Does Kinesiophobia Modify the Effects of Physical Therapy on Outcomes in Patients With Sciatica in Primary Care? Subgroup Analysis From a Randomized Controlled Trial

Background

A higher level of kinesiophobia appears to be associated with poor recovery in patients with sciatica.

Objective

The aim of this study was to investigate whether kinesiophobia modifies the effect of physical therapy on outcomes in patients with sciatica.

Design

This was a subgroup analysis from a randomized controlled trial.

Setting

The study was conducted in a primary care setting.

Patients

A total of 135 patients with acute sciatica participated.

Intervention

Patients were randomly assigned to groups that received (1) physical therapy plus general practitioners' care (intervention group) or (2) general practitioners' care alone (control group).

Measurements

Kinesiophobia at baseline was measured with the Tampa Scale for Kinesiophobia (TSK) and a single substitute question for kinesiophobia (SQK). Pain and recovery were assessed at 3- and 12-month follow-ups. Regression analysis was used to test for interaction between the level of kinesiophobia at baseline and treatment allocation. Subgroup results were calculated for patients classified with high fear of movement and for those classified with low fear of movement.

Results

Kinesiophobia at baseline interacted with physical therapy in the analysis with leg pain intensity at 12-month follow-up. Kinesiophobia at baseline did not interact with physical therapy regarding any outcome at 3-month follow-up or recovery at 12-month follow-up. When comparing both treatment groups in the subgroup of patients with high fear of movement (n=73), the only significant result was found for leg pain intensity difference from baseline at 12-month follow-up (intervention group: X̅=−5.0, SD=2.6; control group: X̅=−3.6, SD=2.7).

Limitations

The post hoc study design and relatively small sample size were limitations of the study.

Conclusions

In 135 patients with sciatica, evidence shows that patients with a higher level of kinesiophobia at baseline may particularly benefit from physical therapy with regard to decreasing leg pain intensity at 12-month follow-up.

Sciatica is characterized by radiating leg pain and related disabilities.¹ It affects many people and has significant medical, social, and economic impact. The annual prevalence as reported in 9 epidemiologic studies ranges from 2.2% to 34%.² The natural course is generally favorable.^3,4 It is important to adequately inform the patient about the diagnosis and prognosis. The advice to stay active was recently reviewed as “likely to be beneficial.”¹ A recent evidence-based clinical guideline of the North American Spine Society states that there is insufficient evidence to make recommendations for or against the use of physical therapy or structured exercise programs for patients with sciatica.⁵

We previously reported the clinical results of a randomized controlled trial (RCT) that compared general practitioners' (GPs') management alone with GPs' management plus physical therapy in patients with sciatica. We observed that additional physical therapy is effective with regard to global perceived recovery at 1-year follow-up, but not more cost-effective compared with GP care alone.^6,7

In recent spine literature, increasing attention has been paid to identifying subgroups of patients with specific prognostic profiles to offer targeted treatments with the aim to improve treatment effects or better predict prognosis, or both.⁸ The presence of fear of movement might be such a subgroup characteristic.⁹ The term “kinesiophobia” was introduced in 1990; this condition was described as an irrational and debilitating fear of physical movement resulting from a feeling of vulnerability to painful injury or reinjury.¹⁰ Kinesiophobia, together with other psychological factors, is reported to play an important role in the development of chronic symptoms and their perception.¹¹ A recent study of 466 patients with sciatica showed that kinesiophobia was associated with nonsuccess at 2-year follow-up.¹² A significant association of kinesiophobia with outcome also was observed in an observational study of the patients participating in the previously described RCT on the effectiveness of physical therapy.¹³ Besides, in this observational study on the feasibility to replace validated questionnaires by a single question to predict outcome, a significant interaction between the single substitute question on kinesiophobia and treatment allocation to physical therapy was found in a logistic regression analysis, with the presence or absence of leg pain at 1-year as outcome measure.¹³ In theory, physical therapy may reduce fear of movement and improve outcome by informing the patient, by reassurance that movement will not harm, by guidance and promotion of mobility, by optimizing functional ability, and by using the existing movement potential and patient-tailored exercises. Patients with fear of movement, therefore, may form a plausible subgroup that especially benefits from physical therapy.

We hypothesized that physical therapy may reduce any negative effect of kinesiophobia on outcome. Therefore, the aim of the present study was to investigate if kinesiophobia modifies the effect of physical therapy on leg pain intensity and recovery at 3- and 12-month follow-ups in patients with sciatica.

