inflation of Type I error due to multiple testing). An analysis of measures of wandering type and intensity over time for all respondents was conducted. To determine if patterns were impacted by baseline wandering behaviors the respondents were divided into three groups "non-wanderers", "low wanderers" (at or below the median score at baseline) and "high wanderers" (above the median score at baseline).


FINDINGS/RESULTS:
Sample description: A convenience sample of 143 dyads was recruited from outpatient clinics in the Tampa VA and the University of South Florida Memory Clinic with 98 dyads completing all four planned assessments. The most common reason for withdrawal (33.0%) from the study was death of the patient (n=13) or caregiver (n=2). Mean age of participants at baseline was 71.5 years (SD= 7.2), 92.3% were male (n=132), 90.2 % (n=129) were white and 86.0% (n = 123) were non-Hispanic. Baseline Characterization of Wandering: At baseline 70% (n=103) of the respondents reported persistent walking, 20.3% (n= 29) spatial disorientation, and 60.8% (n=87) repetitive walking, however, the frequency of the behaviors was low with a mean score on the multi-item scales being 1.4 (SD = 0.05), 1.2 (SD 0.49) and 1.5 (SD 0.60) respectively. The most commonly reported wandering behavior at baseline was "increased amount of spontaneous walking" with 48.2% (n = 66) reporting the behavior. A total of 56 (39.1%) demonstrated eloping behavior and 27 (19.0%) demonstrated mealtime impulsivity, again however, the frequency of these behaviors was low with a mean scores of 1.2 (SD = 0.44) and 1.2 (SD = 0.60) respectively. Nearly half (n = 70, 49.9%) of the respondents reported negative outcomes at baseline, although again the frequency of these events were low with a mean of 1.4 (SD = 0.06). After adjusting for multiple testing, no statistically significant associations were found between types of wandering or intensity of wandering and the measured potential contributing factors at baseline. Statistically significant associations were found between mean values of negative outcomes (falls, fractures and injuries) and the Behavioral Response to Stress, Aggressiveness (r = 0.34, p < .0001), Negative Verbalization (r = 0.32, p = .0001) and Passivity (r = 0.39, p < 0.0001) subscales, the Big Five Personality inventory Conscientiousness subscale (r = -0.31, p = 0.0002), the Modified Scale for Instrumental Activities of Daily Living (IADL) (r = -.35, p = <0.0001) and the Tinetti Gait & Balance Scale (r = -0.48, p < 0.0001) but not for the other immediate adverse outcomes. Wandering Over Time: A total of 113 respondents (84.6%) provided data at two or more times and will be included in the description of wandering over time. While wandering among individual Veterans changed somewhat across the period no discernible pattern over time was observed for measures of wandering type or intensity. A similar pattern was observed after spitting the sample into three groups based on their baseline wandering behaviors.

IMPACT:
This was the first prospective study of wandering behaviors and outcomes of newly diagnosed Veterans with mild dementia living in the community. The results demonstrated that wandering behavior was common in this population but occurred infrequently. For Veterans that completed all interviews, this pattern remained stable across the study period. Study findings should allay caregiver concerns about wandering in mild dementia.

---

External Links for this Project

---

PUBLICATIONS:

Journal Articles

1. King-Kallimanis B, Schonfeld L, Molinari VA, Algase D, Brown LM, Kearns WD, Davis DM, Werner DH, Beattie ER, Nelson AL. Longitudinal investigation of wandering behavior in Department of Veterans Affairs nursing home care units. International journal of geriatric psychiatry. 2010 Feb 1; 25(2):166-74. [view]

VA Cyberseminars

1. Moore DH. Supporting Family Caregivers of Rural Veterans with Dementia Who Wander. National Geriatrics and Extended Care Leads [Cyberseminar]. GRECC. 2016 Apr 27. [view]

Conference Presentations

1. Algase DL, Moore DH, Rowe M, Joseph IV. Hazardous Wandering: Challenges and Solutions. Presented at: University of South Florida Transforming Fall Management Practices Conference; 2010 May 3; Clearwater, FL. [view]

DRA: Aging, Older Veterans' Health and Care, Health Systems, Neurodegenerative Diseases, Mental, Cognitive and Behavioral Disorders
DRE: Diagnosis, Prognosis
Keywords: Caregivers – not professionals, Dementia, Research method
MeSH Terms: none