1 Introduction
1.1 Scope of the problem
1.2 Smart homes to facilitate aging in place
1.3 Specific aim
2 HomeSense: Ambient sensing platform for health monitoring
2.1 Early system design and development
2.2 Community test bed and participant recruitment
Age (mean ± SD) | 75.6 ± 4.8 |
---|---|
Gender (% women) | 71 |
Race/Ethnicity (%) White | 100 |
Education (n, %) | |
High school | 3 (14) |
Some college | 3 (14) |
Associate’s degree | 2 (10) |
Bachelor’s degree | 8 (38) |
Master’s degree | 3 (14) |
PhD/MD | 2 (10) |
M ± SD | Min. | Max. | |
---|---|---|---|
# of medications | 3.6 ± 2.6 | 0 | 8 |
# of comorbid conditions | 2.0 ± 1.6 | 0 | 6 |
MMSEa | 29.7 ± 0.7 | 27 | 30 |
MCFSIb | 1.7 ± 1.3 | 0 | 4 |
MHI-5c | 89.1 ± 9.7 | 64 | 100 |
FSSQd | 4.3 ± 0.6 | 0 | 4 |
PSQIe | 6.2 ± 2.4 | 3 | 11 |
TUGf | 9.3 ± 4.1 | 6 | 18 |
ECOGg | 0.3 ± 0.6 | 0 | 2 |
Desmond Fall Risk Questionnaireh | 2.5 ± 2.2 | 0 | 8 |
AUDIT-Ci (% alcohol misuse) | 38% | ||
Nutrition Screeningj (% High Risk) | 33% |
2.3 System architecture
2.4 Web-enabled user interfaces
2.5 Case management reporting
All Participants | Age Cohorts | Osteoporosis Cohort | |||||
---|---|---|---|---|---|---|---|
Health Domain | (N = 15) Mean ± SD | Less than 70 (N = 8) Mean ± SD | 70+ (N = 7) Mean ± SD | P5 | Yes (N = 3) Mean ± SD | No (N = 12) Mean ± SD | P6 |
Hygiene | |||||||
Time spent in shower1 | 7.8 ± 6.1 | 7.9 ± 5.9 | 7.7 ± 6.8 | 0.95 | 8.1 ± 11.6 | 7.8 ± 4.7 | 0.93 |
Time using washing machine2 | 58.8 ± 37.8 | 41.3 ± 20.3 | 79.8 ± 43.4 | 0.04* | 47.6 ± 29.7 | 61.7 ± 39.9 | 0.59 |
Nutrition | |||||||
Fridge opening(s)3 | 17.5 ± 9.8 | 19.2 ± 9.5 | 15.6 ± 10.4 | 0.49 | 11.9 ± 10.1 | 18.9 ± 9.6 | 0.28 |
Pantry opening(s)3 | 5.4 ± 5.6 | 6.6 ± 6.4 | 4.3 ± 4.9 | 0.47 | 1.3 ± 0.8 | 6.5 ± 5.8 | 0.16 |
Microwave usage3 | 1.9 ± 1.6 | 1.7 ± 1.5 | 2.2 ± 1.8 | 0.57 | 2.1 ± 1.8 | 1.9 ± 1.6 | 0.04* |
Toaster usage3 | 1.2 ± 1.6 | 1.3 ± 2.0 | 1.0 ± 0.7 | 0.83 | 0.4 ± 0.3 | 1.5 ± 1.8 | 0.46 |
Time spent in kitchen4 | 2.6 ± 1.3 | 2.5 ± 0.9 | 2.6 ± 1.7 | 0.93 | 2.6 ± 0.5 | 2.6 ± 1.5 | 0.99 |
Bathroom Usage | |||||||
Time spent in master bathroom1 | 60.8 ± 41.1 | 68.3 ± 47.7 | 52.3 ± 33.6 | 0.48 | 41.5 ± 18.8 | 65.6 ± 44.3 | 0.38 |
Time spent in guest bathroom1 | 14.4 ± 19.5 | 7.5 ± 6.7 | 21.3 ± 25.8 | 0.20 | 37.0 ± 34.6 | 8.3 ± 7.7 | 0.29 |
Daily Activities | |||||||
Time spent in living room4 | 2.8 ± 1.2 | 3.3 ± 1.3 | 2.3 ± 1.0 | 0.14 | 2.0 ± 1.3 | 3.0 ± 1.2 | 0.19 |
Time spent using TV4 | 3.6 ± 3.1 | 2.3 ± 2.8 | 5.3 ± 2.9 | 0.79 | 3.9 ± 3.2 | 2.5 ± 2.9 | 0.04* |
Sleep disruption4 | 0.7 ± 0.7 | 0.5 ± 0.4 | 0.8 ± 0.9 | 0.04* | 1.4 ± 1.3 | 0.5 ± 0.3 | 0.04* |
Time spent in bed4 | 8.1 ± 2.2 | 8.0 ± 1.9 | 8.3 ± 2.6 | 0.79 | 7.3 ± 4.4 | 8.4 ± 1.4 | 0.05* |
Time spent outside4 | 3.8 ± 3.1 | 4.5 ± 3.3 | 3.0 ± 2.6 | 0.38 | 4.5 ± 5.9 | 3.6 ± 2.4 | 0.03* |
2.6 Notifications: A mechanism for health interventions in the short term
2.7 Communication with participants
2.8 Data visualization for understanding health and wellness related behaviors
2.9 Data analytics
3 Supplemental health assessments: A mechanism to elucidate the relationship between home behaviors and health in the long term
Type | # of Items | Short Description |
---|---|---|
Demographic questions | 14 | Assessing age, gender, education, employment, relationship status, residence, primary care status, chronic conditions, family history, self-reported health, medications, and community activity participation. |
DUKE-UNC Functional Social Support Questionnaire (FSSQ) [54] | 8 | Measure of functional support. Higher scores reflect higher perceived support. Cronbach’s α .85. |