Method

Design Overview

The current study was a post hoc analysis of the RCT comparing GP management alone with GP management plus physical therapy in patients with sciatica in primary care.^7,14 Details on the methods are described in the original publications.^6,7,14

Setting and Participants

Between May 2003 and November 2004, participating GPs (N=112) invited patients with acute sciatica to participate in the trial. The most important inclusion criterion was complaints of radiating pain in the leg below the knee of less than 6 weeks' duration, with the severity of complaints scored above 3 on an 11-point numerical rating scale (0=“no complaints” and 10=“maximum complaints”).⁷

Randomization and Interventions

All patients received care from their GP according to clinical guidelines.¹⁴ Physical therapy consisted of exercise therapy in combination with information and advice about sciatica. The treatment protocol was developed in a consensus meeting with the participating physical therapists.^7,14

Fear of movement at baseline was measured using 2 questionnaires: (1) the Tampa Scale for Kinesiophobia (TSK) and (2) a newly devised substitute question to measure fear of movement on a numerical rating scale. The TSK is a validated questionnaire that measures fear of movement and consists of 17 items rated on a 4-point Likert scale.^10,15 The scores range from 17 to 68 points, with higher scores indicating a higher level of kinesiophobia. Although the TSK may be valuable in daily clinical practice, it is time-consuming to administer. Therefore, we decided during the consensus meeting to apply a single question for measuring kinesiophobia (in this article, referred to as the Substitute Question Kinesiophobia [SQK]). The SQK was introduced with the sentence “You visited your general practitioner because of complaints in your back or leg” followed by the question “How much ‘fear’ do you have that these complaints would be increased by physical activity?” This question could be answered on an 11-point numerical rating scale ranging from 0 (“no fear”) to 10 (“very much fear”). In a previous study, we showed that this SQK may be feasible to replace the TSK for predicting outcome in patients with sciatica in primary care.¹³

Outcomes and Follow-up

Both recovery (global perceived effect) and leg pain intensity at 3- and 12-month follow-ups were used as outcome measures. Global perceived effect was measured on a 7-point scale ranging from 1 (“completely recovered”) to 7 (“vastly worsened”).¹⁶ This rating scale was dichotomized as recovery (“completely recovered” and “much improved”) and no recovery (“slightly improved” to “worse than ever”).⁷ Leg pain intensity was scored on an 11-point numerical rating scale ranging from 0 (“no pain”) to 10 (“unbearable pain”).¹⁷

Data Analysis

To test whether there is an interaction effect between the level of kinesiophobia at baseline and physical therapy, we used regression analyses, with the outcomes recovery and leg pain intensity at 3- and 12-month follow-ups. The regression analysis models contained the level of kinesiophobia, the treatment allocation (physical therapy included or not included), and the interaction between them as independent variables and (according to the outcome measured) recovery or leg pain intensity as the dependent variable. The regression analysis with outcome leg pain intensity was adjusted for leg pain intensity at baseline. As the level of kinesiophobia was measured with 2 different questionnaires, all analyses were performed twice (with either TSK or SQK in the model). Basic statistical assumptions for linear regression were tested for the analysis with the outcome leg pain intensity, reported when violated, and handled according to up-to-date knowledge. Statistical significance for the interaction test was defined as P<.05.

In addition, for ease of clinical interpretation, descriptive statistics were calculated for patients classified with high fear of movement and those classified with low fear of movement. In a highly cited Dutch study of patients with chronic low back pain, the median TSK score of 37 was used as the cutoff for dividing the group into low responders (TSK ≤37) and high responders (TSK >37).⁹ Accordingly, the present study used the same cutoff point.^18–20 For both the patients with high fear of movement and those with low fear of movement, differences in leg pain intensity at 3- and 12-month follow-ups between the randomization groups were assessed using the Student t test, and differences in recovery were assessed by using the chi-square test. In addition, the Student t test was used to test leg pain difference from baseline between the randomization groups.

Patients without complete questionnaires at 3- or 12-month follow-ups were excluded from the analyses. At 3- and 12-month follow-ups, clinical outcomes were missing for 7% and 13% of the patients, respectively.⁷ Four patients in the physical therapy group (6%) and 3 patients in the control group (4%) received surgery.⁷ As these numbers of surgical intervention for sciatica were small, we did not correct for them in the analyses. Baseline differences between patients with and without complete questionnaires were (depending on the type of variable) assessed by comparing means or percentages.

Role of the Funding Source

The Dutch Health Care Insurance Board (CvZ) funded the original RCT. The funding source had no involvement in the design, conduct, or reporting of results.

