Journal of Occupational Rehabilitation

, Volume 16, Issue 1, pp 50–59

Development and Validation of a Short-Form Functional Capacity Evaluation for Use in Claimants with Low Back Disorders

Authors

    • Department of Physical TherapyUniversity of Alberta
    • Workers’ Compensation Board-Alberta Millard Health
    • Ph.D, 2‘50 Corbett Hall, University of Alberta
  • Michele C. Battié
    • Department of Physical TherapyUniversity of Alberta
  • Alexander Asante
    • Orion Health
Original Article

DOI: 10.1007/s10926-005-9008-x

Cite this article as:
Gross, D.P., Battié, M.C. & Asante, A. J Occup Rehabil (2006) 16: 50. doi:10.1007/s10926-005-9008-x
  • 116 Views

Objectives: Functional Capacity Evaluations (FCE) are used for making return-to-work decisions, yet FCE's modest predictive ability is currently outweighed by the administrative burden of testing. We attempted to develop a short-form FCE while maintaining comparable predictive ability. Methods: Three databases previously created for evaluating FCE predictive validity were used. Subjects were compensation claimants with low back disorders. FCE measures included items in the Isernhagen Work Systems’ FCE. Days until benefit suspension served as an indicator of return-to-work. Analysis included Cox regression. Results: Three items, floor-to-waist lift, crouching, and standing, were maintained in the short-form FCE. The short-form FCE was found to predict comparably to the entire FCE protocol in two validation cohorts (R2 difference<3%). Subjects meeting job demands on all three items consistently experienced faster benefit suspension. Conclusion: A short-form FCE for determining future work status in claimants with low back disorders was developed. A substantially abbreviated FCE may offer an efficient alternative.

KEY WORDS:

functional capacity evaluationreturn-to-work testingwork capacity evaluationfunctional abilities evaluationpredictionprognosis

INTRODUCTION

Functional Capacity Evaluations (FCE) are commonly used internationally for making return-to-work decisions for injured workers (1). FCEs are also used prior to rehabilitation programs, often in conjunction with other screening tests, to select suitable candidates for rehabilitation, guide treatment planning and determine whether patients are improving with care (2,3). Because of the important consequences of decisions made based on FCE results, these tests must have acceptable measurement properties of reliability and validity (4). Additionally, as FCEs typically take hours to complete often over multiple occasions and cost as much as advanced diagnostic imaging studies, any potential reductions in test burden would be beneficial.

Earlier studies have identified opportunities for greater efficiencies within current FCE protocols and, possibly, greater effectiveness (5). Our research group has studied the reliability and validity of the Isernhagen Work System's FCE (IWS-FCE, Duluth, MN) within samples of workers’ compensation claimants with low back pain (610). The FCE lift and carry tasks were found to have adequate levels of reliability for clinical use (6) and the construct validity of FCE as a test of work-related functional ability was supported (7). However, FCE performance appears to only modestly predict return-to-work or fitness for work, as judged from suspension of time-loss benefits (8). One item from the full FCE protocol related to the matching relationship between subject performance and required job demands, the floor-to-waist lift, alone provided as much predictive value as the entire protocol (8).

This finding related to floor-to-waist lifting is consistent with Matheson et al. who found floor-to-waist lifting to be the only independent predictor when the Isernhagen lift, carry, and handgrip tests were added simultaneously to a multivariable model predicting return-to-work at 6 months in a sample of subjects with unspecified musculoskeletal disorders (11). Additionally, Cutler et al. studied the association between components of a 20-item Dictionary of Occupational Titles FCE and return-to-work following treatment for chronic back pain (12). They found that only a crouching test independently predicted short- and long-term employment outcomes (13). Growing evidence from studies of FCE predictive ability appears to indicate that all or most of the important functional information could be obtained in shorter versions when used for purposes of determining future work status.

Seeing opportunities for potential efficiency improvements, we attempted to empirically derive a short-form FCE while maintaining comparable predictive ability. A short-form FCE would be beneficial for reducing subject and administrative burden duringtesting.

MATERIAL AND METHODS

For the purpose of developing a short-form FCE, we used three databases previously used for evaluating FCE predictive validity in workers’ compensation claimants with low back problems. One database was used for development of the short-form FCE, while the other two were used for validation. The data were gained from claimants undergoing FCE at the major Alberta Workers’ Compensation Board (WCB-Alberta) rehabilitation facility in Edmonton, Canada. The development of these databases, as well as subject characteristics, have been described in detail elsewhere (810,14), but essentially involved merging clinical information from the rehabilitation center with administrative outcomes from the main WCB-Alberta administrative database.

