Journal of Occupational Rehabilitation

, Volume 17, Issue 4, pp 652–666

Factors Explaining Improvement of Isoinertial Lifting-Capacity

Authors

  • Katharina von Garnier
    • Department of Physical Medicine and RehabilitationLudwig-Maximilians-University
  • Thomas Ewert
    • Department of Physical Medicine and RehabilitationLudwig-Maximilians-University
  • Robert Freumuth
    • Department of Physical Medicine and RehabilitationLudwig-Maximilians-University
  • Heribert Limm
    • Department of Physical Medicine and RehabilitationLudwig-Maximilians-University
    • Department of Physical Medicine and RehabilitationLudwig-Maximilians-University
    • ICF Research Branch of the WHO CC FIC (DIMDI), IHRSLudwig-Maximilians-University
    • Swiss Paraplegic Research
Article

DOI: 10.1007/s10926-007-9099-7

Cite this article as:
von Garnier, K., Ewert, T., Freumuth, R. et al. J Occup Rehabil (2007) 17: 652. doi:10.1007/s10926-007-9099-7

Abstract

Introduction A clearer understanding of the factors involved in improving lifting-capacity may assist professional health workers to enhance patient’s functioning and minimize chronic back pain. However, few studies have examined this association. This study is part of a trial comparing two secondary back pain prevention programs. It aims to identify anthropometric, physical, psychic and demographic baseline variables (baseline model), and over time change variables (comprehensive model), which explain the alteration of lumbar isoinertial lifting-capacity, from baseline to post-treatment. Methods The association between these variables’ baseline- or change values, and the change of lifting-capacity (PILE-test) over time, were analyzed with multiple regression analyses. Potential variables for the regression analyses were identified within a standardized stepwise selection process. Results In the baseline model, 35.2% of the variance in lifting-capacity was mainly explained by a low baseline score of lumbar lifting-capacity, high body weight and gender. In the comprehensive model, 41.9% could be mainly explained by the same baseline variables, an increase of perceived exertion during the PILE-tests and decrease of fear-avoidance caused by work. Conclusions The results suggest that treatments to improve lifting-capacity in individuals with mild low back pain should particularly address the reduction of fear-avoidance beliefs. Although strong conclusions cannot be drawn from this study due to methodological limitations, they may be helpful to assign patients to appropriate and most beneficial treatment programs, as well as to develop specific programs. Fear-reduction may be an important target for early interventions in regard to functional capacity.

Keywords

Lifting-capacityProgressive isoinertial lifting evaluation (PILE)Functional capacityMultiple regressionNurse

Introduction

Occupational low back pain (LBP) is a major burden to society [13]. In addition to human suffering, it causes a substantial economic burden due to absence from work and the wide use of medical services [46]. The largest expenses due to LBP have been identified within individuals having recurrent disabling episodes of it, and developing long term pain and disability, which results in activity restrictions [3, 7, 8]. Therefore, there is a need for effective preventive interventions to avoid the development of chronic problems, reduced function and high costs [7, 9, 10]. Nursing is among the high-risk occupations regarding LBP problems [11]. Frequent patient handling appears to be associated with increased incidence of LBP [12], and lifting patients is one of the most serious risk factors for LBP in the nursing population [11].

Information about lifting-capacity may be of value to professional health workers involved in ‘functional capacity evaluation’ and the estimation of the extent of a disability [13]. Additionally, it can serve as a screening tool for the identification of risk patients, or in the return to work process, when hiring new employees, or for the transfer from one job to another [14, 15]. Knowledge about the factors involved in changes of lifting-capacity may assist professional health workers in their understanding of how to improve lifting-capacity, and minimize disability or chronic problems due to back pain. Furthermore, knowledge about the beneficial factors for LBP could help to assign patients to appropriate treatment programs. Since workers with chronic back pain have a lower lifting-capacity than the healthy population [12, 16], an increase in lifting-capacity may be an important target for early interventions, in regard to functional capacity [17].

