The longitudinal relationship between fear of movement and physical activity after cardiac hospitalization: A cross lagged panel model

Background Little is known about the association between fear of movement (kinesiophobia) and objectively measured physical activity (PA), the first 12 weeks after cardiac hospitalization. Purpose To assess the longitudinal association between kinesiophobia and objectively measured PA and to assess the factor structure of kinesiophobia. Methods We performed a longitudinal observational study. PA was continuously measured from hospital discharge to 12 weeks using the Personal Activity Monitor. The PAM measures time spent per day in PA-intensity categories: light, moderate and heavy. Kinesiophobia was assessed with the Tampa Scale for Kinesiophobia (TSK) at four time points (hospital discharge, 3, 6 and 12 weeks). The longitudinal association between PA-intensity and kinesiophobia was studied with a random intercept cross lagged panel model (RI-CLPM). A RI-CLPM estimates effects from kinesiophobia on objectively measured PA and vice versa (cross-over effects), and autoregressive effects (e.g. kinesiophobia from one occasion to the next). Results In total, 116 patients (83.6% male) with a median age of 65.5 were included in this study. On no occasion did we find an effect of kinesiophobia on PA and vice versa. Model fit for the original model was poor (X2: = 44.646 P<0.001). Best model fit was found for a model were kinesiophobia was modelled as a stable between factor (latent variable) and PA as autoregressive component (dynamic process) (X2 = 27.541 P<0.12). Conclusion Kinesiophobia and objectively measured PA are not associated in the first 12 weeks after hospital discharge. This study shows that kinesiophobia remained relatively stable, 12 weeks after hospital discharge, despite fluctuations in light to moderate PA-intensity.


Introduction
After (acute) cardiac hospitalization, only 17% of patients perform the recommended amount of weekly physical activity (PA) which constitutes: 150-300 minutes of moderate-intensity aerobic PA or 75-150 minutes of vigorous intensity aerobic PA [1,2].Physical activity is defined as 'any bodily movement produced by skeletal muscles that results in energy expenditure' [3].One potential explanation for these low PA-levels might be kinesiophobia (fear of movement).
Kinesiophobia is described as an irrational, debilitating fear of movement and is explained by the fear avoidance model (FAM) [4].The FAM is a biobehavioral model which describes how individuals develop avoidance behavior based on pain related fear.
In patients with coronary artery disease (CAD), the prevalence of kinesiophobia varies from 45-75% after hospitalization to 20% after three months.In addition, a longitudinal study revealed that high levels of kinesiophobia are present in 21.1% after 4 months [5] High levels of kinesiophobia are associated with long disease duration, decreased muscle strength and reduced quality of life [6][7][8].In addition, kinesiophobia is associated with low self-efficacy and maladaptive coping strategies, which in turn impede movement behavior [9][10][11].
Although kinesiophobia is associated with disability, low self-efficacy, and self-reported PA [12][13][14], the relation between kinesiophobia and objective measures of PA is less clear.The longitudinal association between kinesiophobia and daily PA, measured with an accelerometer, has not been prospectively investigated.Using an accelerometer allows for objective measurement of patients daily PA, instead of subjective methods which are prone to recall bias [15].Better understanding of the longitudinal association between kinesiophobia and objectively measured PA is necessary to gain insight in the concept of kinesiophobia and may inform the development of future treatment strategies.
The aim of this study was to explore the longitudinal relationship between kinesiophobia and objectively measured physical activity in the first phase (12 weeks) after hospital discharge, in patients with cardiovascular disease.

Materials and methods
In this prospective observational study, we used a cross-lagged panel model (CLPM) to explore the longitudinal association between kinesiophobia and objectively measured PA.CLPM's are a type of discrete time, structural equation model, used to analyze panel data in which two or more variables are both measured at two or more points in time.CLPM's aim to estimate the effects of one variable on another at different points in time [16,17].In order to account for stable, trait-like differences between patients, such that the lagged relations pertain exclusively to within patient fluctuations, a random intercept is added to the model (RI-CLPM) at each point time.To fit a RI-CLPM, data is decomposed into within-patient dynamics and stable between patient differences, where the latter, account for unobserved heterogeneity [18].

Sample size consideration
For this study four times points (T = 4) were used.Given the exploratory nature of this study, a sample size of N = 100 was deemed sufficient to reach model convergence [19].