The short-form FCE development database included information from a cohort of subjects seen for full IWS-FCE between January 1999 and December 2000. One validation cohort was made of subjects undergoing the IWS-FCE between April 2001 and March 2002 for purposes of determining fitness for work. The second validation cohort was comprised of subjects undergoing a modified Isernhagen one-day FCE between January 1999 and December 1999 to assist in assessing the fitness for work and guiding rehabilitation. These subjects subsequently underwent a multidisciplinary rehabilitation program at the center. All subjects were workers’ compensation claimants.

Measures

The principle measures used in this evaluation were individual items from the IWS-FCE protocol used at the WCB-Alberta rehabilitation facility (15). For clients with back pain problems, the protocol includes 27 individual activities representing the physical demands of work outlined in the Dictionary of Occupational Titles, such as lifting, carrying, pushing, pulling, and other tasks. Judgments of safe, maximal performance using this protocol appear reliable (6,16,17). There is also support for construct validity (7), and as mentioned previously, the FCE has demonstrated modest predictive value for time to return-to-work or work readiness (11).

At the WCB-Alberta rehabilitation facility, clinicians rate each client on each FCE item according to whether or not performance meets or exceeds required physical job demands. Subjects are, therefore, given a pass/fail rating on each FCE item. The total number of items rated as failed and the pass/fail rating on the floor-to-waist lift have been previously evaluated as prognostic indicators, and found modestly associated with time to benefit suspension and claim closure (8). Pass/fail ratings for each item in the FCE protocol were used in this study.

Outcomes

Days until suspension of time-loss benefits following FCE served as a surrogate indicator of time to return-to-work or work readiness. Although this outcome cannot be equated to full recovery from pain and disability (18), it is considered a critical outcome indicator within many workers’ compensation administrations (8).

Analysis

Initially, factor analysis and Cox proportional hazards regression were used to create the short-form FCE using the development database. Factor analysis was used to determine whether meaningful factors existed within individual FCE item pass/fail ratings. Promax rotation was used as individual items were conceptualized as intercorrelated, because a subject performing highly on one FCE item may be more likely to perform well on other items (19). We used a factor-loading criterion of 0.4 to examine the underlying factor structure.

Next, a series of multivariable Cox regressions predicting days to suspension of time-loss benefits were performed to determine the items independently predictive of this outcome within each identified factor (while controlling for the other items in the factor) and then between the factors (20). Items remaining independently predictive within the final model at a 0.05 alpha level were retained in the short-form FCE. A separate categorical variable was then created from items remaining in the final predictive model, indicating the number of failed items within the short-form FCE. This indicator's predictive ability was evaluated relative to the number of failed tasks in the entire protocol using Cox regression and R2 statistics within the development and validation databases.

RESULTS

Nine FCE factors were identified in factor analysis (Table I) and labeled Lifting, Carrying, Push/Pull, Low-Level/Positional Activity, Climbing, Standing, Sitting, Hand Coordination and Hand Grip. The clinical relationships between items in the factors are evident (i.e. all positional tests involving standing loaded in one factor while positional tests involving sitting loaded separately). Each of the 27 items loaded onto one of the nine factors; no item loaded on multiple factors.

Table I.

Factor Analysis of Functional Capacity Evaluation Items in Patients with Low Back Pain (n=183)

 