For back pain treatment more specific information about influences on the improvement of lifting-capacity may be very valuable. There are several theoretical models and equations with biomechanical, physiological and psychophysical approaches to safe lifting [1822]. But to the authors’ knowledge, there is no commonly agreed theoretical model explaining improvement of isoinertial lifting-capacity, integrating anthropometric, physical, psychic and demographic variables.

According to Mayer et al. [12], the most appropriate way to test lifting-capacity, in order to measure patient progress in functional restoration programs for spinal disorders, is by use of dynamic isoinertial tests. In order to enhance lifting-capacity, it is important to consider variables with influence on this parameter. There are some physical, demographic and anthropometric variables known as being associated with lifting-capacity: Prior cross-sectional studies in populations with—mostly chronic—back pain reported that perceived effort [23], body height [24] and body weight [24] were positively related to lifting-capacity. Matheson et al. [15] could show in healthy female participants that the predictive power of back muscle strength and aerobic capacity taken together could account for 43% of the variance of lifting-capacity, whereas Carlsöö [14] stated it to be impossible to predict a person’s lifting-capacity from a knowledge of muscle strength, body height and body weight in healthy subjects, due to lack of significant correlation between these variables. Gender [2325] was related to lifting-capacity.

Longitudinal studies in people with chronic back pain could find no significant positive association between general activity and the improvement of lifting-capacity over time [26]. A longitudinal study with healthy subjects [27] showed a significant influence of cardiovascular endurance, maximal force in bending and body height, but no significant influence of muscle strength and trunk-muscle endurance on the improvement of lifting-capacity.

Furthermore, there are some psychic and pain variables known to be associated with lifting-capacity: In cross-sectional studies of populations with—mostly chronic—back pain fear-avoidance [23, 28], back pain severity [29], pain interference and depression [28], these variables were inversely related to lifting-capacity. In contrast back pain duration [23] and coping mechanisms [28] were positively related to this parameter. A longitudinal study in patients with chronic back pain [26] could find no significant association between pain interference, affective distress, depression and fear-avoidance and the improvement of lifting-capacity over time. There is conflicting evidence, whether a decrease in pain severity correlates significantly [26] with an increase of lifting-capacity, or not [30].

To summarize the current position, there is still insufficient knowledge from longitudinal studies about over time change variables explaining improvement of isoinertial lifting-capacity.

This study is part of a trial comparing treatment effects of an exercise versus a multidisciplinary secondary prevention program for nurses with LBP [31].

The objective was to identify anthropometric, physical, psychic and demographic variables explaining change of lumbar isoinertial lifting-capacity in a population with prior episodes of LBP.

In order to explore what kind of attributes participants should have in order to benefit from prevention programs, the ‘initial position’ (baseline) was considered. Knowing more about beneficial attributes that increase lifting-capacity may be helpful to assign patients to appropriate and most beneficial treatment programs. Therefore, a precursor regression model (in the following called ‘baseline model’), identifying baseline variables associated with change of lifting-capacity, was developed.

Subsequently, a comprehensive regression model (called ‘comprehensive model’) was developed in order to find out which change variables (= Δ post-treatment scores − baseline scores), in addition to those in the baseline model investigated variables, were associated with change of lifting-capacity. Therefore, in the comprehensive model baseline variables as well as changeable variables were inserted. The knowledge of variables, whose change is associated to improvement of lifting-capacity within such prevention programs, may be valuable for the refinement of future, and even more effective treatments.

Taking into account the studies mentioned above, the authors hypothesized that in the comprehensive model, especially the following changes in independent variables would contribute to the increase of lifting-capacity: (1) decrease of fear-avoidance from baseline to post-treatment, and (2) decrease of back pain severity from baseline to post-treatment.

Additionally, for both models many other variables with potential association to the increase of lifting-capacity were inserted in an explorative way. These variables were either targeted by at least one of the intervention programs or they were asked for by aforesaid measures.

