Effects of a Rehabilitation Program Combined with Pain Management That Targets Pain Perception and Activity Avoidance in Older Patients with Acute Vertebral Compression Fracture: a Randomised Controlled Trial

This study aimed to investigate the effect of a rehabilitation program combined with pain management targeting pain perception and activity avoidance on multifaceted outcomes in older patients with acute vertebral compression fractures (VCFs). We randomised 65 older adults with acute VCFs to either an intervention group (n = 32), involving usual rehabilitation combined with pain management that targeted pain perception and activity avoidance, or a control group (n = 33), involving only usual rehabilitation. The usual rehabilitation was initiated immediately after admission. All patients were treated conservatively. Pain management aimed to improve the patients' daily behaviour by increasing their daily activities despite pain, rather than by focusing on eliminating the pain. Pain intensity and psychological statuses such as depression, pain catastrophising, and physical activity levels were assessed on admission. Two weeks postadmission and at discharge, physical performance measures were assessed along with the above-given measurements. A significant main effect of the group was observed for the intensity of lower back pain, favouring the intervention group (F = 5.135, p = 0.027). At discharge, it was significantly better in the intervention group than in the control group (p = 0.011). A time-by-group interaction emerged for magnification of the pain catastrophising scale (p = 0.012), physical activity levels (p < 0.001), and six-minute walking distance (p = 0.006), all favouring the intervention group. Rehabilitation programs combined with pain management targeting pain perception and activity avoidance could be an effective conservative treatment for older patients with acute VCFs.


Introduction
Fragility vertebral compression fractures (VCFs) due to osteoporosis have become an important health problem, owing to their increasing incidence in the aging population. A recent survey conducted in Japan reported an annual incidence of clinical VCFs diagnosed not only with imaging but also with clinical symptoms in people aged ≥65 years to be 15.58 per 1000 individuals [1].
Surgery is usually not required for all acute osteoporotic VCFs; hence, conservative treatment, comprising pain relief and rehabilitation is recommended [2]. Owing to a higher incidence of chronic pain due to VCFs and other health problems, efective pain management strategies for the acute phase are needed. Previous studies have indicated that 76% of patients with acute VCFs and treated conservatively experience severe pain even after one year, and 40% of patients continue to have disabling pain [3,4]. Further studies have suggested that patients with VCFs are disabled and have lower activities of daily living (ADL) and quality of life (QOL) [3,5]. Research fndings from our previous study suggested that patients with persistent severe acute lower back pain, resulting from VCFs tend to be depressed and experience pain catastrophising, and persistent severe acute lower back pain, which negatively afects endurance and muscle strength but not ADL [6]. In addition, we previously indicated that lower physical activity during the acute phase leads to higher pain four weeks after treatment [7]. Terefore, pain management strategies for VCFs may require tailored approaches to address these factors.
Regarding pain management for VCFs during the acute phase, early mobilisation should be encouraged, as soon as it can be tolerated [8]. In addition, advice to remain active is the standard management for acute lower back pain [9]. However, owing to the low levels of physical activity observed in hospitalised older patients with VCFs [7], pain management strategies to increase physical activity are required. Furthermore, psychological techniques that alter dysfunctional ways of thinking, modify beliefs and attitudes and increase a person's control over pain and how they interpret pain are important for managing pain in older people [10]. Our previous randomised controlled trial demonstrated that exercise combined with pain management aimed at improving patients' daily behaviour by increasing their daily activities despite pain, rather than focusing on eliminating pain. Tis strategy efectively reduces pain intensity and improves psychological status and physical activity levels in older people with chronic musculoskeletal pain [11]. However, no study has demonstrated whether this pain management program, which targets pain perception and activity avoidance is efective in hospitalised older patients with acute VCFs.
Tis study aimed to investigate the efects of a rehabilitation program combined with pain management, targeting pain perception and activity avoidance on multifaceted outcomes in older patients with acute VCFs.

