Fall risk prediction ability in rehabilitation professionals: structural equation modeling using time pressure test data for Kiken-Yochi Training

Study design

This was a cross-sectional study involving rehabilitation professionals and students.

Participants and data collection

The participants were rehabilitation professionals working in 12 medical facilities and three nursing homes; the study included first- or fourth-year rehabilitation students (physical and occupational therapy students) from two schools (four-year universities). Participants were informed of the purpose and methods of the study and their written consent was obtained. In total, 402 rehabilitation professionals and students participated in the study. Inclusion criteria for rehabilitation professionals were therapists who were licensed physical or occupational therapists. No criteria based on years of experience were established. Second and third-year students were not recruited because the progression of basic and specialized subjects in the education curriculum differs across schools. Since the student participants in this study were affiliated with four-year universities, they were advanced learners in the rehabilitation professional field (Japanese Association of Occupational Therapists, 2020a). Data were collected through face-to-face and group surveys at the participants’ workplaces and training schools. They were conducted in a quiet place to ensure that participants focused on the survey. The surveys were administered by occupational therapists who understood the TP-KYT procedures. This cross-sectional survey was conducted from January 2015 to April 2023. Of the 402 participants, no participants were excluded, so all participants were included in the analysis (Table 1). Of the 402 participants, a total of 176 (102 rehabilitation professionals and 74 rehabilitation students) had also participated in a previous study on a similar topic (Arihisa et al., 2019).

Ethical procedure

This study was approved by Hiroshima University’s Ethics Committee for Epidemiology (E-2838). Participants were informed of the study’s purpose, how the data would be processed, and that they were free to abstain or withdraw from participation without any penalty. Their consent was obtained in writing.

Measures

The study used a participants’ characteristics questionnaire and the TP-KYT.

Participants’ characteristics

Age, gender, job role (physical therapist, occupational therapist, physical therapy student, or occupational therapy student), years of experience, and year in school were collected through a questionnaire.

TP-KYT

The TP-KYT is an evaluation tool that can quantify the FRPA of rehabilitation professionals and students (Arihisa et al., 2019). To enhance the performance of life functions, it includes five scenes with a high frequency of medical accidents involving falls during rehabilitation practice that differ in terms of diseases, symptoms, and environments. Scene 1: Patient is sitting in bed with low arousal after cancer surgery; Scene 2: Patient with left hemiplegia is trying to shift from a wheelchair to a bed; Scene 3: Patient with cerebellar ataxia is trying to sit on a toilet seat in a bathroom; Scene 4: Patient with right hemiplegia is trying to get into the bathtub; Scene 5: Patient with lower limb muscle weakness is doing housework. In accordance with the TP-KYT implementation procedure (Arihisa et al., 2019), participants were given 15 s to read the instructions, and were then required to identify the potential fall risks in the illustrations within 10 s. Participants then wrote down the reasons for identifying specific areas as fall risks. This process was repeated for all five scenes.

The scoring method used weighted scores only when the description corresponded to the 24 scoring items (e.g., 40 points were added if a description corresponded to an item with a score of 40, and 0 points if no description corresponded to an item with a score of 40; Arihisa et al., 2019). The scoring range is from 0 to 425 points. Higher scores indicate higher FRPA. The 24 scoring items and weighting scores were quantified based on the viewpoints of rehabilitation professionals with more than 5 years of clinical experience, thus confirming their content validity (Table 2). In this study, after identifying the FRPA factors, corrected scores were calculated for each factor and used in subsequent analyses. Thus, the total score of the 24 scoring items was not used.

Analysis procedure

First, to identify FRPA factors included in the TP-KYT, factors were extracted from the contents of the 24 scoring items constituting the TP-KYT, as presented by Arihisa et al. (2019) through a qualitative analysis (Table 2). The procedure involved summarizing the contents of each scoring item and categorizing them into groups, which were then classified according to their similarity and coded. The codes were then aggregated into categories (Creswell & Plano Clark, 2010). In this study, the FRPA factors that are measured by the TP-KYT were aggregated as categories and are henceforth referred to as factors. The first and second analysts (both occupational therapists) conducted inductive analyses separately. Disagreements were discussed with the four collaborators to reach a consensus. The six occupational therapists involved in the analysis have been licensed for more than five years and are all experts in their field (Unsworth, 2001). This ensured the validity of the analysis (Creswell & Plano Clark, 2010).