Methods

Design

A randomized controlled parallel-group study of an exercise program compared to a multidisciplinary prevention program was conducted at the Ludwig-Maximilians-University of Munich, Germany. Two hundred and two nurses were recruited from several Munich hospitals between July 2003 and February 2005. The participants were consecutively randomized with a sealed envelope. The study used a test-retest design. The interval in between was 13 weeks.

Participants

Inclusion criteria were amongst others at least one LBP episode in the last 2 years and employment status as nurse. Exclusion criteria were acute or chronic pain leading to a sick certificate. For a detailed description of the inclusion and exclusion criteria see Limm et al. [31].

Interventions

The exercise program was similar to the one described by Klaber Moffett et al. [32] and consisted of 11 sessions of 1 h, over a period of 13 weeks. It included stretching, strengthening and relaxation exercises and a swimming session.

The multidisciplinary prevention program consisted of 17 sessions of 1¾ h and one session of half an hour, over 13 weeks. Additionally to the same exercise classes, but without the swimming session, the participants received eight workplace specific interventions (e.g. ergonomic advice and work hardening), seven sessions of segmental stabilizing exercises [33] and five cognitive-behavioral interventions about coping with pain and stress.

Measures

Before the baseline assessment demographic data such as age, gender, educational and professional status, family background as well as data about health care utilization was collected. With a self-administered Comorbidity Questionnaire (modified from Sangha et al. [34]), the participant was questioned about 17 potential comorbidities.

At baseline assessment (beginning of first week) and post assessment (end of 13th week), the severity and grade of pain were assessed using the 11-item Korff-index [29] with its numerical scale ranging from zero to 10. Additionally, the participants evaluated in self-administered questionnaires the number of days in pain in the previous 12 months, their average sports frequency per week and in the previous weeks as well as social support to sports by their families and friends. Anthropometric data like body weight (noted in kg) was measured with a standard clinic balance scale that was calibrated for weight and the participants were asked for their body height (noted in cm).

Lifting-capacity was measured with the progressive isoinertial lifting evaluation (PILE). In prior studies [12, 16, 25, 3538], the PILE has been described as a standardized psychophysical lifting test with good psychometric properties, relevant to functional daily living tasks and frequent lifting, able to distinguish between disabled subjects due to back pain and healthy control subjects. The test involves a lumbar and cervical component. In this study, only the lumbar component was performed by the subjects. They lifted defined, increasingly heavy weights in a box from the floor onto a table of 75 cm and back, regardless of their body height. Every 20 s weight was added while they performed four lifting cycles. The obtained variable ‘isoinertial lumbar lifting-capacity’ expresses uppermost weight lifted in the last acceptable cycle, noted in kg. The test has several endpoints described in detail by Mayer et al. [12]. It is terminated by the tester if the subject (1) reports the inability to continue (psychophysical end point); (2) reaches the designed heart rate limit based on age (aerobic end point, controlled via a Polar Sport Tester; (3) reaches a predetermined level endpoint at 55% of body weight that is undesirable to exceed in a rehabilitation setting (safety end point); or (4) exceeds 20 s during four lifting cycles.

The rate of perceived exertion during completion of the PILE was rated by the Borg Scale [39]. The original scale is graded from 1 to 21, whereas in this study a modified scale graded from 1 (very, very light) to 10 (very, very heavy) was used. Static trunk-muscle extension endurance was investigated with the Biering-Sørensen test [40]. The participant was positioned prone with his hips over the edge of the examination couch and his lower extremity fixed to the couch by belts. He was asked to maintain the unsupported upper body as long as possible in a horizontal position and his maximum holding time was noted in seconds. Isometric muscle strength (flexion and extension) of the knee and elbow joints was measured by a professional health worker using a hand-held pull gauge with a continuous scale. The participant was instructed to increase muscle strength gradually to his limit. A muscle strength index was calculated as mean of the subjects’ individual standardized scores [41].