Participants.
We enrolled patients with acute VCFs aged >65 years who were admitted to the emergency unit of Nagasaki Memorial Hospital, Nagasaki, Japan. Patients who were eligible for enrolment in the study were diagnosed with acute vertebral fractures based on acute pain in the back and lower back regions, a deformed vertebral body on radiography, and abnormal intensity within the vertebral bodies on magnetic resonance imaging. All fractures were confrmed to have originated from low-energy trauma. Patients were excluded if they were diagnosed with an unstable fracture, which afected all three columns (e.g., chance and burst fractures), were transferred from other hospitals, were unable to walk independently with or without a walking aid before hospital admission, unable to complete the questionnaire due to cognitive impairment, had a Mini-Mental State Examination (MMSE) score from 0 to 15, were defned as having moderately severe or severe cognitive impairment [12,13], had complications that inhibited rehabilitation (for example, were haemodynamically unstable), had any other fracture type, were scheduled to be discharged within four weeks, or did not agree with the study protocol. Te study protocol was approved by the Research Ethics Committee of the Graduate School of Biomedical Sciences of Nagasaki University (Approval no. 18110805). All the participants were informed about the study procedures and outcome measures and provided signed informed consent. Tis study was registered with the UMIN Clinical Trials Registry (UMIN000035426).

Design and Randomization.
Tis randomised controlled trial was conducted between January 2019 and May 2021. Te participants who met the inclusion criteria were randomised into two groups (1 : 1) using a computergenerated randomisation list. Block randomisation with a fxed block size of two was used to ensure similar sample sizes across the conditions. Te groups included an intervention group, involving a usual rehabilitation program combined with pain management that targeted pain perception and activity avoidance, and a control group, involving only the usual rehabilitation program. An independent investigator (J. Y.) performed the randomisation after baseline assessment. A physical therapist in charge of each patient assessed the participants and implemented their usual rehabilitation and/or pain management programs.

Usual Rehabilitation Program.
Te usual rehabilitation program for patients in both groups was initiated immediately after admission (Table 1). All the patients were treated conservatively, as described in our previous study [6]. One week after admission, the patients were instructed to use a reclining bed and were permitted to move on the bed and walk around the room without a brace, but with the assistance of a healthcare worker to access the toilet. In the rehabilitation sessions, methods to reduce spinal pain during movement, such as turning over on the bed, were taught, and nonweight-bearing exercises of the upper and lower extremities were performed on the bed without bracing. After mild rehabilitation at the bedside, patients started mobilisation and exercises under bracing with a rigid or soft orthosis, which was adjusted according to the patient's injury level or with no orthosis. Rehabilitation after mobilisation included gait exercise, muscle strength training, balance exercise, and ADL exercises to prevent falls and allow a return home. Te patients were instructed to wear the orthosis for 8-12 weeks. Furthermore, to manage acute low back pain, international guidelines recommend that general practitioners provide education and reassurance Table  1: Usual rehabilitation program and pain management program.

Usual rehabilitation program (for the intervention and control groups)
Mild rehabilitation at the bedside (frst week after admission) (i) Bed mobility exercise and teaching the way to reduce spinal pain during movement (ii) Nonweight-bearing exercises of the upper and lower extremities were performed on the bed without bracing Rehabilitation after mobilisation (second week after admission) (i) After mild rehabilitation at the bedside, gait exercise, muscle strength training, balance exercise, and ADL exercise to prevent falls and allow a return home were performed Education (from frst to second week) (i) About the mechanism of VCFs and acute pain due to VCFs, conservative treatment and management of VCFs, and multifaceted pain Pain management program (for the intervention group) Aim (i) To improve patients' daily behavior by increasing their daily activities despite pain, rather than focusing on eliminating pain Goal setting (i) Rehabilitation goals were set based on shared decision-making, involving patients and physical therapists Physical activity approaches (i) Pedometers (Yamax digiwalker SW-200; Yamasa Tokei Keiki Co., Ltd., Tokyo, Japan) and diaries were used (ii) Patients were asked to wear the pedometer when awake and record their daily pain intensity, step counts, and behaviour in a diary (iii) Step targets were increased by 5% every week if there was no worsening pain or no increase in fatigue (iv) Step targets were not increased if patients complained of worsening pain or increased fatigue Pain Research and Management [14]. Tus, participants in both groups received education on the mechanism of VCFs and acute pain due to VCFs, conservative treatment and management of VCFs, and multifaceted pain.

