Physical Functioning Scale of the ShortForm (SF) 36: Internal Consistency and Validity With Older Adults

Bohannon, Richard W. PT, DPT, EdD, NCS, FAPTA, FAHA; DePasquale, Louis PT, MA

Journal of Geriatric Physical Therapy:
doi: 10.1097/JPT.0b013e3181d0735e
Research Reports

Background and Purpose: There is a need to characterize the physical functioning of older adults. As the Physical Functioning Scale (PFS) of the 36-Item Short-Form Health Survey has not been utilized widely by therapists for this purpose, we sought to investigate its reliability and validity in this population.

Methods: This study involved the secondary analysis of PFS, physical performance, and fall history data obtained from 19 men and 39 women, 65 to 94 years old.

Results: Reliability (internal consistency) of the PFS was supported by a Cronbach α of .82 between its items. Known groups' validity was demonstrated by a significant difference in PFS scores of individuals with and individuals without a history of falls and by an area under the receiver operating characteristic curve of 0.850. Convergent validity was upheld by significant Pearson correlations between the PFS and single limb stance time (r = 0.42), Timed Up and Go test (r = −0.70) performance, and gait speed (r = 0.75).

Conclusions: The simplicity and brevity of the PFS as well as its internal consistency and known groups and convergent validity provide support for its broader use by physical therapists among older community-dwelling adults.

Author Information

1Program in Physical Therapy, Department of Kinesiology, University of Connecticut, Storrs.

2Visiting Nurse Service of New York, New York.

Address correspondence to: Richard W. Bohannon, PT, DPT, EdD, NCS, FAPTA, FAHA, Department of Physical Therapy, University of Connecticut, U-2101, Storrs, CT 06269 (

Article Outline
Back to Top | Article Outline


Activity limitations are a chief target of physical therapy interventions for a variety of pathologies, disorders, and conditions. The limitations can be characterized by using physical performance measures or self-report surveys. There are numerous such surveys available, one of which is the Physical Functioning Scale (PFS) of the 36-Item Short Form Health Survey (SF-36).1 The PFS is composed of 10 items encompassing a hierarchical range of difficulties (Table 1). Each item is scored on the basis of the limitations perceived by surveyed individuals. Item scores (1, 2, or 3) are summed to obtain a total score, which can then be scaled relative to its range. Normative scale values are available for the PFS for the US population and for specific subpopulations (eg, women aged 55–64 years, adults with type II diabetes).1 The PFS has been reported to have good internal consistency (Cronbach α ≥ .81).1 The PFS has been used as a “stand-alone” instrument in research describing activity limitations among diverse groups, including older adults living in the community2,3 and patients with stroke,4,5 multiple sclerosis,5 amyotrophic lateral sclerosis,5 cancer,6 human immunodeficiency virus infection,7 peripheral arterial disease,8 hip fracture,9 joint replacement,10 and chronic fatigue syndrome.11

Its use in research, simplicity, and parsimony notwithstanding, our experience does not reveal the PFS to be widely employed by physical therapists working with older adults who are likely to have reduced balance and be at risk for falls. If it is to be advocated for use with such individuals, it must be shown to be clinimetrically sound. We therefore sought to determine 2 clinimetric qualities of the PFS, that is, reliability and validity. The aspect of reliability investigated was internal consistency. The aspects of validity examined were known groups and convergent validity.

Back to Top | Article Outline


This descriptive investigation involved the secondary analysis of data from a study of the Spring Scale Test.12 The study was approved by the institutional review board of the Visiting Nurse Service of New York Center for Home Care Policy and Research.

Back to Top | Article Outline

All subjects were independently ambulatory, community-dwelling adults who provided written consent prior to participation. Nineteen were men and 39 were women. They ranged in age from 65 to 94 years, mean age 80.8 (SD = 7.2). By self-report, 29 of the 58 subjects had fallen at least once over the previous 2 years.

Back to Top | Article Outline

A written copy of the PFS of the SF-36 (version 1.0) was given to each subject to complete. All surveys were completed without omissions. Scores (1 = yes, limited a lot; 2 = yes, limited a little; or 3 = no, not limited at all) for each of the 10 items were recorded. Thereafter, total and scale scores were determined for each subject.

To enable the examination of convergent validity, 3 performance-based measures were obtained. Single limb stance (SLS) time was used as an indicator of balance.13 The eyes were open; the preferred limb was used. Timing began upon withdrawal of support and continued until the uplifted foot returned to the floor, the position of the weight-bearing foot changed, or 30 seconds elapsed. The best of 2 trials was recorded. Gait speed was determined by timing subjects as they completed a single walk over 3 m at their normal comfortable speed.14 They were provided with acceleration and deceleration zones before and after the 3 m. The Timed Up and Go test (TUG) was employed as a measure of overall mobility.15 Subjects stood from an armchair, walked 3 m, turned, walked back to the chair, and sat down. They performed 2 trials at a quick but safe speed. Timing began on the command go and ceased when they returned to a seated position. The best of 2 trials was recorded.

Back to Top | Article Outline
Statistical Analysis

Conventional descriptive statistics were used to provide an overall summary of the PFS scores and the performance measures. Descriptive statistics were also used to summarize the scores and measures for fallers and nonfallers and for comparable individuals for whom reference values were available.1,13–15 The Cronbach α was used to examine the internal consistency of PFS scores. Known groups' validity was determined by using a t test and receiver operating characteristic (ROC) curve to compare the PFS scale scores of fallers and nonfallers. Pearson product moment correlations were calculated to describe the convergent validity of the PFS scores with SLS, gait speed, and TUG performance.