Pain severity and pain interference (perceived impact of pain on a patient’s life) were collected with the German version [42] of the West Haven-Yale Multidimensional Pain Inventory (MPI-D) [43]. The MPI is a comprehensive questionnaire with 12 scales and 60-items. It was developed to assess the experience of chronic pain. The two scales ‘pain severity’ and ‘pain interference’ used in this study consist of 13 items altogether and measure the severity and interference with daily life due to pain on a numerical scale from zero to six. Fear-avoidance beliefs were collected with the German version [44] of the Fear-Avoidance Beliefs Questionnaire (FABQ-D) [45]. The questionnaire was developed to assess a patient’s beliefs about pain and work/physical activity. Its two scales ‘fear-avoidance beliefs about work’ and about ‘physical activity’ have altogether 11 items and a numerical scale ranging from zero to six. Generalized self-efficacy was investigated using the 10-item scale of Schwarzer [46], which has a scale ranging from one to four. Basler’s questionnaire about self-efficacy and good postural habits for the prevention of back pain [47] with its 10 items and a numerical scale ranging from one to four was employed to measure back pain specific self-efficacy. To collect cognitive evaluation of pain the coping strategies questionnaire-revised (CSQ-R) [48] was administered. On a numerical scale from zero to six its 50 items measure cognitive coping strategies on six scales: ‘distracting’, ‘distancing’, ‘coping self instructions’, ‘ignoring pain’, ‘praying’ and ‘pain catastrophizing’ (the latter meaning the degree to which patients have negative self-statements and catastrophizing thoughts and ideation when in pain). Depression was collected by the German version of the Center for Epidemiologic Studies Depression Scale [4951]. It was developed by the Center for Epidemiological Studies of the National Institute of Mental Health and is a 20-item questionnaire with a scale ranging from one to four, measuring depressive symptoms with an emphasis on affective symptoms. Stress was collected using a 16-item-instrument measuring daily hassle on a scale from zero to five [52].

Statistical Analyses

General Analyses

Statistical analyses were conducted using the Statistical Package for Social Sciences (SPSS), version 12.0 [53].

Only participants with data for baseline and post-assessment were included and analyzed in terms of available case analysis according to the intention to treat principles [54]. All variables were tested for normal distribution (Kolmogorov-Smirnov Test). Baseline to post-changes in the variables were tested by T-Tests or Wilcoxon Tests.

Preliminary Selection of Variables for the Final Regression Analyses

In the following, the change (= Δ post-treatment score − baseline score) of lumbar lifting-capacity is referred to as ‘dependent variable’. For the selection of appropriate independent variables for the final regression models physical, psychic and demographic variables were classified to chunks. The first chunk comprised—mostly anthropometric—control variables for which a certain impact on the change of lifting-capacity was assumed. Nominal variables were recoded to dummy variables.

For the two regression models, independent variables with correlations with P-values < .2 of the bivariate (baseline model) or partial correlation coefficient (comprehensive model) were selected.

Then for both models multiple regression analyses were calculated separately within the chunks, considering the control variables and a stepwise insertion of the variables. The reason for this procedure was to separately identify the variables with most influence on the dependent variable within each chunk and to thereby reduce the number of variables for the subsequent regression analyses.

Final Regression Analyses

Multiple forward regression analyses were performed to investigate the individual and joint contributions of the independent variables remaining in the final models of the chunks to the dependent variable.