Pain Management Program.
Patients in the intervention group also received a pain management program that targeted pain perception and activity avoidance one week after the start of rehabilitation (Table 1). Te pain management program aimed to improve patients' daily behaviour by increasing their daily activities despite the pain, rather than focusing on eliminating pain. First, patients in the intervention group set goals for rehabilitation. Sharing decisions about rehabilitation goals with participants can have a positive impact on the patient's health and mental well-being [15,16]; thus, rehabilitation goals were set based on shared decision-making, involving patients and physical therapists (H. K. and K. G.). Second, patients in the intervention group also underwent supervised physical activity using pedometers (Yamax Digiwalker SW-200; Yamasa Tokei Keiki Co., Ltd., Tokyo, Japan) and diaries. Te patients were asked to wear the pedometer when awake and record their daily pain intensity, step counts, and behaviour in a diary [11]. Pain intensity was recorded using an 11-point numerical rating scale (NRS), ranging from 0 ("no pain") to 10 ("worst imaginable pain"). Furthermore, our previous study revealed that the experience of successfully reaching step-count goals was useful in changing older participants' pain awareness [11]. Terefore, step targets were determined based on the frst week and were calculated to increase the patients' average daily step counts by 5% in the second week, in accordance with previous studies [17]. Subsequently, step targets were calculated repeatedly and were increased by 5% every week until discharge if there was no worsening of pain or increase in fatigue. However, if the patients complained of worsening pain or increased fatigue, the step targets did not increase. Tus, the determination of step targets was based on shared decision-making, involving patients and physical therapists.

Baseline
Characteristics. Data collected to characterise the patients at baseline included demographics, instrumental ADL before admission, and cognitive function. Patient demographics, including age, sex, height, body weight, body mass index (BMI), comorbidity, the locations and number of acute and previous fractures, and pain medication were obtained from the patients' medical records. Instrumental ADL before admission was assessed using the Tokyo Metropolitan Institute of Gerontology (TMIG) Index of Competence [18], and cognitive function was assessed using the Japanese version of MMSE [13]. Comorbidity was assessed using the Charlson Comorbidity Index (CCI), which includes 19 disease groups, where a higher score indicates a higher mortality risk [19]. Te number of fractures was categorized into one, two, or more than two acute fractures, and zero, one, two, or more than two previous fractures, which is mild or more deformity according to the Genant classifcation [20].

