Feasibility Trial of Tailored Home Modifications: Process Outcomes

Research Design

This study reports the outcome of a process evaluation conducted with a double-blind, sham-controlled RCT designed to evaluate the effectiveness of a tailored home intervention to improve daily activity performance and reduce falls at home among community-dwelling older adults who had experienced a recent fall. The study was approved by the Human Research Protection Office at Washington University.

Participants

Community-dwelling older adults age 65 yr or older who had a history of falling within 6 mo were recruited from two sources: the Barnes-Jewish Hospital (St. Louis, MO) Emergency Department (ED), a Level 1 trauma center, and flyers in the community. Older adults visiting the ED because of a fall were identified by geriatric screeners or an automatic paging system. Screening and consent occurred in the ED or at home after discharge. A sample of 2,185 older adults was identified through the ED. A sample of 533 older adults was identified through the flyers and subsequent screening over the telephone. All potential participants were blinded to the outcome of the trial and thus were invited to participate in a trial of daily activity performance.

Potential participants with a recent fall and impairment in activities of daily living (ADLs) and in instrumental ADLs (i.e., impairment in two or more daily activities from the Older Americans Resources and Services Multidimensional Functional Assessment Questionnaire [Fillenbaum, 1988] ADL scale) who lived in Missouri within 30 miles of Barnes-Jewish Hospital were eligible. Participants were excluded if they reported that their fall was caused by syncope, had moderate to severe memory problems demonstrated by a score of 10 or less on the Short Blessed Test (Carpenter et al., 2011), had chronic alcohol abuse demonstrated by a score of 2 or higher on the Short Michigan Alcoholism Screening Test–Geriatric Version (Blow, Gillespie, Barry, Mudd, & Hill, 1998), were unable to ambulate independently, resided in a nursing home, or were diagnosed with a neurological condition that resulted in motor performance difficulties (e.g., Parkinson’s disease, recent stroke). A total of 115 participants consented to the study.

Assessment

Although assessment information is fully described elsewhere (Stark, Keglovits, & Somerville, 2016), briefly, participants underwent assessment during a home visit at baseline and at 6 and 12 mo after intervention. All raters were trained occupational therapy practitioners blinded to group allocation throughout the study. Balance and gait were assessed using the Performance-Oriented Mobility Assessment (Tinetti, 1986). Each participant’s prescription medications were recorded. Fall-related indicators included the number of self-reported falls and injurious falls in the past year. The In-Home Occupational Performance Evaluation (Stark, Somerville, & Morris, 2010), a valid and reliable self-report and performance-based assessment of in-home activity performance, was used to quantify the problematic activities for the older adult, current performance level and satisfaction with performance, and objective activity performance in the home environment.

Randomization allocation concealment was achieved by automated randomization using a block size of six. The allocation ratio was 1:1, with blocking intervals of four. Groups were balanced with regard to race and sex. Randomization assignment was elicited after baseline assessment.

Process Outcome Measures

To aid in future dissemination of the intervention, we framed the process evaluation using the RE–AIM framework. We assessed the reach of the intervention by examining the willingness of people to participate in the study. Effectiveness of the intervention to improve daily activity performance was examined. Adoption was measured by examining adherence to the recommended interventions immediately and assessing self-management behaviors such as obtaining additional environmental supports after treatment. The implementation of the intervention, or fidelity to the protocol, was examined by calculating dose (time) and frequency of the intervention. A visit checklist was completed for each element of each visit. Maintenance, or the extent to which participants were adherent to the intervention over a 12-mo follow-up period, was obtained for each home modification strategy provided during the intervention.

Treatment time, goals, and accomplishments were recorded for each session. Costs of the study intervention (including time and materials) were recorded for the intervention group. These methods to evaluate process adhere to the National Institutes of Health framework for the conduct and reporting of dissemination and implementation of research (Neta et al., 2015).

Treatment and Assessment

All treatment participants were scheduled to receive six 90-min sessions of treatment with a certified occupational therapy practitioner delivered over an 8-wk period. Scripted study events were scheduled for each session. The sham group received equal time from an occupational therapist who provided a standardized kit of adaptive equipment for fine motor tasks (e.g., needle threader, large-grip pizza cutter). Treatment time and goals were recorded for each session. Costs of the study intervention (time and materials) were recorded for the treatment group.

