Journal of Gerontological Nursing

Feature Article Supplemental Data

Feasibility of Motivational Interviewing to Engage Older Inpatients in Fall Prevention: A Pilot Randomized Controlled Trial

Hiroko Kiyoshi-Teo, PhD, RN; Kathlynn Northup-Snyder, PhD, RN, CNS; Deborah J. Cohen, PhD; Nathan Dieckmann, PhD; Sydnee Stoyles, MBST; Elizabeth Eckstrom, MD, MPH; Kerri Winters-Stone, PhD

Abstract

In the current 3-month, two arm, unblinded, single site, pilot randomized controlled trial, 120 high fall risk hospitalized older adults (age ≥65) were contacted, and 67 participants were enrolled. The intervention arm received a brief motivational interviewing (MI) intervention. Both arms received routine hospital fall prevention protocols. Measurements were conducted at baseline, 2 days, 1 week, 1 month, and 3 months. MI intervention took approximately 21 minutes and was provided at beginning proficiency level. Approximately 66% of participants completed 3-month data collection. The intervention group reported a greater decrease in fear of falling after the intervention than the control arm (β = −0.856 vs. β = −0.236) and maintained fall prevention behaviors at 3 months (β = 0.001 vs. β = −0.083) (p < 0.05). The current study found brief MI for fall prevention in acute settings feasible and provided preliminary evidence for a positive impact of MI [Journal of Gerontological Nursing, 45(9), 19–29.].

Abstract

In the current 3-month, two arm, unblinded, single site, pilot randomized controlled trial, 120 high fall risk hospitalized older adults (age ≥65) were contacted, and 67 participants were enrolled. The intervention arm received a brief motivational interviewing (MI) intervention. Both arms received routine hospital fall prevention protocols. Measurements were conducted at baseline, 2 days, 1 week, 1 month, and 3 months. MI intervention took approximately 21 minutes and was provided at beginning proficiency level. Approximately 66% of participants completed 3-month data collection. The intervention group reported a greater decrease in fear of falling after the intervention than the control arm (β = −0.856 vs. β = −0.236) and maintained fall prevention behaviors at 3 months (β = 0.001 vs. β = −0.083) (p < 0.05). The current study found brief MI for fall prevention in acute settings feasible and provided preliminary evidence for a positive impact of MI [Journal of Gerontological Nursing, 45(9), 19–29.].

Sustaining an injury from falling while hospitalized continues to be a major issue for older adults. Fall rates are as high as 8.9 falls per 1,000 bed-days (Miake-Lye, Hempel, Ganz, & Shekelle, 2013), and one third of these falls are considered preventable (Cameron et al., 2010). Significant efforts have been made to decrease inpatient falls and fall-related injuries partly because hospitals are not reimbursed for fall-related injuries (Agency for Healthcare Research and Quality [AHRQ], 2013a). Many hospitals have multi-modal fall prevention programs that include standardized fall risk assessments, universal and tailored fall prevention education, and post-fall assessments (AHRQ, 2013b). Past efforts have not focused on patient engagement in fall prevention, and studies have found that older adults perceive fall prevention recommendations as irrelevant and threatening to their independence (McMahon, Talley, & Wyman, 2011). Patient engagement strategies have been successful in other patient safety areas (AHRQ, 2018), and could be adapted to fall prevention (Tzeng & Yin, 2015).

Motivational interviewing (MI) is an evidence-based approach to facilitate patient behavior change by engaging patients in the change process through open-ended questions, affirmation, and reflection on patients' beliefs (Miller & Rollnick, 2012). MI is effective in fostering behavior change with numerous health behaviors, including smoking cessation and weight loss (Copeland, McNamara, Kelson, & Simpson, 2015; Lundahl et al., 2013).

Studies suggest that MI may also support adoption of new behaviors for acute care topics, such as post-stroke medication adherence (Barker-Collo et al., 2015), self-care among cancer patients (Masterson Creber et al., 2016), and behavior change for those with heart failure (Riegel, Masterson Creber, Hill, Chittams, & Hoke, 2016). MI is especially promising for clinical nursing because older adults are often resistant toward making behavior changes for fall prevention (McMahon et al., 2011). MI has the capacity to facilitate behavior change with patients who have varied levels of interest in change and can address multiple fall prevention sub-behaviors. Limited evidence supports the efficacy of MI for fall prevention with community-dwelling older adults to promote physical activity (McMahon & Fleury, 2012; Reuben et al., 2017).

The goal of the current pilot study was to determine the feasibility of a brief MI intervention with hospitalized Veterans for fall prevention. The authors examined the feasibility of recruitment and retention, delivery of the MI intervention, and collection of the study measures (e.g., fear of falling, importance and confidence related to fall prevention, patient activation, fall prevention behaviors, fall rates). In addition, the authors explored the potential efficacy of brief MI in changing perceptions and behaviors of hospitalized older adults for fall prevention.

Method

Design

The authors conducted a 3-month, two arm, unblinded, pilot randomized controlled trial. The treatment arm received one brief MI intervention during hospitalization. The treatment and control arm received routine hospital fall prevention protocols. Measures were collected at baseline, 2 days, 1 week, 1 month, and 3 months.

Setting and Sample

The current study was conducted on three medical-surgical units at a Veterans Affairs (VA) hospital in the northwestern United States and approved by the Oregon Health and Science University Institutional Review Board. At this site, all inpatients received universal fall precaution, including environmental safety, orientation to hospital surroundings, and education to call for assistance when needed. In addition, patients who were identified as high fall risk were visited every 2 hours or given bed exit alarms.