2.5.
Outcomes. Before beginning the study, physical therapists received training from one of the authors (H. K.), regarding assessment protocols. Te primary outcome was lower back pain intensity during movements, such as rolling, standing up, and walking, which was assessed using an 11-point NRS [21]. Secondary outcomes were psychological status, physical activity levels, ADL, QOL, the ratio of the height of each border of the collapsed body to the posterior border of a normal upper vertebral body, and physical performance measures. Psychological status, including depressive symptoms, pain catastrophising, and fear of movement or (re) injury was evaluated. Depressive symptoms were evaluated using the short Japanese version of the 15-item Geriatric Depression Scale (GDS-15) [22]. Pain catastrophising was evaluated using the Pain Catastrophising Scale (PCS) [23], which includes 13 items and three categories: rumination (fve items), helplessness (fve items), and magnifcation (three items). Fear of movement/(re) injury was assessed using the Tampa Scale for Kinesiophobia-11 (TSK-11) [24].
Physical activity levels were monitored using a uniaxial motion counter (Lifecorder GS ® , Suzuken, Japan) placed at the right or left anterior superior iliac portion during admission (baseline). Tis motion counter can measure and store real-time estimates of the frequency, intensity, and duration of total physical activity [25]. Te participants were asked to put on the motion counter whenever they were in the hospital, including during the therapy sessions, as well as in their bedroom. Te motion counter was removed when the participant bathed or showered because it was not waterproof. Data from the motion counters were downloaded to the computer weekly to analyse the activity data (Y. K. and Y. N.). Using the motion counter, daily step counts and physical activity intensities were classifed according to activity levels 1-9, and these activity levels were assigned to metabolic equivalents of 1.8-8.3. Te average times of 1-9 (1.8-8.3) activity levels per week were recorded, according to our previous study [7]. We used the Functional Independence Measure (FIM) to determine the generic ability to perform ADL after admission. Tis performance-based disability tool assesses the disability level when performing basic ADL [26].
Te EuroQOL 5-dimension3-level (EQ-5D 3L) was used to measure health-related QOL. Te EQ-5D 3L descriptive system measures health-related QOL in fve dimensions: mobility, self-care, usual activity, pain/discomfort, and anxiety/depression within three levels (i.e., no problems, moderate problems, and severe problems) and reports the result as a single fve-digit number (from 11111, representing full health, to 33333) [27]. Te single fve-digit number can be converted into a single summary index value, which can range from −0.111 (for the Japanese population) to 1. Terefore, we adopted a single summary index value for statistical analysis in the present study, as in a previous study [28].
Te collapsed vertebral body height was measured at the anterior (A), centre (C), and posterior (P) borders on lateral radiographs [29]. Te posterior border height of the normal upper vertebral body (NP) was used as a reference. To evaluate fracture severity, the ratios of the A, C, and P border heights of the collapsed body to that of the NP were calculated as follows: (height of each border)/(height of NP) × 100%. Radiological measurements were performed by a physical therapist (H. K. or K. G.) who received specialised training from an orthopaedist (K. M.). Inter-rater (1, 1) and intrarater (2, 1) reliability values were high and acceptable, as described in our previous study [6].
Physical performance was evaluated using the chair stand test (CST), timed up-and-go test (TUGT), and 6-min walking test (6-MWT). During the CST, patients were asked to stand up and sit down, as quickly as possible fve times, with their arms crossed in front of their chest. Te foorto-seat height is 45 cm [30]. Te TUGT was performed by timing the ability of the patients to stand up from a chair, walk 3 m, turn around, walk back to the chair, and sit down [31]. Te 6-MWT was used to assess whether functional exercise capacity was correlated with physical ftness. Tis test measures the distance (m) that a patient can walk on a fat and hard surface for 6 min [32].
Secondary outcomes, except for physical performance measures, were evaluated at admission, two weeks after admission, and at discharge. Physical performance was assessed two weeks after the admission and at discharge.

Required Sample Size.
We used G * Power 333 to perform a preliminary test force analysis and estimate the required sample size [33]. Te power was set at 0.8, and the signifcance level (a) was set at 0.05 [34]. Te efect size for the two-way analysis of variance was set at 0.39, according to our previous study [11]. Te power analysis indicated that 54 patients (27 per group) were required for pre-and postevaluations. We expected a 10% dropout rate [11], and required a minimum of 30 participants per group.