RAND Social Health Battery [75] | 11 | Measure of social health and engagement where higher scores reflect higher engagement. Cronbach’s α .84. |
ECOG Performance Status [57] | 1 | Measure of ADL limitation where a higher score indicated increased limits in ADL ability. |
Desmond Fall Risk Questionnaire [58] | 15 | Assessment of falls risk where positive answer to 3 or more items indicates potential risk of falling. |
Alcohol Use Disorders Identification Test (Audit-C) [59] | 3 | Assessment of hazardous drinking and alcohol consumption. Scored on a scale of 0 to 12 with 3 or more indicating risk for women; 4 or more indicating risk for men. |
Alcohol, Smoking and Substance Involvement Screening Test (ASSIST) [76] | 2 | Measure of lifetime and recent tobacco use. Average test-retest reliability coefficients (kappas) had a high of .90. |
Nutrition Checklist [60] | 9 | Measure of risk for poor nutritional status; scores of 3 to 5 indicate moderate nutritional risk, 6 or more high nutritional risk. Sensitivity .72, specificity .94. |
Simplified Medication Adherence Questionnaire (SMAQ) [77] | 4 | Measure of medication adherence where higher scores indicate lower medication adherence. Cronbach’s α .85. |
Mental Health Inventory (MHI-5) [53] | 5 | Assessment of depression and anxiety with a cut score of 76. Cronbach’s α .84. |
Mini Mental State Exam (MMSE) [78] | 12 | Measure of cognitive impairment where lower scores indicate greater likelihood of cognitive impairment. Cronbach’s α .85. |
Cognitive and Affective Mindfulness Scale- Revised (CAMS-R) [79] | 12 | Assessment of mindfulness where higher scores indicate higher mindfulness. Cronbach’s α .81. |
Pittsburgh Sleep Quality Index (PSQI) [55] | 9 | Measure of sleep quality and sleep disturbances where a score of 5 or greater is indicative of poorer sleep quality. Sensitivity of 98.7 and specificity of 84.4 as a marker for sleep disturbances in insomnia patients versus controls. |
Timed “Up & Go” test [80] | 1 | Timed measure of physical mobility. Scores greater than 13.5 s indicate mobility challenges. Sensitivity 0.32 and specificity 0.73. |
Type | # of Items | Short Description | Administration Schedule |
---|---|---|---|
Self-Reported Activity and Health Status | 10 | Identifying changes in self-reported health, health-service utilization, changes in co-morbid conditions or medications, activity participation, travel plans and company at home | Ballots 1–4/Weeks 2–8 |
Patient Health Questionnaire (PHQ-2) [81] | 2 | Screening for potential depression symptomology. Sensitivity of .86 and specificity of .78 for diagnosing major depression | Ballots 1–4/Weeks 2–8 |
Physical Activity Scale (PASE) [82] | 21 | Measure of physical activity and mobility where higher scores indicate greater mobility and activity levels | Ballot 4/Week 8 |
Loneliness (UCLA Loneliness Scale) [83] | 20 | Assessment of loneliness where higher scores indicate greater loneliness | Ballot 2/Week 4 |
Falls Efficacy Scale [84] | 10 | Measure to assess fear of falling, scores over 70 indicate a fear of falling | Ballot 4/Week 8 |
Geriatric Pain Measure [85] | 24 | Assessment of pain in older adults, scores range from 0 to 42 with higher scores indicating more pain. | Ballot 1/Week 2 |
Telephone Interview for Cognitive Status (TICS) [86] | 11 | Phone-based assessment for cognitive status. Performance was significantly correlated with MMSE score (r = 0.86, p < 0.001) | Ballot 3/Week 6 |