Results

A total of 135 patients were included and randomized to treatment groups (Fig. 1); of these patients, 67 received GP care plus physical therapy (intervention group) and 68 received GP care alone (control group) (Tab. 1). Patients in the intervention group reported means of 6.7 and 9.7 physical therapist treatments at 6- and 12-week follow-ups, respectively. At 3 months after randomization, 68% of the patients reported recovery (73% in the intervention group versus 63% in the control group), and patients reported a mean leg pain intensity of 2.6 (2.3 in the intervention group versus 2.8 in the control group). At 12 months after randomization, 73% of the patients reported recovery (82% in the intervention group versus 63% in the control group), and patients reported a mean leg pain intensity of 2.1 (1.8 in the intervention group versus 2.4 in the control group). The missing patients at the 12-month follow-up had a significantly higher level of kinesiophobia at baseline according to the SQK compared with the nonmissing patients (5.2 versus 3.8, respectively; P=.04). There was no significant difference in any of the other characteristics.

Figure 1

Flowchart of participant enrollment, allocation, and analysis throughout the study. GP care=general practitioners' care, PT care=physical therapy care. Some patients were excluded for multiple reasons.

Table 1

Baseline Characteristics of the Study Population^a

Variable	General Practitioners′ Care Plus Physical Therapy (n=67)	General Practitioners′ Care Only (n=68)
Age (y)	42.2 (9.6)	42.9 (11.9)
Male sex, n (%)	29 (43)	41 (60)
Body mass index (kg/m2)	25.6 (4.1)	26.8 (4.9)
Symptom duration (d)	12.1 (10.1)	14.2 (10.2)
TSK score (17–68)	39.0 (5.8)	41.0 (7.1)
High TSK score (<37), n (%)	38 (57)	48 (71)
SQK score (0–10)	4.0 (2.6)	4.0 (2.8)
NRS leg pain score (0–10)	6.3 (2.2)	6.3 (2.2)

^a

Values represent mean (SD) unless otherwise indicated. TSK=Tampa Scale for Kinesiophobia (higher scores indicate more kinesiophobia), SQK=Substitute Question Kinesiophobia (higher scores indicate more kinesiophobia), NRS=numerical rating scale (higher scores indicate more pain).

Interaction Effect

There was no interaction effect between the level of kinesiophobia at baseline (TSK and SQK) and treatment allocation (physical therapy included or not included) in the regression analyses predicting perceived recovery at 3- and 12-month follow-ups (eTabs. 1 and 2, available at ptjournal.apta.org, respectively).

There was no interaction effect between kinesiophobia at baseline (TSK and SQK) and treatment allocation (physical therapy included or not included) in the regression analyses predicting leg pain intensity at 3-month follow-up (Tab. 2). Patients with higher levels of kinesiophobia at baseline reported higher leg pain intensity at 12-month follow-up (P<.01). However, the TSK score at baseline tended to interact with physical therapy in the regression analysis predicting leg pain intensity at 12-month follow-up (P=.07). Furthermore, the SQK score at baseline showed a significant interaction with physical therapy in the regression analysis predicting leg pain intensity at 12-month follow-up (P<.01) (Tab. 3).

Table 2

Linear Regression Analysis With Leg Pain Intensity at 3-Month Fol ow-up (n=126)^a

TSK	Beta (95% CI)	P	SQK	Beta (95% CI)	P
Randomization to PT	3.1 (−2.8, 9.1)	.30	Randomization to PT	0.4 (−1.2, 2.0)	.63
TSK	0.1 (−0.0, 0.2)	.17	SQK	0.2 (−0.0, 0.5)	.07
Interaction between TSK and randomization to PT	−0.1 (−0.2, 0.1)	.23	Interaction between SQK and randomization to PT	−0.2 (−0.6, 0.1)	.19
Baseline leg pain	0.3 (0.1, 0.5)	<.01	Baseline leg pain	0.3 (0.1, 0.5)	.01

^a

TSK=Tampa Scale for Kinesiophobia (17–68; higher scores indicate more kinesiophobia), Cl=confidence interval, SQK=Substitute Question Kinesiophobia (0–10; higher scores indicate more kinesiophobia), randomization to PT=allocation to group receiving physical therapy plus general practitioners' care, baseline leg pain=leg pain intensity on numerical rating scale (0–10; higher scores indicate more pain).