Statistical Analysis.
We used the unpaired t-test or chisquare test to evaluate signifcant diferences in patient characteristics before and after hospital admission between the intervention and control groups. Te chi-square test was used for group comparisons of sex distribution and the proportion of dropouts. To analyse the efects of the rehabilitation program combined with pain management on outcome measures, analysis of variance (ANOVA) was used. A 3 × 2 [time (admission, two weeks postadmission, and at discharge) × group (intervention and control groups)] ANOVA for NRS, GDS, PCS, TSK-11, FIM, and physical activity, and loss of vertebral height was performed. A 2 × 2 [time (admission, two weeks postadmission, and at discharge) × group (intervention and control groups)] ANOVA for physical performance measures was also performed. Posthoc Bonferroni tests were used for specifc comparisons, and signifcance was two-sided. All statistical analyses were performed using SPSS version 22 software (IBM Corp., Armonk, NY, USA). Figure 1 shows a fowchart outlining the study participation. Between 24 January 2019 and 31 May 2021, we screened 166 potential participants aged ≥65 years who were admitted to the hospital with acute osteoporotic VCFs. Ten patients were transferred from another hospital to the Nagasaki Memorial Hospital after acute treatment. Twenty-four patients were unable to walk independently with or without a walking aid before hospital admission, and 38 could not complete the questionnaire because of cognitive impairment. Eighteen patients had severe complications that inhibited rehabilitation, and three had any other fracture type. Four patients were scheduled to be discharged within four weeks of admission, and four did not agree with the study protocol. Finally, we enrolled the remaining 65 individuals in the study and randomly allocated them to either the intervention group (n � 33) or control group (n � 32) (Figure 1). Eight (12.3%) patients withdrew from the trial; four dropped out of the intervention group, and four from the control group. Tere were no signifcant group diferences in study withdrawal (p ≥ 0.99), and no patients dropped out because of the intervention program itself; 57/65 patients completed the intervention: 29 in the intervention group and 28 in the control group. For the intervention and control groups, the length of hospital stay was 51.1 ± 22.0 and 50.1 ± 20.5 days, respectively. Tere was no diference in the length of hospital stay between the two groups (p � 0.863).

Baseline Characteristics.
Tere were no signifcant differences in age, sex, height, body weight, BMI, CCI, TMIG Index of Competence before admission, location and number of acute VCFs, number of previous VCFs, pain medication, orthosis prescription, or MMSE scores between the intervention and control groups. No signifcant diferences in pain intensity, psychological status, ADL, physical activity time, QOL, and the ratios of the A, C, and P border heights of the collapsed body to those of the NP were observed between the two groups ( Table 2).

Effects of Rehabilitation Program Combined with Pain Management on Outcome Measures
In the intervention group, 26 patients (86.2%) achieved the target step count at discharge. Table 3 shows the efects of the interventions on the outcome measures at all time points. In both groups, the mean NRS scores at two weeks and after admission and discharge improved signifcantly compared with baseline. A signifcant main efect of the group was observed (F � 5.135, p � 0.027), and the mean NRS score at discharge was signifcantly higher in the intervention group than in the control group (p � 0.011). Te mean PCS rumination, helplessness, and total scores at discharge in both groups improved signifcantly compared with the baseline, and there were no signifcant differences between the two groups at any time point. Tere were signifcant time-by-group interactions for PCS Pain Research and Management magnifcation (p � 0.012), and the mean score at discharge in the intervention group improved signifcantly compared with the baseline. In both groups, the mean GDS-15 score at discharge improved signifcantly compared with the baseline, and there was no signifcant diference between the two groups at any time point. Te mean TSK-11 score at two weeks and discharge in both groups improved signifcantly compared with baseline, and there was no signifcant difference between the two groups at any time point.
Tere were signifcant time-by-group interactions for daily step count (p < 0.001) and physical activity time (p < 0.001). In both groups, the mean step count and activity time at discharge improved signifcantly compared with those at two weeks after admission. Although we found no signifcant diference in the mean step counts and activity time two weeks after admission between the two groups, the mean values at discharge in the intervention group were signifcantly higher than those in the control group.
In both groups, the mean FIM score and EQ-5D 3L single summary index value at two weeks and discharge improved signifcantly compared with the baseline, and there were no signifcant diferences between the two groups.
Tere were signifcant time-by-group interactions for ratios of the height of the C and P borders of the collapsed body to the posterior border of a normal upper vertebral body (p � 0.006 and p � 0.002, respectively), but not for the A border. In both groups, the ratios of the A and C border heights of the collapsed body to those of the NP at two weeks and discharge decreased signifcantly compared with the baseline. In the intervention group, the ratio of the P-border height of the collapsed body to that of the NP at two weeks and discharge decreased signifcantly compared with the baseline. We found no signifcant diference in the ratios of the A and C border heights of the collapsed body to those of the NP at two weeks and at discharge between the two groups.
Acute VCFs patients admitted to the hospital between January 2019 to May 2021 assessed for eligibility n = 166 Excluded n = 101 transfers from other hospitals (n = 10) unable to walk independently with or without a walking aid before their hospital admission (n = 24) could not complete the questionnaire due to cognitive impairment (n = 38) severe complications that inhibited rehabilitation (n = 18) had any other fracture type (n = 3) scheduled to discharge within 4 week (n = 4) did not agree with the study protocol (n = 4)   Pain Research and Management Te mean CST, TUGT, and 6-MWT values at discharge in both groups improved signifcantly compared with those at two weeks after admission. Tere were no signifcant diferences in the mean CST and TUGT values at discharge between the two groups. Tere were significant time-by-group interactions for the 6-MWT (p � 0.006), and the mean value at discharge in the intervention group was signifcantly longer than that in the control group.