Next, to verify the model of FRPA factors based on whether participants did or did not have rehabilitation professional licenses, structural equation modeling (SEM) was conducted by setting FRPA as a latent variable and the correction scores of factors as observed variables, then drawing a path from job role to FRPA. In the analysis, correction scores were calculated for each of the aggregated factors, and the observed variables for the job role (rehabilitation professionals = 1, rehabilitation students = 0) were set and used. The correction score was calculated by dividing the factors score (acquisition score) by the factor score (full score), which is the sum of the weighted scores of the scoring items (0 to 1). This data processing was conducted to correct the full scores of the different factors, considering that the scoring items are binary in their weighting scores (Little et al., 2013).

The SEM was analyzed using the generalized least squares method. The goodness of fit was evaluated using root mean square error of approximation (RMSEA), goodness-of-fit index (GFI), and adjusted GFI (AGFI), using the following criteria: RMSEA<0.05, GFI>0.95, and AGFI>0.90 (Schermelleh-Engel, Moosbrugger & Müller, 2003). Statistical significance was defined as p < 0.05. Analyses were performed using IBM SPSS Statistics 28 and Amos 28 (IBM Corp., Armonk, NY, USA).

Participants’ characteristics

The characteristics of the 402 participants are shown in Table 1. The participants consisted of 184 rehabilitation professionals, 101 of whom were men (54.9%), and 218 rehabilitation students, 109 of whom were men (50.0%), with mean ages of 31.4 (SD =6.5) and 19.9 (SD =1.9), respectively. The mean years of experience of the rehabilitation professionals was 8.0 (SD =5.8). Of the rehabilitation students, 112 (51.4%) were first-year students.

FRPA factors measured by the TP-KYT

By extracting 14 types of codes from 24 items, the following three FRPA factors were aggregated: (1) Patient ability (PA): risk of falls caused by patients’ physical and mental functions during rehabilitation practice; (2) Physical environment (PE): risk of falls caused by the physical tools surrounding the patient during rehabilitation practice; and (3) human environment (HE): risk of falls caused by someone who is not prepared to help if a patient falls during rehabilitation practice or by having no one near the patient. Table 2 lists the codes that comprise the factors and their concepts.

FRPA factors based on whether participants did or did not have rehabilitation professional licenses

The model analysis results are shown in Fig. 1 (Table 2 shows the full scores for PA, PE, and HE; Table 3 shows the corrected scores for PA, PE, and HE). The standardized coefficient from job role to FRPA was 0.85 (p < 0.001); the standardized coefficients from FRPA to the three factors were 0.58 for PA, 0.64 for PE, and 0.46 for HE (p < 0.001). The model demonstrated a good fit, with RMSEA<0.001, GFI =0.998, and AGFI =0.990.

FRPA factors measured by the TP-KYT

PA was related to the risk of falls due to patients’ physical and mental functions. In supporting patients, rehabilitation professionals work on basic, applied, and socially adaptive abilities to expand life functions, including physical and mental functions (Ministry of Health, Labour and Welfare, 2008). PA was extracted as a factor because rehabilitation practices need to predict the risk of falls based on PA assessment and provide quality support programs according to patient characteristics. The PE was related to the risk of falls due to the physical tools surrounding patients. As environmental adjustments by rehabilitation professionals are highly effective in preventing falls (Stevens & Lee, 2018), PE was identified as a factor and had the highest standardized coefficient. HE was related to the risk of falls due to medical staff not being prepared to help when the patient falls or having no one near the patient. For example, in a survey of factors contributing to the incidence of falls in hospitalized patients, it was reported that 18.4% of falls occurred because medical staff did not accurately assess the patient’s fall risk while providing medical care (Liu, Zhu & Song, 2021). Furthermore, it has been shown that many medical accidents occur in relation to human factors, such as a lack of confirmation or inattention by medical staff (Reason, 2000; Japan Council for Quality Health Care, 2012). In this study, the codes for HE included “The therapist’s position is far away” and “The therapist is not observing the patient” (Table 2). In other words, therapist judgment and behavior may be related to the occurrence of falls, and measures focusing on human factors of rehabilitation professionals are required. Therefore, clinical practice considering HE, including rehabilitation professionals, was extracted as a factor because it leads to fall prevention. Of these three factors, while PA was similar to the assessment of gait, balance, and mobility, PE was similar to the assessment of environmental hazards, which are common components in clinical practice guidelines for fall risk assessment (Williams-Roberts et al., 2021). However, HE was not included in the common components of clinical practice guidelines for fall risk assessment and was a distinguishing factor of the TP-KYT. Since falls are more likely to happen when the various factors occur together (Campbell & Robertson, 2006), rehabilitation practices that focus on the interaction of the three factors, HE, PA, and PE, may contribute to fall prevention.

