Feasibility of Problem-Solving Training During Inpatient Rehabilitation in Patients With Stroke

Overview and Design

We conducted a single-group feasibility study of PST during the IRU stay of patients with stroke. A trained interventionist, an occupational therapy student, delivered PST to patients in the IRU. The student was trained by the first author (Candice L. Osborne) using a previously developed training protocol developed by the third author (Shannon B. Juengst), who has led multiple PST studies and developed the training protocol, which will be included in the final training manual (Juengst, Silva, et al., 2019). Interventionist training consisted of a combination of didactic instruction, self-study, and active practice with feedback. A member of the study team used a previously developed checklist to assess fidelity, screening 10% of the audio-recorded sessions. Intervention delivery adherence and competence were measured. A description of the fidelity checklist and calculations used were previously published by Juengst, Silva, et al. (2019). Participants received up to six PST sessions. If they were unable to complete all PST sessions during their IRU stay, the final session was completed over the phone within 1 wk of IRU discharge. Sessions were generally scheduled in the late afternoon in the patient’s room. Only the patient and interventionist were present. Outcome assessments were conducted at baseline, at the completion of PST, and at 3 mo postdischarge (final assessments were completed electronically or by phone). Postdischarge, participants received weekly text messages or emails for 3 mo that contained a link to the RedCap-based goal-setting boosters. Boosters prompted participants to set goals and report goal accomplishments once a week. The University of Texas Southwestern Medical Center Institutional Review Board approved all research procedures, and the participants provided informed consent.

Participants

Participants were patients who had received a primary diagnosis of stroke and who had been admitted to an IRU at an academic or county hospital. Electronic medical records were screened daily, and patients were approached about the study if they met the following criteria: (1) less than 4 wk since stroke, with a planned discharge home; (2) English fluency; (3) age 18 yr or older; (4) a Western Aphasia Battery score of 50 or higher; (5) a Montreal Cognitive Assessment (MoCA) score of 20 or higher; (6) the capacity to self-consent; and (7) the ownership of, and ability to use (independently or assisted), a smartphone.

Measures

To characterize the sample, we collected data on age, sex, race, ethnicity, education, and stroke severity using the National Institutes of Health Stroke Scale. Feasibility data (on discharge from the IRU; Aim 1) included the following: (1) number of participants recruited, (2) number and percentage of patients who consented, (3) reasons for refusal, (4) number of PST sessions completed, (5) length of sessions, and (6) scores on the Client Satisfaction Questionnaire–8 (CSQ–8; Atkisson & Greenfield, 1999), a validated self-reported measure of satisfaction with health- related services. At baseline, on discharge from the IRU, and at 3 mo postdischarge (Aim 1), we collected data using (1) the General Self-Efficacy Scale (GSE), a 10-item self-report measure of self-efficacy as it relates to coping with stressful events, validated with a wide range of rehabilitation populations (Luszczynska et al., 2005), and (2) the Patient Health Questionnaire–8 (PHQ–8), an eight-item questionnaire validated among patients with stroke (Trotter et al., 2019) that measures the frequency of eight symptoms from the Diagnostic and Statistical Manual of Mental Disorders (4th ed., text revision; American Psychiatric Association, 2000) that define a major depressive episode in the past 2 wk. Goal Attainment Scaling (GAS; Kiresuk & Sherman, 1968) was conducted on discharge from the IRU (Aim 1) and weekly for 3 mo (Aim 2). GAS is a method for developing patient-centered goals and measuring goal achievement that has demonstrated validity, reliability, and sensitivity in an IRU population (Malec, 1999). At 3 mo postdischarge from the IRU, we also collected data using the Telehealth Usability Questionnaire (TUQ; Aim 2), a 21-item questionnaire that has been validated across a range of clinical patient populations and measures technology usefulness, ease of use, effectiveness, reliability, and satisfaction (Parmanto et al., 2016).

Problem-Solving Training Intervention

PST is a metacognitive strategy technique, which means that the training focuses on a systematic approach to thinking through the steps of problem solving rather than focusing on solving a specific problem. PST was provided in addition to regular inpatient rehabilitation treatment and consisted of up to six sessions (30–45 min each) that follow a structured format (for session content, see Table 2; Juengst, Silva, et al., 2019). Under the guidance of the interventionist, participants applied the ABCDEF steps of PST (Assess, Brainstorm, Consider and Choose, Develop and Do, Evaluate, and Flex) to develop specific action plans to overcome barriers to the goals they choose. The Discharge Planning Assessment Tool (DPAT; Texas Woman’s University, 2020) was administered during Session 2 to identify specific areas where the patient perceived the potential for problems or barriers that might impede a successful transition from hospital to home.