Patient population
Eligible patients had been hospitalized for acute coronary syndrome (ACS), stable angina pectoris (AP), acute heart failure (AHF) or atrial fibrillation (AF) and had undergone a percutaneous coronary intervention (PCI) or electro-cardioversion (ECV).All patients gave written informed consent before participating.Patients were excluded if they were unable to perform physical activity, wear an accelerometer or fill in questionnaires (e.g.due to language problems), or transferred to a nursing home.

Data collection and measurements
Patients were identified, between August 2019 and May 2021, through the electronic health records system of the Amsterdam University Medical Centre.In the hospital, eligible patients were approached by researchers of the Amsterdam University of Applied Sciences and enrolled in this study.The following data were collected from the electronic health records: age, sex, education, marital status, cardiac diagnosis and disease history.Patients were asked by email to complete a questionnaire about kinesiophobia at four time points: at hospital discharge (week 1), 3, 6 and 12 weeks post-discharge.In addition, patients were asked to wear an accelerometer (see below) to assess PA-levels for 12 weeks directly after hospital discharge.
Objectively measured physical activity.At hospital discharge, patients were asked to wear a PAM-AM300 © Personal Activity Monitor (PAM) at waist level (belt, trouser or pocket) during all daily activities.The PAM is an uniaxial accelerometer and measures PA intensity by calculating the time spent doing light, medium or heavy PA.Categories of PA are based on metabolic equivalents of tasks (MET's) and are multiplies of the resting metabolism, reflecting the metabolic rate during PA.Light PA comprise activities such as cooking and doing groceries (<3 MET), moderate PA comprise activities such walking or cycling (3-6 MET) and heavy PA comprise activities as aerobics, running or cycle racing (>6 MET) [20].
The PAM is a validated tool to assess physical activity intensity in patients with various conditions [21].The PAM has substantial reliability (ICC: 0.80; 95% CI: 0.28 to 0.92) and significantly correlates with the Actigraph accelerometer (R spearman : 0.82).Patients were asked to wear the PAM-sensor for 90 days.Scores are presented as minutes of PA per day/per week, in the categories light, moderate, heavy and total activity.

Data analysis
Descriptive statistics.Descriptive data were presented as frequencies (proportion), mean (SD) or median (IQR).Associations between physical activity and kinesiophobia were analyzed with correlational analysis using the spearman rank correlation coefficient.Magnitude of the association was interpreted as small (0.00 to 0.30), medium (>0.30 to 0.50) and large (>0.50) [22].Differences in kinesiophobia and objectively measured PA between hospital discharge and 12 weeks follow up were assessed with the Wilcoxon signed rank test.
The basic RI-CLPM.For this study, TSK-NL Heart scores and total physical activity scores (average minutes per day/per week) of week 1 (hospital discharge), 3, 6 and 12 after hospital discharge were used.Observed scores (kinesiophobia and PA) were decomposed into grand means, stable between-components and fluctuating within-components.This model is illustrated in Fig 1 .In this model, the blue squares represent the observed kinesiophobia and PA scores.The between-components (random intercepts) of kinesiophobia and PA capture the persons' deviations from the grand mean and represent stable differences between patients.The random intercepts are specified as a latent variable with the repeated measures as its indicators, fixing all factor loadings to 1.The within-components (red ovals) are the differences between a patient's observed and expected score, based on the grand means and its random intercepts.The following structural relations between within components were specified: firstly: Autoregressive effects (e.g. from T0-kinesiophobia at T1-kinesiophobia) represent the within carry-over effect from one occasion to the next.If this effect is positive, this implies that an individual who experiences elevated kinesiophobia at time = t, relative to his/her own score, is likely to experience kinesiophobia at time = t+1.Secondly, The cross-lagged effects represent the cross-over effects from one domain to the other (e.g.T0-kinesiophobia to T1-PA).A positive effect implies that deviation from an individual's level of kinesiophobia will likely be followed by a positive deviation in PA.Autoregressive and cross-lagged effects are presented as standardized beta effect size estimates (β).Effect sizes were interpreted as small (<0.29), moderate (0.30-0.49), large (� 0.50) [22].Model fit was assessed with Chi square test for model fit (X 2 ), Comparative Fit Index (CFI) and Tucker Lewis Index (TLI).Model fit was deemed acceptable if X 2 p >0.05, and CFI/TLI >0.95 [23].Missing values of the TSK-NL Heart were: Hospital discharge: 34 (22.8%), 3 weeks: 37 (24.8%), 6 weeks: 42 (28.2%),12 weeks: 54 (36.2%).Of the 149 patients, 33 (22.1%) did not wear the PAM accelerometer.In total, 116 patients were included in the final sample.In the PAM accelerometer data, the total amount of missing data was 14.4%.Little's MCAR test was used to determine patterns of missing data.(Little's MCAR Test Chi Square = 4871,310 DF = 4995, Sig = 0.893).Missing data were handled using Full Conditional Specification Multiple imputation (FCS-MI).Multiple imputation and descriptive statistics were performed in SPSS V28 and the RI-CLPM was performed in Mplus V8.