Results

Participants

About 19 (9.4%) of the 202 participants randomized for the study dropped out before baseline-assessment, 11 (5.4%) from the exercise group and 8 (4.0%) from the multidisciplinary group. Of the remaining 183 nurses, 21 (11.5%) did not attend the post-assessment. Therefore, 162 (88.5%) subjects could be included in the data analysis. Their baseline characteristics are shown in Table 1. At baseline, only for lumbar lifting-capacity (P = .018) and perceived exertion during the PILE (P = .000) higher scores have been found in the multidisciplinary group (n = 80) than in the exercise (n = 82) group.
Table 1

Baseline characteristics of the study group (n = 162)

Characteristics

Age [years]

39.98 (11.2)a

Gender (female)

150 (92.6%)

Partnership (yes)

88 (54.3%)

High school diploma graduate

29 (17.9%)

Working fulltime

127 (78.4%)

Body height [cm]

166.92 (7.0)a

Body weight [kg]

72.01 (14.8)a

Sports frequency per week

2.21 (2.0)a

Comorbidity score

1.49 (1.6)a

Pain intensityb

    0

62 (38.3%)

    1–2

87 (53.7%)

    3–4

12 (7.4%)

    5–6

1 (.6%)

Pain gradesc (n = 161)

    Grade I

107 (66.5%)

    Grade II

37 (23.0%)

    Grade III

13 (8.0%)

    Grade IV

4 (2.5%)

Days in pain (LBP) in the last 12 months

71.66 (95.8)a

Persons with sick leave in the last 12 months

20 (12.3%)

Health care utilization in the last 12 months

   Persons reporting physician visits

70 (43.2%)

   Persons reporting non-physician service visitsd

51 (31.5%)

   Persons reporting use of analgesic medication

88 (54.3%)

aMeans (SD)

bAccording to MPI-D; 0 = no pain; 6 = unbearable pain [37]

cGrade I: low disability, low pain intensity; grade II: low disability, high pain intensity; grade III: high disability, moderately limiting; grade IV: high disability, severely limiting [29]

dIncluding physiotherapist, occupational therapist, psychological therapist and alternative practitioner visits

LBP = low back pain

A withdrawal analysis comparing the 21 withdrawals to the remaining 162 participants showed no differences (P < .05) concerning baseline characteristics except for the fact that the withdrawals reported more sick leave days in the last 12 months (P = .026) and a higher sports frequency per week (P = .030).

Preliminary Analyses

Most variables showed improvement from baseline- to post-treatment (Table 2). Lumbar lifting-capacity improved over time in both programs (P = .000). Missing variables varied between zero and 13 variables per person and instrument. The variables’ distribution allowed the use of parametric procedures. Therefore, multiple regression analyses were computed. The preliminary stepwise variable selection is presented in Table 3.
Table 2

Baseline to post-treatment changes in the outcome measures (n = 162)

Chunks

Variables

Baseline (mean ± SD)

Post-treatment (mean ± SD)

P-value

Control variables

Lumbar lifting-capacity

20.99 (8.34)

26.07 (10.79)

.000b

Body weight

72.03 (14.81)

72.34 (15.06)

.138b

Physical variables

Perceived exertion during PILE

5.82 (1.73)

6.34 (1.66)

.007b

Muscle strength index

58.86 (11.30)

60.47 (11.49)

.012a

Static trunk-muscle extension endurance

92.59 (54.88)

112.66 (53.89)

.000b

Pain severity

1.52 (.98)

1.19 (.96)

.000b

Pain interference

1.45 (1.08)

1.02 (.91)

.000b

Psychic variables

Depression

11.51 (8.93)

10.60 (8.46)

.040b

Generalized self-efficacy

2.90 (.45)

2.96 (.44)

.029b

Back pain specific self- efficacy

3.06 (.62)

3.31 (.58)

.000a

Social support to sports by family

2.05 (.69)

2.07 (.72)

.632a

Social support to sports by friends

1.92 (.68)

1.93 (.70)

.925b

Stress

1.51 (.76)

1.44 (.75)

.058a

CSQ—Distracting

2.63 (.88)

2.61 (.88)

.654a

CSQ—Distancing

.90 (.87)

1.00 (1.00)

.374b

CSQ—Coping self instructions

2.81 (1.02)

2.50 (1.03)

.000a

CSQ—Ignoring pain

2.55 (.88)

2.46 (.92)