Patients were eligible for the current study if they were age ≥65, high fall risk as indicated by Morse Fall Scale (MFS) score (≥45 on the most recent nursing documentation), and hospitalized for at least 24 hours on medical-surgical units. In addition, patients had to be alert and oriented to time, place, and person, and able to carry on a verbal conversation in English. Critical care and psychiatric units were excluded as routine hospital fall prevention differs on these units. No sample size calculations were performed for the current pilot study. The authors aimed to enroll 70 participants, with at least 20 participants from each hospital unit. The authors considered this to be sufficient to test the feasibility and practicality of study procedures and measures.

Study Procedure

All newly admitted patients to the study units received a study invitation letter by hospital staff as part of their routine hospital admission process during the study recruitment period (January to August 2016). A study team member (H.K.T.) and a research assistant approached the newly admitted patients who met the screening criteria to invite them to participate. After signing consent, they were immediately assigned a study ID.

To ensure balance in each arm from the three hospital units, computer generated randomization was conducted in blocks of 10. Study team members and patients were not blinded to their arm assignments. Within 24 hours of consent, baseline data were collected. Study staff members administered survey questions at bedside at a patient preferred time and entered responses into a secure HIPAA-compliant online REDCap database (n.d.).

Brief MI interventions were conducted after obtaining consent. MI interventions were audio recorded with permission from participants and transcribed by research staff. After discharge from the hospital, participants responded to study measures over the telephone during the 3-month follow-up period. Chart reviews were conducted to gather demographic variables. No incentives were offered for the current study.

MI Intervention Arm

The intervention arm received brief MI for fall prevention at the baseline visit. The MI-Based Communication Guide for Fall Prevention (Table 1) includes semi-structured questions and prompts based on MI approaches that are adapted for fall prevention. MI approaches, such as open-ended questions, prompts for evoking behavior change, affirmations, and reflections with a focus on fall prevention (Miller & Rollnick, 2012) were included. The structured part of the MI intervention ensured that permissions were sought from patients before asking questions or offering information. The MI intervention also included the following two questions: “What does fall prevention mean to you? and “What are one or two things that may be useful or meaningful to you for us to discuss about fall prevention?” For Question 2, the Menu of Options for Fall Prevention (Table A, available in the online version of this article) was shared so that participants could choose a specific fall prevention topic for discussion. An important part of the intervention was to let participants identify specific behaviors that could help reduce falls (e.g., using a walker, asking for help) rather than discussing the umbrella concept of fall prevention.

Motivational Interviewing (MI)–Based Communication Guide for Fall Prevention: MI Approaches to Address Fall Prevention and Rationales

Table 1:

Motivational Interviewing (MI)–Based Communication Guide for Fall Prevention: MI Approaches to Address Fall Prevention and Rationales

Menu of Options for Fall Prevention

Table A:

Menu of Options for Fall Prevention

All MI interventions were provided by the principal investigator (PI) (H.K.T.), a RN with ≥40 hours of MI training by a Motivational Interviewing Network of Trainers (MINT) trainer. The coinvestigator reviewed the current study materials to ensure alignment of the study processes with MI concepts and provided biweekly MI coaching for the PI during the data collection period.

Control Arm

Control arm participants received a fall prevention intervention as part of their routine hospital stay and interacted with study staff members for purposes of data collection.

Measures

Details of the current study measures are reported elsewhere (Kiyoshi-Teo et al., 2019).

All participants were surveyed at bedside for cognition (Montreal Cognitive Assessment-Basic [MOCA-B]) (Nasreddine, 2014), fear of falling (Falls Efficacy Scale International-Short [FESI-S]) (Kempen et al., 2008), fall prevention daily behaviors (Modified Falls Behavioural [M-FaB] Scale) (Clemson, Bundy, Cumming, Kay, & Luckett, 2008; Clemson, Cumming, & Heard, 2003), and level of engagement with their own health (Patient Activation Measure [PAM]) (Hibbard & Greene, 2013; Hibbard, Mahoney, Stockard, & Tusler, 2005). Permissions to use or modify all study instruments were obtained prior to data collection. In addition, participants were asked about their fall history and to rate the importance related to their confidence and fall prevention.

The original Falls Behavioural Scale was modified for the current study to capture fall prevention behaviors relevant to participants' specific living situations. This modification was possible because the scale score is a calculated mean of all responses. The goal of this modification was to capture the frequency of context-specific behaviors to prevent falls as patients transition through different care settings (e.g., hospital, home, care facility) depending on their health status.

Thus, the authors used an 18-item or 29-item version of the M-FaB depending on participants' living situation. The 18-item M-FaB (i.e., the inpatient version of the Falls Behavioural Scale) has been previously published (Kiyoshi-Teo et al., 2019) and included questions applicable to fall prevention behaviors relevant for hospitalization, such as checking to see if anything is attached to self before getting up (e.g., oxygen tubes, intravenous lines). The 29-item M-FaB included inpatient questions and 11 additional post-discharge questions to capture fall prevention behaviors specific to the home or care facility. Post-discharge questions included: checking the soles of shoes when buying shoes, crossing the street only where there is signage, and not using a ladder (Table B, available in the online version of this article).

Modified Fall Behavioral Scale: Inpatient and Post-Discharge Questions

Table B:

Modified Fall Behavioral Scale: Inpatient and Post-Discharge Questions

A chart review was conducted to gather baseline demographic variables such as age, gender, education, diagnoses, medications, and mobility. The numbers of diagnoses and medications were used as surrogate measures of disease burden. The AHRQ High Fall Risk Medication Score was used to assess risk of falls due to medications (≥6 is considered high fall risk) (AHRQ, 2013c).