Discussions
Tis study examined the efects of a rehabilitation program combined with pain management targeting pain perception and activity avoidance on multifaceted outcome measures in hospitalised older patients with acute VCFs. Currently, there is insufcient evidence regarding the use of conservative treatment for acute VCFs [35], and there are no universally accepted treatment methods for this condition [36]. Our rehabilitation program, combined with pain management, efectively reduced pain intensity, improved pain catastrophising, and increased exercise tolerance and physical activity levels. Tus, this intervention strategy could be an efective conservative treatment for older patients with acute VCFs.
In our pain management program, the patients' rehabilitation goals were frst set based on shared decisionmaking. Systematic reviews investigating the efectiveness of goal setting in rehabilitation settings have shown that goal setting can improve patient adherence to treatment regimens; however, evidence for improved outcomes remains inconsistent [37]. In contrast, in patients undergoing a lower back pain rehabilitation program, goal setting has been found to have a positive efect on adherence to exercise and   Pain Research and Management self-efcacy, but no efect on treatment outcomes in terms of physical function in rehabilitation settings [38]. Tus, goal setting may be efective in improving adherence to rehabilitation in patients in the intervention group. Pain intensity, PCS magnifcation, and physical activity levels improved signifcantly in the intervention group compared to those in the control group. A systematic review revealed the efect of behavioural change interventions on physical activity levels in hospitalised patients, and the most common equipment used was accelerometers for feedback [39]. In addition, pedometers have been suggested as a successful motivational tool for increasing ambulatory activity in older adults [17]. Tus, our method would have been appropriate for older patients with VCFs to increase physical activity levels. Regarding acute pain and physical activity, it has been reported that patients who underwent lower limb arthroplasty with a substantial correlation between physical activity and postoperative day showed signifcant improvement in pain compared to patients with no correlation between them [40]. Furthermore, we previously reported that low physical activity during the acute phase is associated with higher pain at four weeks [7]. According to these previous studies, increasing physical activity levels is considered a key factor for improving pain in patients with VCF. In a review evaluating cognitive behavioural therapy, Hassett and Williams revealed that increasing physical activity levels is efective for managing chronic pain [41]. Furthermore, we previously investigated combined exercise and pain management that aimed to improve patients' daily behaviour by increasing their daily activities despite pain rather than focusing on eliminating pain, as advised by physical therapists for older people with chronic musculoskeletal pain [11]. Improvements in psychological status and physical activity levels have been reported [11]. It is indicated that professional feedback is important for increasing physical activity [42]; thus, our intervention using a pedometer and determination of step targets based on shared decision-making, involving participants and physical therapists, would be appropriate for patients with VCFs.
Although pain rumination and helplessness in both groups improved signifcantly compared with the baseline, pain magnifcation improvement was detected only in the intervention group. Pain magnifcation indicates a heightened perception of the threat represented by pain symptoms [43]. It has been suggested that the magnifcation component of catastrophising may be a risk factor for heightened pain in the acute phase [44]. In addition, pain magnifcation constitutes a psychological risk factor for chronic postsurgical pain in all surgical models [45]. However, a previous study revealed that an activity diary with goal setting during occupational therapy improved PCS magnifcation along with walking pain and physical activity levels in patients after total knee arthroplasty [46]. Tis result is consistent with our study; thus, improving the magnifcation of PCS