Electronic Boosters

Electronic boosters consisted of weekly check-ins where participants stated whether they achieved their goal for the week (responses were “yes,” “no,” or free text). If they did not accomplish the goal, they were provided a drop-down menu from which they could select a reason why they did not accomplish the goal, or they could provide a free-text answer. The third prompt asked whether the participant would like to continue working on the previous week’s goal or create a new goal. If the participant created a new goal, they were prompted to describe the new goal using free text. Weekly boosters took approximately 5 min to complete.

Statistical Analysis

To describe feasibility, we report the number and percentage of recruited participants, number of participants who gave consent, reasons for ineligibility, reasons for refusal, and number and length of PST sessions. To identify a change in clinical outcomes, we present the average preintervention–postintervention changes in scores for depressive symptoms (on the PHQ–8) and self-efficacy (on the GSE) for changes both from baseline to discharge and from baseline to 3 mo postintervention. We also present the summary statistics (M ± SD) of participant satisfaction with PST (recorded using the CSQ–8). For Aim 2, we present the percentage of attained goals (recorded using GAS) out of all the attempted goals for each participant’s summary statistics (M ± SD) for technology usability (recorded using the TUQ).

Feasibility of the Intervention

Figure 1 shows the flow of participants through the study, from the total number of patients who were screened between October 2019 and June 2020 to the total number who gave consent, with reasons for ineligibility and refusal. Compared with the patients who were screened, those who gave consent were more likely to be African American and less likely to be Hispanic. Table 1 presents demographic information for participants who consented. Four (27%) participants withdrew: 1 after three PST sessions; 1 after discharge, stating that she did not want to participate any further; and 1 for unknown reasons during the first PST session. We withdrew 1 participant because of worsening medical conditions that affected her cognition before the first PST session. Of those who did not withdraw, 1 was lost to follow-up at 3 mo. Eleven participants completed three or more PST sessions. The PST sessions lasted, on average, 30.46 min.

OPEN IN VIEWER

Figure 1.

Participant flowchart.

Participants (n = 10) were generally very satisfied with PST (CSQ–8, M = 29.3, SD = 4.4). Participants reported moderate depression, on average, at baseline (PHQ–8; M = 11.0, SD = 6.1), decreasing to mild depression at the time of discharge (M = 5.6, SD = 3.8) and at 3 mo postdischarge (M = 8.3, SD = 5.5). Participants rated their self-efficacy at baseline as moderately high (GSE; M = 31.1, SD = 7.3), with an increase at the time of discharge (M = 36.2, SD = 3.3) and at 3 mo postdischarge (M = 34.1, SD = 4.2). Figure 2 shows changes in depression and self-efficacy scores from pre-PST to post-PST for each participant. All participants were able to set measurable goals with guidance from the interventionist, and 81% of the participants (n = 11) who set at least one goal during the hospital stay achieved at least one goal.

OPEN IN VIEWER

Figure 2.

Individual participant change in depressive symptoms and self-efficacy.

Feasibility of Electronic Booster Follow-Up

Five participants (54%) used electronic boosters at least once. Overall, the participants who used boosters were satisfied with them (85%; TUQ–Satisfaction, M = 6.0, SD = 1.4; range = 1 [disagree] to 7 [agree]), although they found boosters to be only moderately useful (67%; TUQ–Useful, M = 4.9, SD = 2.9). Most participants found it easy to use (85%; TUQ–Ease of Use, M = 6.3, SD = 1.2), were satisfied with the level of interaction with the interventionist (85%; TUQ– Interaction with Therapist, M = 6.6, SD = 1.6), and were satisfied with the interface quality (85%; TUQ– Interface Quality, M = 6.3, SD = 1.5).

Participant Eligibility

The most common reason for ineligibility was lack of fluency in English. Most of the patients who were excluded spoke Spanish, reflecting the growing number of Spanish speakers who live in and receive rehabilitation services in the United States. To prevent ongoing health disparities that the lack of Spanish-language services may create, we have translated PST into Latin American Spanish (Vega et al., 2020), with pilot testing and cultural adaptation underway.