A fall was defined as an unplanned descent to the floor with or without injury to patients (National Dataset of Nursing Quality Indicators, 2012). At baseline, participants were asked about any fall incidents that occurred in the previous 3 months and in the past 1 year. During the follow-up period, participants were asked if and how many times they had fallen since the last study encounter.

Analysis

The feasibility of the current study was examined by descriptive statistics (frequency and standard deviation) for recruitment and retention of study participants. In a meta-analysis of fall prevention interventions, the median recruitment rate was 48.5% and attrition rate was 16.2% at 12 months using study data of hospital inpatients and residents of nursing care facilities (Nyman & Victor, 2011). Thus, considering the acutely ill state of current study participants, the authors conservatively considered recruitment and retention rates ≥50% to be a success.

All MI interventions were timed, recorded, and transcribed, and field notes were completed to assess quality of the intervention. Four external MINT members trained on the MI Treatment Integrity Coding Manual (MITI) 4.1. (Moyers, Manuel, & Ernst, 2014) assessed approximately 30% of recorded MI interventions to determine fidelity. For the current study, beginning proficiency quality was considered appropriate.

The authors tested study outcome measurements (FESI-S, importance and confidence related to fall prevention, PAM, M-FaB, and fall rates) and examined baseline differences between the intervention and control arms using standard t tests and chi-squared tests. The reliability of study instruments was determined by Cronbach's alpha.

To examine the potential impact of MI over time on study outcome measures, a linear mixed effects modeling approach with the nlme package (Pinheiro et al., 2018) for the R statistical computing environment was used (R Core Team, 2018). Because this is a feasibility study with a small sample size, efficacy findings must be interpreted with caution. However, linear mixed effect modeling is the most appropriate analysis approach because it allows researchers to examine changes over time within individuals, does not require all time points to have the same sample size during the longitudinal follow-up period, and adequate sample size will be dependent on the number of control variables included in the models (Hedeker & Gibbons, 2006).

The authors expected that MI would have an immediate effect (Lundahl et al., 2013) to change fall prevention behaviors during participants' hospitalization. The authors were less certain about the long-term effect. Thus, separate analyses were conducted for short-term effects (baseline to 2 days) and long-term effects (2 days to 1 week, 1 month, and 3 months). The unadjusted model included fixed effects for the study arm (intervention vs. control), time, and the study arm × time interaction. PAM was only measured at baseline, 1 month, and 3 months, thus included only with the model that examined long-term effect. Control variables for the adjusted model were selected based on theoretically meaningful variables that could contribute to changes in the outcome. Control variables tested for the short-term model included: fall history at baseline, number of diagnoses, MOCA-B score, AHRQ High Fall Risk Medication Score, and age.

The long-term models tested these control variables as well as occurrence of falls since last follow up, change in health status, living situation, and mobility. These potential control variables were added individually to the models to test for significant relationships with each of the primary study outcome variables (confidence, FESI-S, M-FaB, PAM). When the relationship was significant using a p < 0.05 cut off for at least two outcome measures, the variables were included in the adjusted model. Thus, for the short-term adjusted model, 3-month fall history and number of diagnoses were included as control variables. For the adjusted long-term model, falls during follow up and out-patient status were included as control variables. Differences in fall rates between the study arms were assessed using incidence density to consider time of participation in the current study (Oleckno, 2015). The alpha was set at p < 0.05 for all analyses.

Results

Feasibility of the Study

The authors contacted 120 patients to seek their interest in the current study. Thirty-one patients declined and 18 deferred to participate. Thus, 71 participants (59%) were consented and randomized in the study. Subsequently, four participants withdrew from the study between the consent and baseline data collection due to changes in participation interest or condition (Figure 1). Furthermore, three participants were unexpectedly unable to receive any study intervention for reasons such as unplanned discharge, procedure, or presence of visitors. For study purposes, those participants were reassigned to the control arm.

Participant enrollment and attrition.* Among the 49 patients who were excluded, 31 declined to participate and 18 deferred the study for another time.** Unable to obtain data from three participants due to study withdraw and participant withdraw.*** Unable to obtain data from one participant due to participant withdraw.+ Three participants from the intervention arm were unable to receive the study intervention and were reassigned to the control arm. Inpatient status: Participants may be readmitted to a health facility after discharge.Incomplete reasons: D = deceased, U = unable to reach, W = withdrew, and M = unable to answer questions due to medical condition (e.g., medications, surgery, illness, pain).

Figure 1.

Participant enrollment and attrition.

* Among the 49 patients who were excluded, 31 declined to participate and 18 deferred the study for another time.

** Unable to obtain data from three participants due to study withdraw and participant withdraw.

*** Unable to obtain data from one participant due to participant withdraw.

+ Three participants from the intervention arm were unable to receive the study intervention and were reassigned to the control arm. Inpatient status: Participants may be readmitted to a health facility after discharge.

Incomplete reasons: D = deceased, U = unable to reach, W = withdrew, and M = unable to answer questions due to medical condition (e.g., medications, surgery, illness, pain).

Sixty-seven participants completed the baseline data collection: 31 participants completed baseline data collection and received the study intervention; 36 participants completed just the baseline data collection as control arm participants. For the longitudinal follow up, approximately 66% of participants participated in the 3-month data collection, control arm (n = 24) and intervention arm (n = 20). Of those who were unable to complete the study, six participants died during the study period, and an average of 11 participants per study time point could not be reached with three attempts via telephone. From the current study, FESI-S had an unstandardized Cronbach's alpha of 0.89, whereas PAM had Cronbach's alphas of 0.68, respectively. The Cronbach's alpha for the 18-item M-FaB (baseline during hospitalization) was 0.76 and the 29-item M-FaB ranged from 0.70 to 0.74 during 2-day, 1-week, 1-month, and 3-month data collection.