The Person–Environment–Occupation (PEO) model explains human occupational performance in terms of three aspects: person, which includes the individual’s mental and physical functions and values; environment, which includes the human, physical, and social environment; and occupation, which includes tasks and engagement in life activities (Law et al., 1996). During rehabilitation practice, when viewed as task characteristics that enhance patients’ performance of life functions, the TP-KYT’s five scenes correspond to occupations. More specifically, the three factors (PA, PE, and HE) aggregated in this study are similar to the person and environment in the PEO model. Therefore, by focusing on the three factors that constitute the TP-KYT’s FRPA, it may be possible to contribute to the construction of a support system that enhances the performance of life functions.

FRPA factors based on whether participants did or did not have rehabilitation professional licenses

The results of the model analysis showed a good fit (Fig. 1). The student participants were in the process of receiving advanced education in rehabilitation. However, since the three FRPA factors were higher for professionals than for rehabilitation students, receiving professional training on fall prevention as students is necessary. In clinical settings, clinical reasoning is important for predicting and preventing behaviors that lead to patient falls (The Japanese Society for Fall Prevention, 2023). Clinical reasoning is classified into three domains: cognitive, psychomotor, and affective (Huhn et al., 2019). The cognitive domain is related to knowledge; the psychomotor domain to examination, measurement, and treatment skills, such as observation and palpation skills; and the affective domain to professionals’ attitudes and emotional state (Huhn et al., 2019; da Silva Araujo et al., 2022). As the TP-KYT requires participants to read descriptions of patients’ medical conditions and view illustrations that demonstrate the specific scenes to predict fall risks (Arihisa et al., 2019), the ability to integrate knowledge and observational skills is required. In addition, the TP-KYT score is quantified based on the risk prediction of experts with more than five years of clinical experience (Arihisa et al., 2019), which expresses the depth of experience in rehabilitation practice (Arihisa et al., 2019; Matsushita et al., 2023). In other words, we believe that job role and FRPA were positively related because differences in knowledge and observation of the three FRPA factors affected the depth of experience. Tests using rating scales have shown that clinical reasoning ability is higher in physical therapists than in students (Huhn et al., 2011). Furthermore, the results of a literature review investigating the differences between novice and expert occupational therapists indicated that clinical reasoning is an ability that develops continuously, with progression from students to professionals (Unsworth & Baker, 2016). The positive association between job role and FRPA in the present study was a result shared by previous clinical reasoning studies.

Among Japanese rehabilitation professionals, about one-quarter have less than five years of clinical experience (Japanese Association of Occupational Therapists, 2020b; Japanese Physical Therapy Association, 2023). Falls that occur during rehabilitation practice are more frequent among therapists who have less than five years of experience than among those with more experience (Maeda et al., 2018). Providing safe, high-quality rehabilitation that avoids falls is one of the quality indicators identified by the World Federation of Occupational Therapists  (2020). Additionally, focusing on professional training to increase FRPA during the student years is an important patient safety measure. Continuous staff training after certification is a factor that contributes to reducing the risk of falls (Tricco et al., 2019). The perspective of rehabilitation professional training has the potential to effectively enhance FRPA from the student stage by including the three factors identified in the content in this study. Specifically, these findings can be used in lectures, interactive learning activities, simulation education, and so forth (Shaw, Kiegaldie & Farlie, 2020) to develop knowledge and observation skills that form the foundation for preventing falls in rehabilitation practice settings. Moreover, the TP-KYT can continuously measure the effects of professional training to increase FRPA, from rehabilitation students to professionals. Furthermore, the fact that HE was extracted as a characteristic factor emphasizes the importance of professional training focusing on this factor, or human factors, during rehabilitation practice. The three factors may be applied as components for rehabilitation professionals to develop assessment methods for disease-specific risk scenes for both their target patients and the diverse situations in their patients’ lives. However, the factors “because the position of the bath board is far away, there is a risk of falling as the patient moves” and “the position of the shower chair is set to sit down from the paralyzed side” related to PE may need to be modified for applicability in countries where the lifestyle is different from that in Japan.