Feasibility of Participant Retention

We aimed to enroll 15 participants with a goal of each completing at least three PST sessions before IRU discharge. Although PST is generally provided in six sessions, previous studies with other populations demonstrated efficacy after only three sessions and also demonstrated that neither the number of hours devoted to PST nor the length of follow-up was associated with treatment effect sizes, suggesting that the core principles of PST may be rapidly incorporated into daily life and that the effects are maintained over time (Elliott & Berry, 2009; Juengst, Osborne, et al., 2019). Most of the patients we approached consented to participate (88%), and most completed three or more PST sessions during their IRU stay. The main reason for withdrawal from the study was feeling overwhelmed or too tired to participate after a day of therapy. Although patients generally reported high satisfaction with PST, it may be beneficial in the future to include their care partners in a dyadic PST intervention, leveraging the previously documented benefits of collaborative problem solving (Portillo et al., 2009). Patients may feel more supported in the problem-solving process if a care partner is also participating. We previously assessed the feasibility of PST in care partners during the patient’s IRU stay, with promising results (Juengst, Osborne, et al., 2019).

Participants who withdrew were younger, more often female, and reported less independence with ADLs before the stroke than those who completed the study. Notably, all who withdrew had a right-sided hemorrhagic stroke and were all of the participants with hemorrhagic stroke who had been recruited. Although the sample is small, this suggests that those with poorer awareness of their deficits—which is most common in the early stages of recovery after a right-sided stroke (Jehkonen et al., 2006)—may not see the potential benefit of this type of intervention.

Feasibility of the Problem-Solving Training Intervention on an Inpatient Rehabilitation Unit

A short length of stay was the primary reason why patients received fewer than six sessions. This could be mitigated in the future if the PST sessions are continued throughout the transition from hospital to home by means of phone or video calls. Previous work in this area supports the efficacy and feasibility of a phone-based PST intervention for patients with acquired brain injury (Bell et al., 2011).

The DPAT was used as part of the PST intervention to identify and assess potential problems. We found that it provided a structured and comprehensive format for patients to consider and evaluate the problems they anticipated upon returning home. Patients set goals to address the following areas: driving, grooming, transferring to and from the shower, passing through narrow doorways, carrying items while using a walker, and sit-to-stand transfers.

Feasibility of Problem-Solving Training Mobile Boosters

The mHealth booster sessions, which were designed to encourage participants to continue setting goals and implementing PST steps after discharge, were less well received. Only half of the participants tried the booster at least once. Those who used it were mostly satisfied but found it only moderately useful. We trained participants to use boosters during their last in-person PST session. Introducing boosters earlier may reduce confusion and encourage participants to establish a routine. Reminders may also increase compliance. Past studies using similar mHealth platforms trained participants for 2 to 4 wk and closely monitored compliance, assisting with any issues that arose during the training (Nussbaum et al., 2019; Powell et al., 2016). Implementing these changes may encourage more consistent use of boosters throughout the transition from hospital to home and may provide more one-on-one support for booster use. Patients with stroke may also face unique challenges in using mHealth, and although developments in wireless technology have addressed some key factors (e.g., enlarged text, screen brightness, voice dialing), patients with stroke report having difficulty using small screen displays and buttons, as well as confusion because of unclear symbols and complex features (Morris et al., 2010). Developing more user-friendly boosters may also increase compliance.

Study Limitations

Although our results support the feasibility of PST during the IRU stay for patients with stroke, this was a pilot study in a single setting; therefore, the findings cannot be applied across all settings. We could greatly expand our recruitment pool and extend the application of our results by including Spanish speakers, using a validated translation of PST. Twenty percent of the participants withdrew from the study (we withdrew 1 participant). Those patients tended to have greater functional impairment before the stroke and felt overwhelmed by the intervention after a long day. A dyadic intervention that includes the patient and their care partner could provide additional support for the patient during the intervention in the future. Although mHealth boosters were well received by those who used them, compliance was low. We identified several strategies to combat this issue in future studies, including earlier training and supervised implementation. More research into how to best leverage smartphones to support patients throughout the transition from hospital to home is warranted. Last, volunteer bias may have positively skewed the results, because those who are more motivated to participate may also be more motivated to complete proactive problem-solving tasks.