Baseline Measurements

Characteristics of the control and intervention arms were not significantly different at baseline (Table 2). Participants were inpatient males, 73.1 (SD = 6.4) years old, hospitalized for 4.3 (SD = 4) days, at high risk for falling (mean MFS score = 68.4 [SD = 15.4]), and cognitively oriented (mean MOCA-B score = 25.9 [SD = 2.9]). Approximately 58.3% of the control arm and 45% of the intervention arm had at least one fall within the 3 months prior to baseline data collection (p = 0.485).

Participants' Baseline Characteristics (N = 67)

Table 2:

Participants' Baseline Characteristics (N = 67)

At baseline, participants reported the level of importance to prevent falls as high at 9.12 (SD = 2.03; range = 0 to 10). Because the reported level of importance to prevent falling had small variability in participant responses and a ceiling effect, it was excluded from the longitudinal analyses. The reported level of confidence to prevent falling was lower at 7.23 (SD = 2.54; range = 0 to 10). Participants were moderately concerned about falling (FESI-S score = 17.8 [SD = 6.6; range = 7 to 28]). They reported moderate engagement in their own health with a mean PAM score of 64.3 (SD = 13.6) (1 to 100 possible score). The frequency of daily behaviors to prevent falling was moderate with M-FaB of 3.06 (SD = 0.46; range = 1 to 4).

MI Intervention

The MI intervention took approximately 21 minutes. Four participants had their intervention briefly interrupted due to hospital procedures and patient needs (i.e., pain, fatigue, toileting) but still responded to the two pre-structured questions for MI. No harm was reported from the intervention.

MI provider fidelity as measured by MITI coding indicated beginning proficiency in all aspects measured: (a) global scales (1 to 5) including cultivating change talk = 3, softening sustain talk = 3, partnership = 3.5, and empathy = 3.5; (b) skills including the ratio of reflections to questions = 1.6:1 (goal is 2 or 3:1), total MI adherent score (seeking collaboration + affirm + emphasizing autonomy) = 7.8 (moderately high); and (c) total MI non-adherent score (persuade + confront) = 3.5 (only used persuasion, goal is as few as possible).

Short-Term Results. The unadjusted model tested the study arm (intervention vs. control), time, and study arm × time interaction. Between baseline and 2-day follow up, there were significant differences in daily change between the intervention and control arms in both fear of falling (FESI-S, p = 0.010) and fall prevention behaviors (M-FaB, p = 0.031). When adjusted for the number of falls in the past 3 months and number of diagnoses, FESI-S showed a decrease in the intervention arm (−1.94) and a small increase in the control arm (0.37) (Table 3). In addition, M-FaB showed a small increase in the intervention arm (0.04) and a slightly larger increase in the control arm (0.19). These results were consistent with the unadjusted models. Daily change in confidence was not significantly different from baseline to 2-day follow up between the intervention and control arms. The incident density for falls was not evaluated due to the brevity of the follow-up period.

Difference Across Time by Group for Perceptions Related to Falls and Fall Prevention Behaviors (N = 67)

Table 3:

Difference Across Time by Group for Perceptions Related to Falls and Fall Prevention Behaviors (N = 67)

Long-Term Results. The unadjusted model tested the study arm (intervention vs. control), time, and study arm × time interaction. There was a significant sustained difference between the arm for FESI-S in the unadjusted model (p = 0.040), but this significance did not remain once adjusted for falls during follow up and outpatient status (p = 0.262). Furthermore, using the adjusted model, between 2 days post-intervention and 3 months, M-FaB showed a significant difference (p = 0.043) between the intervention and control arms for a sustained effect of MI. There was no practical change at 3 months in the intervention arm, but a small decrease was noted in the control arm (−0.08). Confidence and PAM did not show significant differences in change between the two arms in unadjusted or adjusted models. Further details of the mixed effects models are provided in Table 3. For fall rates, the authors found that the incidence density of falls during the 3-month follow-up period was 26.1% higher for the intervention arm. The incident density for falls for the control arm was 0.15 (95% confidence interval [CI] [0.057, 0.243]), and 0.18 for the intervention arm (95% CI [0.072, 0.307]). However, this difference was not significant (Table 4).

Fall Incidents During the Follow-Up Period

Table 4:

Fall Incidents During the Follow-Up Period

Discussion

The primary goal of the current pilot study was to evaluate the feasibility of MI in engaging hospitalized older adults in fall prevention. As per the authors' knowledge, the current study is one of the first to evaluate the application of brief MI in hospital settings for fall prevention. The recruitment rate was comparable to prior studies (Nyman & Victor, 2011) and met a success criterion of 50%; the authors also met their retention criteria. However, further improvement with retention is needed. The median attrition rate for institutionalized older adults (hospitalized/at nursing homes) was 10.4% (Nyman & Victor, 2011). The authors believe that acutely ill patients may have had limited capacity to participate in the current study due to their condition or having other priorities to address. Thus, recruiting more participants at baseline may be especially important for future studies. However, despite the acuity of the illness, 67 inpatients participated in the 8-month recruitment period and collectively voiced the importance of fall prevention.

The quality of MI met pre-determined criteria. The amount of training for the interventionist was comparable to other studies (Lundahl et al., 2013) and quality of MI was appropriate for the current feasibility study. The goal for future studies would be to provide repeated MI at advanced proficiency to offer additional value. The quality of MI is seldom reported in other health care MI studies (Lundahl et al., 2013). Thus, it will be important to continue evaluating the fidelity of MI interventions.