Demographic and clinical characteristics
In total, 188 patients were assessed for eligibility.After inclusion, two patients died and 149 patients completed the TSK-Heart NL questionnaires.Of these patients, 33 did not wear the PAM accelerometer.Finally, 116 patients were included in the analyses with a median (IQR) age of 65.5 (15) years.The majority of patients were male (83.6%) and lived with a partner (78.4%).Most patients had been admitted for an elective intervention (53.4%) and had undergone PCI (81%).In this study, 45 (38,8%) initiated cardiac rehabilitation (CR).Those that initiated CR had lower levels of light PA (at T0 and T1) compared to those that did not initiate CR. (Table 1).An overview of kinesiophobia levels at each time point is presented in Fig 2 .Median TSK scores (IQR) decreased from week 1: 27.4 (7.28), to 12 weeks: 25 (5.90) (P<0.002).Median light activity (minutes/day/per week) (IQR) improved from week 1: 54.14 (32.04)

Random intercept cross lagged panel model
Large statistically significant autoregressive effects (β > 0.5) for kinesiophobia and total PA were found at all occasions, indicating the presence of large carry-over effects from one occasion to the next.Patients who had elevated levels of total PA at time = t were also likely to have elevated levels of total PA at time = t+1.Similar carry-over effects were found for total PA.One small cross-over effect was found from Total PA-week1 to TSK-Week3: β = 0.15 (95%CI: 0.01 to 0.29) indicating that total physical activity at three weeks was associated with increased kinesiophobia at 6 weeks.Model fit for this model was poor (Chi Square = 44.646P<0.001) (Table 3).This model is presented in

Alternative model: Random intercept model without cross lagged effects
As described above (see demographics and clinical characteristics) kinesiophobia scores decrease slightly, but significantly.Therefore, after evaluation of the previous model, a new model was proposed by the researchers, where kinesiophobia was modelled as a random intercept (latent variable) and PA as autoregressive variable.This resulted in a model with good model fit (Chi Square: 27.541, P<0.12) (Table 4).This model assumes that kinesiophobia is stable over time and is considered a trait-like construct while PA is considered as a dynamic process with PA levels varying over time.No association was found between objectively measured PA and kinesiophobia.This model is presented in