.126a

CSQ—Praying

1.17 (1.31)

1.07 (1.25)

.253b

CSQ—Pain catastrophizing

.92 (.72)

.75 (.73)

.000b

Fear-avoidance beliefs about work

2.31 (1.25)

2.25 (1.33)

.500a

Fear avoidance beliefs about activity

2.88 (1.17)

2.63 (1.19)

.014a

aT-test

bWilcoxon test

Table 3

Preliminary variable selection for final regression models

Chunks

Independent variables

Baseline regression model

Comprehensive regression model

Bivariate correlations

Chunkwise regression analyses

Final model

Partial correlations

Chunkwise regression analyses

Final model

Control variables

Lumbar lifting-capacity

−.545***

X

X

1,000***

X

X

Affiliation to one of the programs

−.114

X

X

X

X

Gender

−.134

X

X

X

X

Age

.084

X

X

−.095Δ

X

X

Body height

.164*

X

X

X

X

Body weight

.148

X

X

.104Δ

X

X

Physical variables

Perceived exertion during PILE

−.258***

  

.260***Δ

X

X

Muscle strength index

.191**

X

 

.070Δ

  

Static trunk-muscle extension endurance

−.109

  

.004Δ

  

Pain severity

.034

  

−.129Δ

  

Pain interference

.029

  

−.155Δ

  

Psychic variables

Depression

.061

  

−.173

  

Generalized self-efficacy

.049

  

.033Δ

  

Back pain specific self-efficacy

.080

  

.121Δ

  

Social support to sports by family

.082

  

.097Δ

  

Social support to sports by friends

.062

  

.003 Δ

  

Stress

.050

  

.021Δ

  

CSQ—Distracting

.006

  

−.150Δ

  

CSQ—Distancing

.003

  

.054Δ

  

CSQ—Coping self instructions

.013

  

−.165

  

CSQ—Ignoring pain

.003

  

.003Δ

  

CSQ—Praying

.084

  

.027Δ

  

CSQ—Pain catastrophizing

.015

  

−.189

X

 

Fear-avoidance beliefs about work

.072

  

−.188

X

X

Fear-avoidance beliefs about activity

.004

  

.263***Δ

  

Demographic variables

Comorbidity score

.019

     

Sports frequency in previous 4 weeks

.010

     

High school diploma graduate

.123

     

Partnership

.046

     

Δ = Change score (post-treatment score − baseline score)

X = Remaining in selection procedure for subsequent regression analyses

Bold = Linear or partial regressions (P < .2)

Italic = Control variables entered in all models

Bold italic = Linear or partial regressions (P < .2), Control variables entered in all models

* = P ≤ 0.05

** = P ≤ 0.01

*** = P ≤ 0.001

Regression Analyses

In the final baseline model, 35.2 % (adjusted R2) of the variance in lumbar lifting-capacity could be mainly explained by the control variables: low baseline score of lumbar lifting-capacity, gender and high body weight (Table 4). In the final comprehensive model, 41.9% (adjusted R2) of the variance could be mainly explained by the same control variables contributing significantly in the final baseline model, as well as an increase of perceived exertion during the PILE and decrease of fear-avoidance caused by work (Table 5). No multi-collinearity (r > .8) between the independent variables was found. In both models the residuals were normally distributed, showed no autocorrelation (Durbin-Watson-Coefficient 1.89 and 1.88; [55]) and had homoscedasticity. All variables of the regression models had tolerance values considerably >.1, indicating that the data were not affected by collinearity [55].
Table 4

Final baseline regression model: baseline scores of independent variables explaining change in lifting-capacity (n = 153)

Independent variables

R2

adj. R2

Standardized β coefficient

P

1. Control variables

.378

.352

  

    1.1 Lumbar lifting-capacity (baseline)

  

.594

.000

    1.2 Affiliation to one of the programs

  

.015

.822

    1.3 Gender

  

.194

.012

    1.4 Age

  