Factors

FCE item

1

2

3

4

5

6

7

8

9

Lifting

 Floor-to-waist lifting*

0.79

 0.16

−0.13

−0.07

−0.06

−0.04

−0.04

−0.07

−0.02

 Waist-to-overhead lift

0.56

−0.02

 0.12

 0.01

 0.01

−0.03

−0.05

−0.01

 0.11

 Horizontal lift

0.85

−0.09

−0.03

 0.05

−0.001

−0.03

 0.03

−0.02

−0.03

 Front carry

0.80

−0.13

 0.01

−0.01

−0.10

−0.001

 0.01

 0.001

 0.15

Low-Level/Positional Activity

 Elevated work

 0.01

0.53

−0.06

 0.62

−0.17

−0.02

 0.15

−0.01

 0.05

 Crawl

−0.27

0.67

−0.01

−0.09

−0.05

 0.14

 0.07

 0.08

 0.18

 Kneel

−0.06

0.89

−0.16

−0.004

−0.07

−0.02

 0.11

 0.02

 0.10

 Crouch*

 0.12

0.61

 0.38

−0.04

 0.06

−0.04

−0.03

−0.03

−0.13

 Squatting

 0.14

0.65

 0.21

 0.04

 0.12

−0.02

−0.12

−0.09

−0.19

Climbing

 Stair climbing

−0.06

−0.12

0.84

−0.03

−0.04

 0.08

 0.003

−0.05

 0.07

 Step-ladder climbing*

−0.05

 0.02

0.71

 0.09

−0.07

−0.09

 0.02

 0.12

 0.16

 Balance

−0.03

 0.04

0.82

−0.01

 0.02

−0.001

 0.01

 0.01

−0.02

Standing

 Bending in standing

 0.02

 0.01

−0.05

0.70

−0.07

 0.01

 0.02

 0.07

 0.05

 Rotation in standing*

 0.002

 0.08

−0.05

0.53

−0.06

 0.14

 0.03

 0.19

−0.06

 Standing*

 0.04

 0.02

−0.06

0.83

 0.02

−0.05

 0.04

−0.11

 0.02

 Walking

−0.08

−0.11

 0.19

0.82

 0.05

 0.02

 0.02

−0.04

−0.05

Hand Grip

 Right hand grip

−0.001

−0.08

−0.002

−0.03

0.97

 0.05

−0.01

 0.04

−0.01

 Left hand grip

−0.003

−0.06

−0.05

−0.01

0.95

−0.02

 0.12

 0.01

 0.09

Table I.

Continued

 

Factors

FCE item

1

2

3

4

5

6

7

8

9

Push/Pull

 Static push

 0.03

 0.03

 0.03

 0.02

−0.08

0.94

−0.03

 0.03

−0.01

 Static pull

 0.02

−0.01

 0.05

 0.05

 0.04

0.91

−0.02

−0.07

 0.003

Hand Coordination

 Right hand coordination

 0.05

 0.08

 0.04

 0.02

 0.04

 0.04

0.89

 0.01

−0.12

 Left hand coordination

−0.02

 0.12

−0.03

 0.07

 0.09

−0.09

0.85

−0.05

 0.08

Sitting

 Bending in sitting

 0.04

 0.16

−0.15

 0.13

 0.11

−0.07

−0.15

0.66

−0.003

 Rotation in sitting

−0.02

−0.01

 0.03

 0.06

 0.02

−0.04

−0.17

0.79

 0.08

 Sitting

 0.05

−0.14

 0.15

−0.16

−0.05

 0.05

 0.28

0.77

−0.13

Carrying

 Right carry

 0.08

 0.08

 0.07

−0.02

 0.07

 0.01

−0.07

−0.04

0.89

 Left carry

 0.08

 0.04

 0.05

−0.05

 0.01

 0.01

 0.02

 0.01

0.90

Note. Items within each factor are highlighted in italics. Promax Rotation was used as items were deemed correlated.

*Significant at 0.05 when added with other items within factor to a multivariable Cox regression model predicting time to benefit suspension.

Items within each factor were then added to a Cox regression model to determine which items within the unique factors independently predicted time-to-benefit suspension. Five items (floor-to-waist lifting, rotation in standing, crouching, standing, and ladder climbing) were independently predictive at a 0.05 alpha level (Table I). No FCE items within the Hand Grip, Hand Coordination, Push/Pull, Sitting or Carrying factors were associated with time-to-benefit suspension at a statistically significant level. When the five predictive items from within the various factors were entered simultaneously into a Cox regression, only three remained independently predictive (Table II). These three items were maintained within the short-form FCE and include floor-to-waist lifting, crouching, and standing.
Table II.

Final Developmental Multivariable Prediction Model (n=183)

Failed FCE item

Hazard rate ratio (95% confidence interval)

Floor-to-waist lift

0.64 (0.42–0.97)

Crouching

0.59 (0.38–0.91)

Standing

0.68 (0.48–0.95)

Note. Total explained variation (R2) in days to time-loss benefit suspension = 25%.

Table III.

The Predictive Value of a Short-form FCE within Development and Validation Cohorts

  

Failed items

 

Variation explained (%)

Cohort

n

3

2

1

Passed all

Short-form FCE

Full FCE

Development

183

1.0

1.57 (0.98–2.53)

3.56 (2.17–5.83)*

4.70 (2.70–8.21)*

22

19

Validation

138

1.0

0.98 (0.61–1.57)

1.53 (0.97–2.39)

2.86 (1.60–5.11)*

10

10

Rehabilitation

228

1.0

1.31 (0.95–1.80)

1.43 (1.00–2.07)*

1.89 (1.07–3.32)*

3

5

Note. All values represent hazard rate ratios (95% confidence intervals).