Home Modification Intervention

The purpose of the tailored home modification intervention was to address daily activity limitations by providing environmental support to compensate for functional impairments. Supplemental Figure 1 shows examples of home modification interventions (available online at http://otjournal.net; navigate to this article, and click on “Supplemental”). The intervention targeted participants’ problematic daily activities. Participants used a systematic process to identify and address these daily activity problems. The standardized components of the intervention included

The intervention is designed to be delivered (after baseline assessment) over six treatment visits lasting 90 min each, with approximately 1 wk between visits. Figure 1 outlines the objectives for each visit.

OPEN IN VIEWER

Figure 1.

Tailored home modification intervention.

The treatment theory, or mechanism of change, guiding this intervention is the competence-press model (Lawton & Nahemov, 1973), which postulates that changes to the environment (e.g., a railing in a stairwell) matched with a person’s personal abilities (e.g., poor balance) improve activity performance (e.g., safely descending the stairs). The essential ingredient of the intervention is the delivery of environmental support and an education and self-management component (Dobkin, 2004) to learn to use environmental solutions and perform activities safely in the home. The active ingredients of the intervention include motivational enhancement techniques (to support a client-centered decision process and to increase adherence with intervention recommendations; Cummings, Cooper, & Cassie, 2009) and shared decision making (to consider the best evidence and the participant’s values and preferences; Schwarzer, 2008).

The manualized intervention is available from the authors. The intervention begins with a performance-based assessment of in-home activities to identify problematic daily activities and environmental barriers (Stark et al., 2010). The competence-press framework is then used to generate potential environmental solutions. These solutions are refined using a clinical reasoning algorithm to consider unique environmental circumstances (e.g., available social support, layout of home) and personal preferences and values (e.g., concern for aesthetics, use of personal assistance; Stark, Somerville, Keglovits, Smason, & Bigham, 2015).

The final set of solutions for each problematic daily activity is presented to the participant using a shared decision-making tool (i.e., a clinical decision analysis [CDA] worksheet; see Supplemental Appendix A, available online at http://otjournal.net; navigate to this article, and click on “Supplemental”). The tool includes the outcomes, benefits, and liabilities of each solution. A script is provided to interventionists to deliver the CDA to the participant and family and includes a specific algorithm to determine the final home modification intervention. Home modifications are obtained and installed. Active practice of daily activities includes an explanation of the purpose of the modification or equipment, demonstration of use, supervised participant practice, caregiver training (if applicable), and problem solving until the participant is able to safely complete the activity. Intervention plans were reviewed by all intervention staff weekly to ensure consistency between interventionists.

Interventionist Training

Study interventionists (n = 7) included both occupational therapists and certified occupational therapy assistants. Interventionists participated in 15 hr of training that comprised self-directed reading, didactic classroom lecture, case studies, and mentored learning in the field.

Analyses

All data entry was completed using REDCap, a secure web-based application (Harris et al., 2009). All data analysis was completed using IBM SPSS Statistics (Version 21; IBM Corp., Armonk, NY). Descriptive analysis was used to describe demographic, baseline characteristics of both groups; fidelity to the intervention protocol; cost; and adherence to and benefit of treatment. An independent t test was used to examine whether there was a significant group difference of characteristics at baseline. We also examined whether there was a difference between the enrolled participants and those who dropped out before the 6-mo postintervention assessment.

Daily activity performance measures were collected at baseline and at 6 mo and 12 mo after intervention completion. A factorial mixed analysis of variance for repeated measures was used to analyze significant effects of group, time, and interaction of time with groups. The criterion for significance is at the probability level of .05 in all tests. In the presence of a significant main effect, Bonferroni post hoc pairwise comparisons were explored.

Reach

The flow of participants through the study is provided in Figure 2. A total of 115 participants consented to the study. The 23 dropouts were compared with the 92 remaining participants on their baseline scores for demographic and outcome variables. There were no differences between groups except for age; younger participants were more likely to drop out. Reasons for not participating in the follow-up included illness, death, becoming ineligible (moving to nursing home), no longer being able to be contacted, withdrawal by the principal investigator (S. Stark; did not disclose exclusion criteria at screening), and declining to continue. Demographic characteristics of the participants are provided in Table 1. Participants were characteristic of older adults at high risk for falling (Tinetti & Williams, 1997; e.g., older, poor balance and strength, previous history of falls). There were no significant differences between groups regarding the number or severity of falls in the previous year.