The semi-structured format of MI questions and prompts and menu of options for fall prevention helped facilitate patient-centered fall prevention discussions in limited time. These study tools are simple to incorporate into routine clinical practice when inquiring patients about their history of falls. Specifically, the Menu of Options for Fall Prevention (Table A) appeared to create an opportunity for clinicians to affirm what patients are already doing to be safe and to facilitate discussions related to barriers or adding other fall prevention strategies. Letting patients guide the conversation using the MI approach to bypass usual clinician-directed information giving shows promise in improving patient engagement in the fall prevention discussion (Apodaca et al., 2016).

The FESI-S and M-FaB tools were appropriate to evaluate perception and behavior changes related to preventing falls, and thus recommended for future studies. Clinicians can use questions from the FESI-S and M-FaB to explore emotional and behavioral barriers for fall prevention in their routine clinical practice. Older adults often undermine their perception of fall risks (McMahon et al., 2011), thus it is important to ask specific questions about their potential concerns related to falls in a way that is caring and normalizing. From the current study, there were positive impacts of MI on FESI-S and M-FaB scores. Even with a single instance of MI provided at beginning proficiency, participants who received the intervention reported less fear of falling during hospitalization as measured by the FESI-S, and less likely to abandon fall prevention behaviors at 3-month follow up as measured by the M-FaB. Higher M-FaB scores for the control arm may indicate that study questionnaires alone may have impacted fall prevention practices during hospitalization.

There was no significant impact on fall rates or PAM from the MI intervention. Participants continue to fall during the 3-month follow-up period. Because falls are caused by multiple factors, repeated MI at advanced proficiency level to address breadth and complexity of issues would be valuable.

Results provided early evidence that brief MI can have immediate and long-term effects in preventing falls. The authors expected brief MI interventions to have an immediate impact (Lundahl et al., 2013) for fall prevention behaviors during hospitalization. The current study results indicate the positive impact of MI to reduce inpatients' fear of falling.

On the contrary, the authors were uncertain about the impact of MI after discharge, especially as patients may transition through different living situations. However, the study found that MI had a lasting and positive impact on minimizing decline in fall prevention behaviors after discharge. In addition, MI showed some evidence to reduce fear related to falling post-discharge as well. The authors measured long-term effect because it would be ideal for inpatient clinical practices to have a long-term positive impact on patients' lives. Miller (2018) suggests MI is successful, especially with vulnerable populations, simply because a clinician is present and accepting of the individual without pushing a specific agenda or treatment, which perhaps suggests why even a beginner MI skill can have an impact on patient outcomes. A MI approach allows for clinicians to meet patients at their point of fall prevention engagement. Study materials such as the MI-Based Communication Guide for Fall Prevention or Menu of Options for Fall Prevention offer a great starting point to engage patients at their readiness to discuss fall prevention.

Limitations

There are several limitations to the current study. This pilot study examined a small group of participants from one VA facility. Thus, participants represented a population of cognitively oriented, mostly male in-patients who are at high risk for falling, which limits the generalizability of study findings. The authors used restrictive criteria to choose the control variables in the adjusted mixed effects models considering their sample size. Although the authors tested five potential variables in the short-term model and nine variables in the long-term model, they included only two control variables in each model to ensure an adequate participant-to-variable ratio for parameter estimation.

The exploratory nature of the current study, and the small sample size, caused the authors to compare actual treatment against non-treatment. The three participants who were unavailable to receive any intervention due to unexpected clinical situations were reassigned to the control arm. In the future, the current study should adhere to the gold standard of randomized controlled trials to use intention-to-treat principles (McCoy, 2017). The intention-to-treat approach will reduce the likelihood of exaggerating the estimate of results.

Some participants were not available via telephone for follow-up data collection. A total of 23 participants did not complete the 3-month data collection. The authors examined the reasons for the attrition: 12 participants could not be reached after three telephone call attempts; five participants withdrew; three participants were unable to participate due to medical conditions; and three participants died. These results speak to the vulnerability of older adult inpatients who are at high risk for falling. Thus, these study results are limited to those who had reliable access to telephones and may have overrepresented participants with better health.

Participants in the current study were Veterans who had an average of 10 diagnoses. For future studies, partnerships with patient advisors to gain insight into study retention to minimize attrition of non-health reasons will be valuable, and enrolling a larger number of participants in anticipation of changes in health status of older adults may be warranted.

There are potential biases due to the non-blinded study design. The PI delivered all MI interventions and supervised the analysis. The data collector was not blinded to the intervention. These factors may have caused bias in results toward finding the significant impact of the intervention. In future studies, interventions can be provided by multiple people with processes in place to establish reliability. Data collectors can be blinded to the intervention to reduce potential bias toward positive results.

The control group received routine hospital fall prevention. Use of an active comparator may have yielded valuable data, such as differential attrition and more accurate analysis of outcomes.

Despite limitations, the current study was novel in exploring the feasibility of MI provided to hospitalized patients for fall prevention.

Conclusion

Brief MI for fall prevention in an acute setting was feasible to deliver. The current pilot study provided insight into suitable study procedures for future trials and revealed challenges related to recruitment and retention for patients who experience acute illnesses. Study measures were appropriate, and the authors found the beneficial impact of MI during hospitalization and after discharge. MI intervention for fall prevention offers a practical way to improve communication with patients about fall prevention practices by meeting patients at their level of interest and engagement. The MI-Based Communication Guide for Fall Prevention and the Menu of Options for Fall Prevention offer a starting point to incorporate aspects of MI into routine clinical practice for fall prevention.