*Significant at 0.05 level.

A separate categorical variable was then created from the three independent predictors, indicating the number of failed FCE items (0–3 failed items). When entered into a Cox regression model within the development database, this variable predicted 3% more of the variation in days to suspension of benefits than the number of failed tasks in the entire protocol according to R2 for Cox regression statistics (Table III). Subjects passing all three items in the short-form protocol returned to work faster than those failing one or more (shown graphically in Fig. 1). The predictive ability of the short-form FCE variable was then tested within the two independent validation datasets and was again a statistically significant predictor and found to predict comparably to the entire FCE protocols used (Table III). While the amount of variation explained within each cohort was variable, likely due to the different context of assessment and normal variation arising between independent cohorts, the predictive abilities of the short-form and entire FCE protocols were consistently comparable. Similar relationships were seen in the cohorts, with subjects passing all three items experiencing faster suspension of benefits (Figs. 2 and 3).
https://static-content.springer.com/image/art%3A10.1007%2Fs10926-005-9008-x/MediaObjects/10926_2005_9008_Fig1_HTML.gif
Fig. 1.

Kaplan-Meier curve of the prognostic value of the short-form FCE within the development cohort (n = 183).

https://static-content.springer.com/image/art%3A10.1007%2Fs10926-005-9008-x/MediaObjects/10926_2005_9008_Fig2_HTML.gif
Fig. 2.

Kaplan-Meier curve of the prognostic value of short-form FCE within the validation cohort (n = 138).

https://static-content.springer.com/image/art%3A10.1007%2Fs10926-005-9008-x/MediaObjects/10926_2005_9008_Fig3_HTML.gif
Fig. 3.

Kaplan-Meier curve of the prognostic value of short-form FCE within the rehabilitation cohort (n = 228).

DISCUSSION

The development and validation of a short-form FCE for purposes of determining future work status in claimants with low back disorders has been described. We found that three items within the broader Isernhagen FCE protocol predict time-to-benefit suspension comparably to information from the full protocol related to the matching relationship between FCE performance and required job demands. The three items within the short-form protocol include clinical ratings of whether subjects met or exceeded required physical job demands during floor-to-waist lifting, a 60-s crouch test on a flat surface, and a 30-min standing tolerance test. While very straightforward and potentially done within less than 1h in the clinic, the three-item FCE has met stringent tests of its predictive ability. The consistency of the short-form FCE's predictive value within three separate cohorts is strongly indicative of its robustness in terms of comparability to the full FCE as apredictor.

Other researchers have independently reported that test items equivalent or comparable to those within the short-form protocol are the most prognostic items within larger batteries. Matheson et al. reported on the prognostic superiority of the floor-to-waist lift test when evaluating the Isernhagen lift, carry and handgrip tests (11), while Cutler et al. reported that a crouching test is the item most predictive of return-to-work within the Dictionary of Occupational Titles FCE (13). Also, the Isernhagen crouch test is similar to an item in the Functional Assessment Screening Test protocol described by Ruan et al. (21) which may possibly explain its importance as a predictor. Ruan et al. reported that subjects who performed poorly on the nonstrenuous items in their assessment (including a crouching/squatting test) had more dysfunctional coping mechanisms, pain avoidance, depression, and self-reported disability. Poor performance on such nonstrenuous tests may be a behavioural manifestation of psychological or cognitive trouble rather than limited physical capacity, and subjects with such problems may be at risk for delayed return-to-work.

While we have presented some empirical data in support of the clinical use of a short-form FCE, the protocol's effectiveness in guiding clinical decisions regarding injured workers and facilitating return-to-work needs to be further examined. It is possible that the items making up the short-form protocol have predictive value only when conducted within the context of the full Isernhagen FCE. For this reason, further research is underway to evaluate the clinical application of the short-form FCE in comparison to the regular Isernhagen protocol in terms of return-to-work outcomes, satisfaction rates and costeffectiveness.

CONCLUSION

A short-form FCE for purposes of determining future work status in claimants with low back disorders was developed and underwent preliminary validation. The new protocol predicted comparably to information from the full FCE protocol from which is was developed, suggesting that the substantially reduced FCE protocol may provide an efficient alternative with comparable predictive ability.

ACKNOWLEDGMENTS

Data collection was facilitated by the Alberta Workers’ Compensation Board/Millard Health.

Copyright information

© Springer Science+Business Media, Inc. 2006