Back to Top | Article Outline


Table 2 provides a summary of the PFS scores and other measures obtained in this study as well as reference measures from potentially representative populations. The items of the PFS had good internal consistency (Cronbach α = .82).

The PFS scores of participants who fell (mean = 63.3) and participants who did not fall (mean = 86.2) were significantly different (t = −5.56, P < .001). The area under the ROC curve was 0.850, with a PFS scale score cut point of 72.5 (sensitivity 0.931, specificity 0.655). Significant relationships (P < .001) were found between the PFS and other variables. The correlations ranged from 0.42 (SLS time) to 0.75 (gait velocity).

Back to Top | Article Outline


The PFS, which all subjects were able to complete without assistance, was shown to be both reliable and valid. Reliability was supported by a Cronbach α value of .82, a value almost identical to the one reported elsewhere (.81).1 Granting that internal consistency does not inform as to test-retest reliability or responsiveness,16 it is an important aspect of reliability. Both known groups and convergent validity were supported. Known groups' validity was upheld by the results of the t test and ROC curve analysis. The area under the ROC curve (0.850) was greater for the PFS than for 3 performance measures (SLS, tandem stance, and gait speed) that were examined in a previous study.12 However, the area under the ROC curve was less for the PFS than for 2 of the performance measures (Spring Scale Test and TUG) in that study.12 The convergent validity of the PFS, a self-report measure, was confirmed by its significant correlations with 3 physical performance measures. It is interesting that the correlations of the PFS, which includes a preponderance of mobility items, were highest with performance tests of mobility (gait speed and TUG) (Table 3).

The results of this study need to be interpreted in light of several limitations. Chief among these is the representativeness of the sample. A convenience sample was used. In spite of its inclusion of participants who fall and participants who do not fall, the sample's PFS scores and physical performance scores show it to be higher functioning than would be expected for comparably aged older adults of the general population. The utility of the PFS, particularly among older adults with impaired cognition, might be limited. Moreover, its use among older adults not living freely within the community (eg, nursing home residents) may not be warranted.

Back to Top | Article Outline


The PFS is reliable and valid among community-dwelling older adults. It may warrant broader application by physical therapists.

Back to Top | Article Outline


1. Ware JE, Snow KK, Kosinski M, Gandek B. SF-36 Health Survey: Manual and Interpretation Guide. Boston, MA: The Health Institute, New England Medical Center; 1989.
2. Bohannon RW, Brennan PJ, Pescatello LS, Marschke L, Hasson S, Murphy M. Adiposity of elderly women and its relationship with self-reported and observed physical performance. J Geriatr Phys Ther. 2005;28:10–13.
3. Rigler SK, Studenski S, Wallace D, Reker DM, Duncan PW. Co-morbidity adjustment for functional outcomes in community-dwelling older adults. Clin Rehabil. 2002;16:420–428.
4. Lai S-M, Duncan PW, Keighley J. Prediction of functional outcome after stroke: comparison of the Orpington Prognostic Scale and the NIH Stroke Scale. Stroke. 1998;29:1838–1842.
5. Dallmeijer AJ, de Groot V, Roorda LD, et al. Cross-diagnostic validity of the SF-36 physical functioning scale in patients with stroke, multiple sclerosis and amyotrophic lateral sclerosis: a study using Rasch analysis. J Rehabil Med. 2007;39:163–169.
6. Given CW, Given B, Azzouz F, Stommel M, Kozachik S. Comparison of changes in physical functioning in elderly patients with new diagnoses of cancer. Med Care. 2000;38:482–493.
7. Roubenoff R, Wilson IB. Effect of resistance training on self-reported physical functioning in HIV infection. Med Sci Sports Exerc. 2001;33:1811–1817.
8. McDermott MM, Guralnik JM, Greenland P, et al. Inflammatory and thrombic blood markers and walking-related disability in men and women with and without peripheral arterial disease. J Am Geriatr Soc. 2004;52:1888–1894.
9. Latham NK, Mehta V, Nguyen AM, et al. Performance-based or self-report measures of physical function: which should be used in clinical trials of hip fracture patients? Arch Phys Med Rehabil. 2008;89:2146–2155.
10. Fortin PR, Penrod JR, Clarke AE, et al. Timing of total joint replacement affects clinical outcomes among patients with osteoarthritis of the hip or knee. Arthritis Rheum. 2002;46:3327–3330.
11. Powell P, Bentall RP, Nye FJ, Edwards RHT. Randomized controlled trial of patient education to encourage graded exercise in chronic fatigue syndrome. BMJ. 2001;322:1–5.
12. DePasquale L, Toscano L. The Spring Scale Test: a reliable and valid tool for explaining fall history. J Geriatr Phys Ther. 2009;32:159–167.
13. Bohannon RW. Single limb stance times: a descriptive meta-analysis of data from individuals at least 60 years of age. Top Geriatr Rehabil. 2006;22:70–77.
14. Bohannon RW. Population representative gait speed and its determinants. J Geriatr Phys Ther. 2008;31:49–52.
15. Bohannon RW. Reference values for the Timed Up and Go test: a descriptive meta-analysis. J Geriatr Phys Ther. 2006;29:64–68.
16. Puhan MA, Bryant D, Guyatt GH, Heels-Ansdell D, Schunemann HJ. Internal consistency reliability is a poor predictor of responsiveness. Health Qual Life Outcomes. 2005;3:33.

geriatrics; measurement; physical performance

© 2010 Academy of Geriatric Physical Therapy, APTA