Table 3

Linear Regression Analysis With Leg Pain Intensity at 12-Month Follow-up (n=117)^a

TSK	Beta (95% CI)	P	SQK	Beta (95% CI)	P
Randomization to PT	4.0 (−1.0, 9.0)	.11	Randomization to PT	0.8 (−0.5, 2.1)	.24
TSK	0.1 (0.0, 0.2)	<.01	SQK	0.4 (0.2, 0.6)	<.01
Interaction between TSK and randomization to PT	–0.1 (−0.2, 0.0)	.07	Interaction between SQK and randomization to PT	–0.4 (−0.7,−0.1)	<.01
Baseline leg pain	0.1 (−0.1, 0.3)	.20	Baseline leg pain	0.1 (−0.1, 0.3)	.32

^a

TSK=Tampa Scale for Kinesiophobia (17–68; higher scores indicate more kinesiophobia), Cl=confidence interval, SQK=Substitute Question Kinesiophobia (0–10; higher scores indicate more kinesiophobia), randomization to PT=allocation to group receiving physical therapy plus general practitioners' care, baseline leg pain=leg pain intensity on numerical rating scale (0–10; higher scores indicate more pain).

High and Low Fear of Movement

Patients with low fear of movement had a mean TSK score of 33.2 (SD=3.1) and a mean SQK score of 2.9 (SD=2.4). Patients with high fear of movement had a mean TSK score of 43.9 (SD=4.4) and a mean SQK score of 4.6 (SD=2.6). Table 4 presents the subgroup results of the patients with high fear of movement and those with low fear of movement at 3- and 12-month follow-ups. Figure 2 illustrates the subgroup scores for leg pain intensity at baseline and at 3- and 12-month follow-ups. The only significant result was found at 12-month follow-up when comparing both treatment groups for leg pain intensity difference from baseline in the patients with high fear of movement (intervention group: X̅=−5.0, SD=2.6; control group: X̅=−3.6, SD=2.7; P=.027). All other comparisons between the treatment groups of the patients with high fear of movement revealed nonsignificant results (although eyeballing showed a possible trend for better outcomes for the patients in the intervention group). Of the patients with high fear of movement, 72% of those in the intervention group and 57% of those in the control group reported being recovered at 3-month follow-up. A mean difference between the randomization groups of reported leg pain intensity at 3 months of 0.8 was seen in favor of the intervention group compared with the control group (2.3 versus 3.1, respectively). At 12-month follow-up, 76% of the patients in the intervention group reported recovery compared with 55% of those in the control group. A mean difference of 1.0 in reported leg pain intensity at 12 months was seen in favor of the intervention group compared with the control group (1.8 versus 2.8, respectively).

Table 4

Subgroup Results for Patients Classified With High Fear of Movement and Low Fear of Movement^a

Variable	Patients With High Fear of Movement (n=80 at 3 mo/n=73 at 12 m)				Patients With Low Fear of Movement (n=46 at 3 mo/n=44 at 12 mo)
	General Practitioners' Care Plus Physical Therapy (n=36 at 3 mo/n=33at 12 mo)		General Practitioners' Care Only (n=44 at 3mo/n=40 at 12 mo)		General Practitioners' Care Plus Physical Therapy (n=28 at 3 mo/n=27 at 12 mo)		General Practitioners' care Only (n=18 at 3 mo/n=17 at 12 mo)
		Difference From Baseline		Difference From Baseline		Difference FromBaseline		Difference From Baseline
Recovery, 3 mo, n (%)	26 (72)		25 (57)		21 (75)		14 (78)
NRS leg pain score (0–10), 3 mo	2.3 (2.4)	−4.3 (3.0)	3.1 (3.1)	−3.3 (2.8)	2.3 (2.4)	−3.7 (3.0)	2.1 (2.7)	−4.2 (3.5)
SQK score (0–10), 3 mo	3.4 (3.1)	−1.2 (4.3)	3.9 (3.2)	−0.6 (3.0)	2.7 (2.9)	−0.4 (2.8)	1.3 (1.8)	−1.2 (2.6)
Recovery, 12 mo, n (%)	25 (76)		22 (55)		24 (89)		14 (82)
NRS leg pain score (0–10), 12 mo	1.8 (2.0)	−5.0 (2.6)b	2.8 (2.5)	−3.6 (2.7)b	1.7 (1.9)	−4.4 (3.1)	1.5 (2.3)	−4.7 (3.0)
TSK score (17–68), 12 mo	37.4 (7.3)	−5.8 (8.8)	37.4 (7.2)	−6.7 (6.6)	32.4 (5.3)	−1.0 (4.9)	30.7 (7.0)	−2.0 (6.5)
SQK score (0–10), 12 mo	2.4 (2.6)	−2.2 (3.9)	3.1 (2.8)	−1.2 (2.9)	2.1 (2.4)	−1.0 (3.0)	1.1 (2.3)	−1.4 (2.6)

^a

Values represent mean (SD) unless otherwise indicated. NRS leg pain score=leg pain intensity on numerical rating scale (higher scores indicate more pain), SQK=Substitute Question Kinesiophobia (higher scores indicate more kinesiophobia), TSK=Tampa Scale for Kinesiophobia (higher scores indicate more kinesiophobia; the TSK score was not measured at 3-month follow-up).

