1. Provider-Caregiver Interactions: Evaluation and Use of the Therapeutic Relationship Index for Caregiver Interventions
Yeon Kyung Chee, Marie P. Dennis, Laura N. Gitlin
Community and Homecare Research Division
Thomas Jefferson University, Philadelphia
Paper Presented at the Annual Meeting of the Gerontological
Society of America, November 23, 2003
Funded by NIA R01 AG10947 and NIA/NINR U01 AG13265

2. Objectives
Examine psychometric properties of the Therapeutic Relationship Index (TRI)
Examine utilities of the TRI for in-home caregiver behavioral interventions

3. Effective Caregiver Interventions: In-home Interventions vs Traditional Medical Settings
Personal relationship with caregiver
Individualized approach/Customization
Sense of empowerment
Shared responsibility
Mittelman et al, 1996; Pusey & Richard 2001; Schulz et al, 2002; Clark et al, 1995; Gitlin & Gwyther, 2003; Toth-Cohen et al, 2001

4. Collaborative Treatment Approach
Patient-Centered Care (Institute of Medicine, 2001)
Responsive to patient values, preferences
Making health care decisions jointly with patients
Positive association with compliance:
Improved health status/role performance
Reduced health services utilization
Satisfaction with medical care
Holman & Lorig, 2000; Stewart, 2001; Von Korff et al, 1997; Greenfeld et al, 1985; Lorig et al, 2001; Roter 2000; Stewart 2001

5. Specific Aims
Evaluate psychometric properties of the Therapeutic Relationship Index
Examine changes over time and differences in TRI scores between intervention completers vs dropouts
Examine the relationship TRI scores and caregiver characteristics

6. Study Participants: Inclusion Criteria
Drawn from randomized clinical trials:
Recruited from Philadelphia Corporation on Aging, memory clinics, social services, media announcements
21 years of age or older
Primary caregiver of persons with a diagnosis of ADRD or MMSE <24
Living with care recipient
Deficiencies in 1+ ADLs
Deficiencies in 2+ IADLs

7. Study Participants Sample 1: Dementia Management Study
Sample 2: REACH I (Resources for Enhancing Alzheimer’s Caregiver Health) at the Philadelphia site
Environmental Skill-building Program (ESP)
Gitlin et al, 2001; Gitlin et al, 2003; Wisniewski et al, 2003

8. Study Participants Sample 1: 99 of 100 intervention group
8 did not receive ESP
27 dropouts (< 5 visits)
Refusal, Bereavement, Placement
No statistical or large differences between dropouts (n=27) vs completers (n=91)

10. Study Protocol Sample 1: Sample 2: Five 90-min visits over 3 months
Active: Five 90-min visits, one phone contact over 6 months
Maintenance: One 90-visit, three phone contacts over 6 months

11. Interventions Communication Setting up home environment
Managing difficult behaviors
Alleviating stress
Treatment documentation completed by interventionists within 48 hours following each visit

12. Therapeutic Relationship Index (TRI)
14 item, 5-point scale
Provider perceptions
Patient-centered care, dementia caregiving, ethical considerations in caregiving
Focus group discussions with occupational therapists
Higher scores = More collaborative

13. Measures of Construct Validity
Caregiver readiness (Prochaska & Velicer, 1997)
1 precontemplation
2 contemplation
3 preparation
4 action/maintenance
Caregiver anxiety (STAI)
Care recipient behavioral problems (RMBPC)
Care recipient cognitive impairment (MMSE)
Care recipient functional dependence (ADL)

14. Scale Structure Sample 2 Principal Components Model:
Factor loadings ranged from .50 to .84 (all 14 items)
Openness
Connectedness
Information-Seeking

16. Reliability Internally consistent: α = .78 (Sample 2)
Openness: α = .77
Connectedness: α = .66
Information-Seeking: α = .72
Six-month test-retest: Five visits (Sample 2)
Ranged .78 to .84
Intraclass correlations: Five visits (Sample 2)
α = .90

17. Construction Validity
Sample 2
Total index and subscales
Openness
Connectedness
Information-Seeking
Caregiving-related measures

19. Changes in TRI scores over time: Completers vs Dropouts
Sample 2
Repeated measures analysis with the Huyne-Feld solution
Significant interaction (F (1.95, ) = 9.89, p < .001)
Completers showed improvements (M = 42.02)
Dropouts showed a drop (M = 38.72)

22. Association of TRI scores to Caregiver Characteristics
Sample 2
Caregiver age, gender, race, income, education, length of caregiving, relationship to care recipient
Caregiver characteristics were not associated with TRI scores
(R2 = .090, F (7, 110) = 1.546, p = .159)

24. Conclusions and Implications
Lack of an adequate measure to evaluate therapeutic relationships for home-based interventions
14-item TRI is valid and internally consistent
Openness
Connectedness
Information-Seeking
TRI is associated with readiness, anxiety, behavioral problems not with functional status
Use of entire scale vs subscales: Therapeutic Relationships are complex and nuanced

25. Conclusions and Implications
Factors other than demographic profile influence the collaborative treatment process
TRI is sensitive to detect improvement or decline in therapeutic relationships with caregivers among completers and dropouts
For development of strategies to enhance compliance and customize intervention to match caregiver needs
Determine factors that contribute to poor collaboration and premature dropout from intervention
Examine relationships between TRI scores and treatment outcomes