.102

.132

    1.5 Body height

  

.088

.276

    1.6 Body weight

  

.147

.037

Table 5

Final comprehensive regression model: baseline scores of control and baseline variables and change scores of independent variables explaining change in lifting-capacity (n = 150)

Independent variables

R2

adj. R2

Standardized β coefficient

P

1. Control variables

.387

.352

  

    1.1 Lumbar lifting-capacity (baseline)

  

.577

.000

    1.2 Affiliation to one of the programs

  

.037

.589

    1.3 Gender

  

.151

.048

    1.4 Age

  

.056

.397

    1.5 Body height

  

.089

.252

    1.6 Body weight

  

.142

.037

    1.7 Perceived exertion during PILE (baseline)

  

.086

.287

    1.8 Fear-avoidance caused by work (baseline)

  

.008

.910

2. Physical and psychic variables

    

    2.1 Δ Perceived exertion during PILE

.435

.399

.259

.001

    2.2 Δ Fear-avoidance caused by work

.458

.419

.166

.017

Δ = Change score (post-treatment score − baseline score)

Discussion

In this study the influence of several anthropometric, physical, psychic and demographic baseline and changeable variables explaining change of lumbar isoinertial lifting-capacity was investigated.

In the baseline model, outside the control variables low baseline score of lumbar lifting-capacity, gender and high body weight, none of the other investigated baseline variables significantly explained additional variance of lifting-capacity. These three attributes seem to be beneficial when assigning participants to treatment programs with the objective of increasing lifting-capacity.

In the comprehensive model, increase in lifting-capacity could be basically explained by the same control variables, but additionally by increase of perceived exertion during the PILE and decrease of fear-avoidance caused by work. When interpreting this model’s results one has to consider that the researchers controlled for all examined baseline variables. Thus, also possibly confounding variables (such as ‘affiliation to one of the programs’) were regarded and it is thereby unlikely to overestimate the variance explained by the change variables.

Regarding perceived exertion during the PILE, it seems reasonable that subjects straining themselves more at post-test then reached higher lifting-capacity scores than baseline. This shows in accordance to Carlsöö [14], that lifting-capacity is also dependent on motivational aspects. However, the change of perceived exertion may also be explained as being a simple test reactivity phenomenon with familiarity with the test, allowing better performance for those participants who over-restricted themselves at baseline. Nevertheless, considering that the sample was composed of nurses for whom lifting is a very frequent task in their daily job, the assessors experienced a high-test motivation already at baseline of the PILE.

Moreover, it could be that, as initially predicted, a decrease of fear-avoidance beliefs about work went together with improved lifting-capacity. The reason for this being that probably participants with reduced fear of pain at post-treatment dared to lift more weight. These results are in line with those of Waddell et al. [45] and Geisser et al. [23], who stated that pain-related fear is being increasingly recognized as an important contributor to poorer performance during functional activity and disability among persons with LBP. These findings are not in accordance with those of a longitudinal study about people with chronic back pain [26], who could find no significant association between fear-avoidance and the improvement of lifting-capacity over time.

Our results are compatible with those of McCracken et al. [26] in a longitudinal study with a chronic back pain population, but they are in contradiction to the second of our initial hypotheses. In the population used for this study, with mild pain symptoms and degrees of disability, reduction of pain severity and pain interference did not explain any change of lifting-capacity. Whereas the cross-sectional studies of Ljungquist et al. [25] and Ruan et al. [28], both with subjects with chronic back pain, found inverse relations between pain, or pain interference, and lifting-capacity. Because of the differences in methods and variable’s compositions in the cited study’s regression analyses a direct comparison is difficult.