^b

P=.027for comparing both treatment groups for leg pain intensity difference from baseline at 12-month follow-up in patients with high fear of movement.

Figure 2

Graph showing leg pain intensity on an 11-point numerical rating scale for 4 subgroups at baseline and at 3- and 12-month follow-ups. HPT=patients classified with high fear of movement randomized to group receiving general practitioners' care plus physical therapy, HGP=patients classified with high fear of movement randomized to group receiving general practitioners' care alone, LPT=patients classified with low fear of movement randomized to group receiving general practitioners' care plus physical therapy, LGP=patients classified with low fear of movement randomized to group receiving general practitioners' care alone.

Discussion

Our study provides some indication that patients with a higher level of kinesiophobia at baseline may particularly benefit from physical therapy with regard to decreasing leg pain intensity at 12-month follow-up. The SQK score at baseline significantly interacted with physical therapy in the linear regression analysis with leg pain intensity at 12-month follow-up. Kinesiophobia at baseline did not interact with allocation to physical therapy with any outcome at 3-month follow-up or with recovery at 12-month follow-up. When comparing both treatment groups in the subgroup of patients with high fear of movement, the only significant result was found for leg pain intensity difference from baseline at 12-month follow-up (−5.0 versus −3.6, in favor of the patients in the intervention group).

A limitation of the present study is the post hoc study design (ie, we did not a priori specify the study question, but our interest arose in response to recent literature). Also, because a study question formulated post hoc may complicate the interpretation of results, confirmation of the present results is needed. Although multiplicity may introduce bias in subgroup analyses,²¹ we believe this factor is less relevant for our study because we limited our analyses to baseline kinesiophobia only. Another limitation is the relative small sample size. The small sample size especially limits the interpretation of the additional subgroup analyses where patients were classified into 1 of 4 categories dependent on randomization group and the dichotomized scale of kinesiophobia.

Presenting results on patients classified as “with kinesiophobia” or “without kinesiophobia” eases clinical interpretation. However, important disadvantages are the loss of information by dichotomizing a continuous scale and the difficult choice of the cutoff point. We decided to use the cutoff point most frequently reported in the literature. However, this cutoff point was based on a median TSK score in a different patient population (with chronic low back pain), resulting in 65% of the patients of our study population defined as “suggestive of high fear of movement.” This high number of patients suggests that the cutoff point used may have resulted in a too wide definition of high fear of movement, blurring the interpretation of results.

To our knowledge, this is the first study to investigate if kinesiophobia modifies the effect of physical therapy on outcome in patients with sciatica in primary care. However, the relationship between kinesiophobia at baseline and outcome in patients with sciatica has been studied in secondary care, especially in patients undergoing spine surgery. One study of 466 patients with sciatica of which one-third were treated surgically showed an association between kinesiophobia and nonsuccess.¹² Most, but not all, of the studies investigating the influence of kinesiophobia on outcome after lumbar disk surgery showed an association between a higher level of kinesiophobia at baseline and worse outcomes after lumbar disk surgery.^22–26

As treatment with physical therapy was most intensively given in the weeks after randomization, we would particularly have expected differences in results at 3-month follow-up instead of 12-month follow-up. This contradiction, the relatively small number of included patients, and the post hoc study design make it difficult to draw firm conclusions on any clinical implications. However, the results from the present subgroup analysis show sufficient basis for further research on special treatment effects for patients with kinesiophobia. A larger sample size will be an important requirement for future research.

In conclusion, we found preliminary evidence that patients with a higher level of kinesiophobia at baseline may particularly benefit from physical therapy with regard to decreasing leg pain intensity at 12-month follow-up.

The Erasmus Medical Center Ethics Committee approved the procedures and design of the trial.

The original trial was funded by the Dutch Health Care Insurance Board (CvZ). The present study is partly funded by a program grant of the Dutch Arthritis Foundation.

The original trial was registered at www.controlled-trials.com (ISRCTN68857256).

Author notes

Dr Verwoerd and Dr Luijsterburg provided concept/idea/research design. Dr Verwoerd, Dr el Barzouhi, and Dr Verhagen provided writing. Dr Luijsterburg provided data collection and participants. Dr Verwoerd and Dr Verhagen provided data analysis. Dr Verwoerd, Dr Koes, and Dr Verhagen provided project management. All authors provided consultation (including review of manuscript before submission).