Discussion
The results of this study show that kinesiophobia and objectively measured PA are not associated with each other in the first 12 weeks after hospital discharge.In addition, we found that kinesiophobia remains stable over time and might be seen as a trait, while light and moderate PA-levels increased over time, especially in the first weeks after hospital discharge.
Concerning the association between kinesiophobia and objectively measured PA, contradictory findings are reported in the literature.A recent study, in patients with low back pain, also shows that kinesiophobia is not associated with objectively measured PA, measured with an accelerometer [13].In addition, other studies did not find an association between physical capacity measures such as walking endurance and maximal oxygen consumption and kinesiophobia [24][25][26].In contrast, Ba ¨ck et al., found that patients with high levels of kinesiophobia took fewer steps than those without kinesiophobia 3 to 10 months after cardiac hospitalization [6].In addition, Ba ¨ck et al. reported that cardiac patients with high levels of kinesiophobia avoid high levels of (self-reported) PA [6].In our study, we found that in the first 12 weeks after hospital discharge, all patients avoid activities with high intensity.An explanation might be that patients are physically active but only avoid certain activities that cause distressing  body signals, which are in turn related to kinesiophobia [9].The results of our study support the absence of an association between kinesiophobia and objectively measured PA in the first 12 weeks after hospital discharge.An explanation for these different findings, might be that all cardiac patients perform less PA in the first weeks after cardiac hospitalization, thereby making discrimination between patients with and without kinesiophobia difficult.Nevertheless, kinesiophobia is consistently associated with disability [13,26,27].In our previously conducted path analysis, we found an association between kinesiophobia, anxiety, educational level and self-efficacy, which are in turn associated with disability [11,28,29].
In this study, we used a random intercept cross-lagged panel model to account for between-patient unobserved heterogeneity by decomposing the data into between-person and within-person variation at various time points.Our final model clearly showed that kinesiophobia is best explained as a stable between-factor and PA-intensity as a dynamic process.This is an important finding and suggests that high levels of kinesiophobia might continue to exist in the months after hospital discharge.
Although the results of our study suggest that kinesiophobia does not impact PA, it is important to target kinesiophobia since patients with kinesiophobia have less self-efficacy, higher levels of anxiety and are less likely to initiate cardiac rehabilitation [9,10].Our previous qualitative study, revealed that kinesiophobia is related to negative beliefs about PA ("By being careful with unnecessary movements I can prevent my heart problems from worsening" and "If I tried to be physically active/exercise my heart problem would increase").This study also showed that patients who were exposed to PA, and supported by an informal caregiver during PA in the early phase after hospital discharge, had lower levels of kinesiophobia [9].Gradual exposure to PA after cardiac hospitalization might improve kinesiophobia.Currently, kinesiophobia, in patients with musculoskeletal complaints, is targeted with exposure in vivo [30][31][32][33].Exposure in vivo, in the form of CR, might help to alleviate fear avoidance beliefs in cardiac patients by gradually exposing patients to PA and thereby altering negative beliefs about PA [33].Unfortunately, exposure to PA, under the guidance of an experienced physical therapist, is limited to those that are referred to CR.Therefore, adequate referral to CR of those with high levels of kinesiophobia is important.

Strengths and limitations
This study has several strengths.First, we explored the association between kinesiophobia and PA using a longitudinal design.The longitudinal design allowed us to explore the nature and reciprocity of kinesiophobia and PA after cardiac hospitalization.Second, using the PAM sensor allowed us to measure PA-intensity instead of step-count and thereby giving more detailed insight of PA-patterns in the first 12 weeks after hospital discharge.Third, the RI-CLPM in which we decomposed our data in a between and within part, enabled us to capture individual fluctuations and account for unobserved heterogeneity.
Some aspects of this study need consideration.First, in this study we investigated the relation between kinesiophobia measured with the TSK-NL Heart.Although the TSK-NL Heart has been found reliable and valid, it might not capture the whole construct of kinesiophobia [34].The results of our previously conducted qualitative study suggest that kinesiophobia is related to distressing body signals, inconsistent information and passive coping style [9].Future studies should therefore consider testing the association of PA measures with other fear-related constructs such as: the cardiac anxiety questionnaire (CAQ) [35] or fear avoidance beliefs questionnaire (FABQ) [36].Second, patients were included in this study at hospital discharge.In the first phase after hospital discharge, patients with no kinesiophobia also might have low PA-levels, which makes discrimination between patients difficult.In addition, the original PA-levels before hospital were unknown.It is therefore difficult to determine if PAlevels were influenced by kinesiophobia.However, this study does show that improvements in PA-levels are not associated with improved levels of kinesiophobia.Future studies should assess PA-levels before hospital admission, changes in PA and kinesiophobia scores over a longer times period.Third, a substantial proportion of patients (38.3%) dropped out of this study which might have led to a selection of patients and might have influenced our findings.It's unknown of those that dropped out of this study had higher levels of kinesiophobia.Fourth, although we chose to include a heterogenous sample, all patients were discharged after an, acute or elective, interventional procedure.We chose not to include surgical patients since their PA patterns after hospitalization differ from our population.Nevertheless, a variety of TSK-scores and PA-levels were found in our sample and was therefore deemed sufficient to explore the association between kinesiophobia and PA.Future studies should also evaluate the use of graded exposure to improve fear avoidance beliefs in patients with kinesiophobia in studies with a randomized design.

Conclusion
Kinesiophobia is not associated with objectively measured physical activity in the first 12 weeks after hospital discharge.Levels of kinesiophobia remained stable during this study.Physical activity, on the other hand, is a dynamic process in time.In the first 12 weeks after hospital discharge light and moderate levels of PA fluctuated while high levels of PA were avoided by patients.Future studies should investigate the association between kinesiophobia and PA, using different anxiety measures and over a longer time period.