References

  • Agency for Healthcare Research and Quality. (2013a). Preventing falls in hospitals. Retrieved from http://www.ahrq.gov/professionals/systems/hospital/fallpxtoolkit/fallpxtk2.html
  • Agency for Healthcare Research and Quality. (2013b). Tool 3A: Master clinical pathway for inpatient falls. Retrieved from https://www.ahrq.gov/professionals/systems/hospital/fallpxtoolkit/fallpxtk-tool3a.html
  • Agency for Healthcare Research and Quality. (2013c). Tool 3I: Medication fall risk score and evaluation tools. Retrieved from https://www.ahrq.gov/professionals/systems/hospital/fallpxtoolkit/fallpxtk-tool3i.html
  • Agency for Healthcare Research and Quality. (2018). Patient engagement in safety AHRQ patient safety network. Retrieved from https://psnet.ahrq.gov/perspectives/perspective/243/patient-engagement-in-safety
  • Barker-Collo, S., Krishnamurthi, R., Witt, E., Feigin, V., Jones, A., McPherson, K. & Aroll, B. (2015). Improving adherence to secondary stroke prevention strategies through motivational interviewing: Randomized controlled trial. Stroke, 46, 3451–3458. doi:10.1161/STROKEAHA.115.011003 [CrossRef]
  • Cameron, I.D., Gillespie, L.D., Robertson, M.C., Murray, G.R., Hill, K.D., Cumming, R.G. & Kerse, N. (2010). Interventions for preventing falls in older people in care facilities and hospitals. Cochrane Database of Systematic Reviews, 12, CD005465. doi:10.1002/14651858.CD005465.pub3 [CrossRef]
  • Clemson, L., Bundy, A.C., Cumming, R.G., Kay, L. & Luckett, T. (2008). Validating the Falls Behavioural (FaB) scale for older people: A Rasch analysis. Disability and Rehabilitation, 30, 498–506. doi:10.1080/09638280701355546 [CrossRef]
  • Clemson, L., Cumming, R.G. & Heard, R. (2003). The development of an assessment to evaluate behavioral factors associated with falling. American Journal of Occupational Therapy, 57, 380–388. doi:10.5014/ajot.57.4.380 [CrossRef]
  • Copeland, L., McNamara, R., Kelson, M. & Simpson, S. (2015). Mechanisms of change within motivational interviewing in relation to health behaviors outcomes: A systematic review. Patient Education & Counseling, 98, 401–411. doi:10.1016/j.pec.2014.11.022 [CrossRef]
  • Hedeker, D. & Gibbons, R.D. (2006). Longitudinal data analysis. Hoboken, NJ: Wiley.
  • Hibbard, J.H. & Greene, J. (2013). What the evidence shows about patient activation: Better health outcomes and care experiences; fewer data on costs. Health Affairs, 32, 207–214. doi:10.1377/hlthaff.2012.1061 [CrossRef]
  • Hibbard, J.H., Mahoney, E.R., Stockard, J. & Tusler, M. (2005). Development and testing of a short form of the patient activation measure. Health Services Research, 40, 1918–1930. doi:10.1111/j.1475-6773.2005.00438.x [CrossRef]
  • Kempen, G.I., Yardley, L., van Haastregt, J.C., Zijlstra, G.A., Beyer, N., Hauer, K. & Todd, C. (2008). The Short FES-I: A shortened version of the Falls Efficacy Scale-International to assess fear of falling. Age & Ageing, 37, 45–50. doi:10.1093/ageing/afm157 [CrossRef]
  • Kiyoshi-Teo, H., Northup-Snyder, K., Cohen, D.J., Dieckmann, N., Stoyles, S., Winters-Stone, K.M. & Eckstrom, E. (2019). Older hospital inpatients' fall risk factors, perceptions, and daily activities to prevent falling. Geriatric Nursing, 40, 290–295. doi:10.1016/j.gerinurse.2018.11.005 [CrossRef]
  • Lundahl, B., Moleni, T., Burke, B. L., Butters, R., Tollefson, D., Butler, C. & Rollnick, S. (2013). Motivational interviewing in medical care settings: A systematic review and meta-analysis of randomized controlled trials. Patient Education & Counseling, 93, 157–168. doi:10.1016/j.pec.2013.07.012 [CrossRef]
  • Masterson Creber, R., Patey, M., Lee, C.S., Kuan, A., Jurgens, C. & Riegel, B. (2016). Motivational interviewing to improve self-care for patients with chronic heart failure: MITI-HF randomized controlled trial. Patient Education & Counseling, 99, 256–264. doi:10.1016/j.pec.2015.08.031 [CrossRef]
  • McCoy, C.E. (2017). Understanding the intention-to-treat principle in randomized controlled trials. Western Journal of Emergency Medicine, 18, 1075–1078. doi:10.5811/westjem.2017.8.35985 [CrossRef]
  • McMahon, S. & Fleury, J. (2012). External validity of physical activity interventions for community-dwelling older adults with fall risk: A quantitative systematic literature review. Journal of Advanced Nursing, 68, 2140–2154. doi:10.1111/j.1365-2648.2012.05974.x [CrossRef]
  • McMahon, S., Talley, K.M. & Wyman, J.F. (2011). Older people's perspectives on fall risk and fall prevention programs: A literature review. International Journal of Older People Nursing, 6, 289–298. doi:10.1111/j.1748-3743.2011.00299.x [CrossRef]
  • Miake-Lye, I.M., Hempel, S., Ganz, D.A. & Shekelle, P.G. (2013). Inpatient fall prevention programs as a patient safety strategy: A systematic review. Annals of Internal Medicine, 158, 390–396. doi:10.7326/0003-4819-158-5-201303051-00005 [CrossRef]
  • Miller, W. (2018). Motivational interviewing: A metamorphosis. Retrieved from https://vimeo.com/316279712
  • Miller, W. & Rollnick, S. (2012). Motivational interviewing: Helping people change (3rd ed.). New York, NY: The Guilford Press.
  • Moyers, T.B., Manuel, J.K. & Ernst, D. (2014). Motivational interviewing treatment integrity coding manual 4.1. Unpublished manual.
  • Nasreddine, Z. (2014). Montreal Cognitive Assessment-Basic (MoCA-B) administration and scoring instructions. Retrieved from http://www.mocatest.org/wp-content/uploads/2015/03/MoCA-B-Instructions-final.pdf
  • National Dataset of Nursing Quality Indicators. (2012). Nursing report card-NDNQI definitions-March 2012. Retrieved from https://www.qualityhealthnd.org/wp-content/uploads/Nursing-Metrics-2012.docx
  • Nyman, S.R. & Victor, C.R. (2011). Older people's recruitment, sustained participation, and adherence to falls prevention interventions in institutional settings: A supplement to the Cochrane systematic review. Age & Ageing, 40, 430–436. doi:10.1093/ageing/afr016 [CrossRef]
  • Oleckno, W. (2015). Epidemiology: Concepts and methods (1st ed.). Long Grove, IL: Waveland Press, Inc.
  • Pinheiro, J., Bates, D., DebRoy, S., Sarkar, D. & Heisterkamp, S.EISPACK authorsR-core. (2018). Fit and compare Gaussian linear and nonlinear mixed-effects models. R package version 3.1-137. Retrieved from https://www.rdocumentation.org/packages/nlme/versions/3.1-137
  • R Core Team. (2018). R: A language and environment for statistical computing. Retrieved from https://www.r-project.org
  • REDCap. (n.d.). REDCap. Retrieved from https://projectredcap.org
  • Reuben, D.B., Gazarian, P., Alexander, N., Araujo, K., Baker, D., Bean, J.F. & McMahon, S. (2017). The strategies to reduce injuries and develop confidence in elders intervention: Falls risk factor assessment and management, patient engagement, and nurse co-management. Journal of the American Geriatrics Society, 65, 2733–2739. doi:10.1111/jgs.15121 [CrossRef]
  • Riegel, B., Masterson Creber, R., Hill, J., Chittams, J. & Hoke, L. (2016). Effectiveness of motivational interviewing in decreasing hospital readmission in adults with heart failure and multimorbidity. Clinical Nursing Research, 25, 362–377. doi:10.1177/1054773815623252 [CrossRef]
  • Tzeng, H.M. & Yin, C.Y. (2015). Patient engagement in hospital fall prevention. Nursing Economic$, 33, 326–334.