The cross-sectional study with healthy subjects of Matheson et al. [15] showed a significant association between muscle strength and lifting-capacity. In this study neither the baseline nor the significant increase of muscle strength over time were significantly correlated to the increase of lifting-capacity (r = −.07). A first explanation therefore could be that the PILE was terminated not only because of the psychophysical end point, which might have reflected strength, but in many cases because of the aerobic or safety end point, so that maximum strength could not be achieved. A second reason could be that the PILE involves an endurance component, which is not represented by the muscle strength index, either. A third possibility could be that “although strength is an important determinant of the capability of an individual to perform an infrequent or occasional lift, capability (maximum-accepted-weight-of-lift) appears to be substantially lower than isometric or isotonic strength maxima” [22, p. 759]. The perceived ‘acceptable’ limit may differ from the ‘real’ strength maxima. Considering variance explained by the variable ‘gender’ in both models, it has to be taken into account that 92.6% of this study’s sample were females.

The results indicate, that factors through which exercise and multimodal back pain programs are successful in increasing functional capacity, might not mainly be an amelioration of physical characteristics such as muscle strength or static muscle endurance. But rather a change of psychic variables, insofar as people correct their irrational appraisals by making experiences differing from their expectations. As stated by Buer and Linton [17], this is already important in early stages of the pain process like the ones this study’s population were in. Persons with a high degree of pain-related fear tend to overestimate the amount of pain they will experience during functional activity, and not actual pain experienced during a task performance [56, 57].

However, the fact that fear-avoidance could be reduced in both an exercise and a multidisciplinary prevention program is interesting, because the exercise program did not involve any psychological intervention. Yet psychic variables showed significant changes, and were dominant in explaining variance in increase of lifting-capacity. May be for these participants the direct experience of lifting without perception of pain was the main mechanism of decrease of fear-avoidance. The finding that only the change score of fear-avoidance contributed significantly, suggests that in this mildly encumbered sample it didn’t matter with how much fear-avoidance the participants started the programs; but rather how much their fear-avoidance could be reduced was the important factor for the increase of lifting-capacity.

As an implication for treatment refinement, this study’s findings propose that in addition to the strengthening of musculature, it might be particularly beneficial to place more emphasis on interventions designed to expose back pain patients to activities they usually avoid, in order to improve the accuracy of anticipated pain during functional activity and thereby by making positive experiences.

The study has a number of limitations. Because of the study’s design, no causal relationships between the independent variables could be drawn. The socio-economic homogeneity of this sample of predominantly female nurses with mild LBP might affect the generalization of the results. Some applied outcome measures have been validated only in chronic back pain populations. Therefore, it is possible that some relevant factors were not uncovered by these measurements and other might be detected by more sensitive measures in a less affected population.

Possibly not all relevant factors explaining change of lifting-capacity, such as self-confidence, motivation, muscular coordination and lifting technique were considered. The three latter may also play a role [14] as well as further anthropometric measures with influence on joint or resisting moments of the spine, such as torso height, pelvic width and pelvic girth [24]. Additionally, some measures relied on self-report which might have introduced a bias in response to social desirability concerns. Finally, in all test situations a learning effect may have improved the post-treatment scores.

Further research and especially longitudinal studies are needed to confirm and broaden the knowledge about causal relationships between the variables and to identify further factors influencing the improvement of lifting-capacity. Additional data about treatment-related processes during follow-up intervals could be helpful.

Conclusions

Lifting-capacity may be of value to professional health workers involved in the measurement of functional capacity evaluation in individuals whose work requires frequent lifting. The results suggest that treatments to improve lifting-capacity in individuals with mild low back pain should also focus on reducing fear-avoidance beliefs about work, although strong conclusions cannot be drawn from this study due to methodological limitations. Fear-reduction may be an important target for early interventions in regard to functional capacity, and could also be useful in the screening of low back pain risk patients.

Acknowledgments

This study was funded by the German Federal Ministry of Health and Social Security (Grant No. 124-43164-1/527). The authors thank Berid Rackwitz, Tina Wessels, Jana John and Benedikt Stark for supporting the accomplishment of the study.

Copyright information

© Springer Science+Business Media, LLC 2007