would be one of the efects of pain management in this study, and it would be efective in preventing the transition from acute to chronic lower back pain. Te 6-MWT, identifed as exercise tolerance, showed a better improvement in the intervention group than in the control group. Although no studies have investigated the effects of rehabilitation programs on exercise tolerance in older patients with acute VCFs, previous studies have indicated that promoting an increase in daily physical activity using a pedometer was efective in improving the 6-MWT in patients with the pulmonary disease [47,48]. In these studies, not only the 6-MWT but also physical activity, such as step counts and improved signifcantly. Tus, increasing daily physical activity is important in improving exercise tolerance. In addition, we previously reported that persistent severe lower back pain during the acute phase of VCFs was associated with poorer 6-MWT results [6]; thus, improving exercise tolerance may be related to reduced pain, as noted in this study.
Considering the low withdrawal rates in the intervention group, it is suggested that the physical activity pacing approach using pedometers and a diary can be widely accepted by older patients with acute VCFs and is a safe and feasible pain management program. In addition, patients in the intervention and control groups had efectively improved pain, physical performance measures, ADL, QOL, and physical activity levels. Tese improvements are considered to result from the efects of the usual rehabilitation program, conservative treatment, and natural course of recovery from acute VCFs.
Tis study has several limitations. First, it lacks blinding. We could not choose patients blindly when deciding to assign them to either the intervention or control group, which was impossible because of the nature of the intervention. It was quite difcult to blind the physical therapists who intervened because the intervention was easily recognised. Furthermore, the same physical therapist assessed all patients who participated in the intervention program. Terefore, our results may have been infuenced by physical therapists' expertise and/or reporting bias. However, all the physical therapists who participated in this study received a similar level of training before the study. Tus, we believe that the physical therapist's expertise had a minimal efect on this study. Second, in the intervention group, the ratio of the P-border height of the collapsed body to that of the NP at two weeks after admission and discharge decreased signifcantly compared with those at baseline. Tis indicates that the progression of the vertebral deformity was strong in the intervention group. However, we found no signifcant diference in the ratios of the A, C, and P border heights of the collapsed body to that of the NP between the two groups at any time point, and there were no adverse events, such as paralysis of the lower limbs. In addition, previous studies have revealed that the progression of vertebral body deformities is not associated with pain severity [6,49]. Terefore, in this study, vertebral body deformities were considered to have no infuence on pain. Tird, because the study period was only during hospitalisation, the long-term efects of the intervention on pain, physical performance measures, and physical activity levels were not clear. Future studies with a longer followup period are required to investigate the efects of interventions on these outcome measures.
Pain Research and Management 9

. Conclusions
A rehabilitation program combined with pain management that targeted pain perception and activity avoidance signifcantly reduced pain intensity, improved pain catastrophising, and increased exercise tolerance and physical activity levels in hospitalised older patients with acute VCFs compared to the usual rehabilitation program. Our results indicated that this intervention program may be benefcial and could be an efective conservative treatment for patients with acute VCF. Furthermore, it may prevent the transition from acute to chronic pain.

Data Availability
Te data associated with the paper are not publicly available but are available from the corresponding author upon reasonable request.

Conflicts of Interest
Te authors declare that there are no conficts of interest regarding the publication of this paper.