Motivational Interviewing (MI)–Based Communication Guide for Fall Prevention: MI Approaches to Address Fall Prevention and Rationales

Approaches and ExamplesRationales
Ask for permission to have a conversation about fall prevention.a

Empower patients.

Show respect.

Ask open-ended question such as:

What does fall prevention mean to you?a

What do you currently do for fall prevention?

How has your past fall impacted you?

Let patients share their thoughts and perspectives about fall prevention.

Explore the focus of the conversation.

Identify stages of change.b

Ask for permission to share information.

I have this information brochure that I think would be helpful for you, may I share this with you?a

Show respect.

Let patients choose the topic to discuss about fall prevention.

What are one or two things that may be useful or meaningful to you for us to discuss about fall prevention from this information brochure?a

Empower patients.

Show respect.

Let patients be in control.

Offer advice in a non-persuasive and non-directive way.

For someone with similar issues, I may recommend XXX.

Empower patients.

Show respect.

Let patients be in control.

Affirm patients' strengths.

I am so glad to hear that you are already being so careful not to fall.

Empower patients.

Identify strengths.

Help establish trust.

Build confidence.

Reflect on patients' stories.

It seems like you don't want to look old and weak by using the walker.

Indicate that patients' stories are valued and heard.

Evoke patients to think about how they might want to change to be safer.

What would it look like if you were to start on XX today?

How motivated are you to XX?

Prompt behavior change.

If patients are disinterested or hesitant, ask:

You seem XX, can you tell me why?

Address behavior change in relation to patients' readiness to change.

Participants' Baseline Characteristics (N = 67)