OPEN IN VIEWER

Figure 2.

Study participant flow.

Mechanism of Fall Prevention: Improve Daily Activity Performance

To determine whether the targeted mechanism of the home modification intervention was effectively delivered, daily activity performance was assessed at baseline and at 6 and 12 mo after intervention (Table 2). All of the activity performance subscores had significant interaction effect of group × time, indicating that home modification participants demonstrated significantly improved subjective and objective activity performance scores compared with controls over time, which suggested that the intervention was successfully delivered. There was a significant main effect of group, F = 8.9, p < .01, and time, F = 56.17, p < .01, on self-rated performance. Post hoc tests indicated that regardless of group, there was a significant improvement between baseline and 6 mo and between baseline and 12 mo. However, there was no significant difference between 6 mo and 12 mo. Self-rated performance scores were higher after the treatment. There was a significant interaction effect of group × time, indicating that change in self-rated performance differed by group over time. The home modification group showed significantly greater improvement in self-rated performance.

There was also a significant main effect of group, F = 5.4, p = .02, and time, F = 99.3, p < .01, on self-rated satisfaction with performance. The home modification group had a higher satisfaction score than the sham control group. Satisfaction scores were higher after the treatment. Similar to self-rated performance, participants did not show significant improvement between 6 mo and 12 mo in satisfaction with improvement. A significant interaction effect of group × time indicates that the home modification group had a greater change in satisfaction with performance compared with the sham group over time.

The objectively rated activity performance scores were significantly improved after intervention. Performance scores did not reach zero, indicating that barriers were still present in participants’ homes. There is no main effect of group on objective performance scores. There was a significant main effect of time, F = 30.6, p < .01, indicating a significant decrease of barriers between baseline and 6 mo and between baseline and 12 mo but not between 6 mo and 12 mo. The interaction effect showed a difference between the two groups over time. Contrasts were performed to compare total barriers of two groups in each time point. There was no statistical difference between groups at 6 mo, F = 3.9, p = .052, and 12 mo, F = 3.3, p = .073, for objective performance.

Intervention Protocol Fidelity

The intervention protocol stipulated that six 90-min visits be completed in participants’ homes over the course of 60 days (8 wk). The participants in the treatment group received an average of 5.9 sessions (range: 4–9 sessions) for 84.39 min (range: 48–121 min), completed over the course of 91.24 days (13 wk). Two participants with protocol deviations as a result of winter travel to Florida for 5 mo and an inpatient hospital stay were excluded from the analysis. The participants received an average of 496.43 min (range: 315–965 min) compared with 540 min of planned treatment.

All treatment was completed in participants’ homes. On a procedure checklist completed after each session, interventionists reported carrying out more than 90% of all session components according to the protocol. All participants received a tailored prescription and home modification interventions (the essential ingredient of the intervention). Interventionists reported no major deviations from the protocol and no unanticipated events or serious adverse events during the treatment period. Reasons for not completing the intervention within 8 wk included participant health (including hospitalization); shipping delay for adaptive equipment; delay of modification installation; and participant travel, bereavement, and holiday activities.

Cost

The mean cost of materials and labor installation for each participant in the intervention group was $931, with a range of $120–$2,790, with an average of 6.37 problematic activities addressed per participant. Therapist time to deliver the intervention averaged 85 min per session (a total of 510 min per participant). Before delivering the intervention, therapists participated in 8 hr of self-directed, didactic, and online learning about the delivery of the home modification intervention (including certification in assessments).

Home Modification Adherence and Benefits

Adherence was calculated as the proportion of intervention strategies used divided by the intervention strategies implemented. Participants reported using 94% of received home modifications at baseline and 91% at 12 mo. Participants who did not have 100% adherence reported various reasons for not accepting or using their modifications: Controls were too difficult to understand or adjust, the modifications or strategies were forgotten, and the items were no longer needed. Participants scored intervention usefulness between 2 (somewhat useful) and 3 (very useful) both immediately after the intervention (2.77, standard deviation [SD] = 0.24) and at 12 mo after intervention (2.84, SD = 0.21). A subset of 23 participants (46.9%) reported independently obtaining additional home modifications after the completion of treatment.