Characteristicn (%) or Mean ± SDp Value
Overall (N = 67)Intervention Group (n = 31)Control Group (n = 36)
Demographics
  Male65 (97)29 (93.5)36 (100)0.408
  Age (years)73.13 ± 6.3572.83 ± 673.43 ± 6.780.704
  Time since admission (days)4.34 ± 3.963.90 ± 2.794.72 ± 4.760.386
  Admission due to a fall8 (11.9)5 (16.1)3 (8.3)0.088
  Number of diagnoses10.37 ± 4.8311.10 ± 4.679.75 ± 4.940.256
  Dependent mobility (requiring at least 1-person assistance)29 (43)16 (52)13 (36)0.401
  Morse Fall Scale68.36 ± 15.4168.06 ± 15.7968.61 ± 15.280.886
  Montreal Cognitive Assessment Basic Score25.85 ± 2.8926.23 ± 2.57825.03 ± 3.050.087
Fall prevention education during admission
  Documented at admission56 (83.6)29 (94)27 (75)0.088
  Documented for the most recent 24 hours of hospital stay61 (91)28 (90)33 (92)0.888
Medications
  Number of routine medications15.24 ± 6.5214.16 ± 6.7616.17 ± 6.240.214
  Patients with AHRQ fall risk medicationsa65 (97)30 (97)35 (97)1.0
  AHRQ High Fall Risk Medication Scoreb9.22 ± 4.438.35 ± 4.229.97 ± 4.520.135
    Score ≥6 (higher risk for falling)65 (97)30 (97)35 (97)1.0
Fall historyc
  Previous 3 months
    ≥2 falls17 (25.3)8 (26)9 (25)0.485
    1 fall18 (26.9)6 (19)12 (33.3)0.485
    Injurious falld23 (34.3)11 (35)12 (33.3)0.981
  Previous 1 year (inclusive of 3 months)
    ≥2 falls30 (44.8)12 (34)18 (50)0.164
    1 fall15 (22.4)5 (16)10 (27.8)0.164
    Injurious falld30 (44.8)13 (42)17 (47.2)1.0
Fall-related perception measures (baseline)
  Importance9.12 ± 2.039.28 ± 1.69 ± 2.340.576
  Confidence7.23 ± 2.547.79 ± 2.366.76 ± 2.630.113
  Fear of falling (Falls Efficacy Scale International-Short)17.81 ± 6.6218 ± 6.9017.65 ± 6.480.114
Patient engagement measure (baseline)
  Patient Activation Measure64.33 ± 13.6464.69 ± 12.2664.01 ± 14.930.848
Fall prevention behavior measure (baseline)
  Modified Falls Behavioural Scale3.06 ± 0.463.11 ± 0.392.92 ± 0.500.449

Difference Across Time by Group for Perceptions Related to Falls and Fall Prevention Behaviors (N = 67)

OutcomeUnadjustedAdjusteda
InterventionControlp ValuebInterventionControlp Valueb
Confidence
  Baselinec7.7786.7658.1907.374
  Daily change from 0 to 2 days0.1510.5060.4710.1430.4830.5
  Monthly change from 2 days to 3 months−0.0540.2320.263−0.0280.3830.119
FESI-S
  Baselinec1818.06617.01116.330
  Daily change from 0 to 2 days−20.3550.009−1.9380.3710.010
  Monthly change from 2 days to 3 months0.039−0.0460.040−0.856−0.2360.262
M-FaB
  Baselinec3.1033.0183.0102.915
  Daily change from 0 to 2 days0.0390.1800.0400.0400.1850.031
  Monthly change from 2 days to 3 months−0.023−0.1070.0410.001−0.0830.043
PAMd, e
  Baselinec64.29764.36981.05983.208
  Monthly change from 0 to 3 months2.4580.1850.1420.786−2.0830.324

Fall Incidents During the Follow-Up Period

Fall IncidentIntervention GroupControl Group
Between baseline and 2-day follow up11
Between 2-day and 1-week follow up21
Between 1-week and 1-month follow up42
Between 1-month and 3-month follow up68
Incidence rates (month)0.20290.2098
Incident density0.180.15
  Incident density ratio1.261
    95% confidence interval[0.072, 0.307][0.057, 0.243]

Menu of Options for Fall Prevention

Modified Fall Behavioral Scale: Inpatient and Post-Discharge Questions

Questions
Inpatient questions
1,I talk with someone I know about things I do that might help prevent a fall
2,I use call-light to get assistance anytime I need to stand
3,I'm often not in a hurry to go use the toilet*
4,I bend over to reach something only if I have a firm handhold
5,I made changes to make the light better
6,I check to see what things are attached to myself before getting up
7,I get help when I things are beyond easy reach
8,I notice spills on the floor
9,When I stand up I pause to get my balance
10I do things at slower pace
11,I hold onto things to stabilize myself
12,I use a walking stick or walking aid when needed
13,I do not hurry to answer the phone*
14,I do not hurry when I do things IN GENERAL*
15,When I am feeling ill I take special care of how I get up from a BED and move around
16,When I am feeling unwell I take particular care doing everyday things
17,I do not turn around quickly*
18,I use a light if I get up during the night
Post-discharge questions
1,When I buy shoes I check the soles to see if they are slippery
2,I get help when I need to change a light bulb
3,To reach something high I do not use the nearest chair, or whatever furniture is handy, to climb on*
4,When I am getting down from a ladder or step stool I think about the bottom rung/step
5,When I walk outdoors I look ahead for potential hazards
6,I avoid ramps and other slopes
7,I go out on windy days
8,When I go outdoors I think about how to move around carefully
9,I cross at traffic lights or pedestrian crossings whenever possible
10I hold onto a handrail when I climb stairs
11,I avoid walking about in crowded places
Authors

Dr. Kiyoshi-Teo is Assistant Professor, Dr. Northup-Snyder is Clinical Assistant Professor, Dr. Dieckmann is Associate Professor, Ms. Stoyles is Research Associate, and Dr. Winters-Stone is Professor, School of Nursing, Dr. Cohen is Professor, Department of Family Medicine, and Dr. Eckstrom is Professor, School of Medicine, and Chief of Geriatrics, Division of General Internal Medicine and Geriatrics, Oregon Health and Science University, Portland, Oregon.

The authors have disclosed no potential conflicts of interest, financial or otherwise. The current project was made possible through the Oregon Health and Science University, School of Nursing Hartford Award for Research and Practice.

Address correspondence to Hiroko Kiyoshi-Teo, PhD, RN, Assistant Professor, School of Nursing, Oregon Health and Science University, 6S, 3455 SW US Veteran's Hospital Road, Portland, OR 97239; e-mail: kiyoshi@ohsu.edu.

Received: January 22, 2019
Accepted: July 09, 2019

10.3928/00989134-20190813-03

Sign up to receive

Journal E-contents