The physical activity scale for individuals with physical disabilities: development and evaluation

Objective: To evaluate the construct validity of a new 13-item physical activity survey designed to assess physical activity in individuals with physical disabilities.

Design: Mail survey requesting information on physical activity, basic demographic characteristics, self-rated health, and self-rated physical activity.

Setting: In February 2000, surveys were sent to 1176 individuals who had used rehabilitative services at a major midwestern university between 1950 and 1999.

Participants: Two hundred twenty-seven men and 145 women with disabilities responded to the mail survey (80%, spinal cord or other locomotor injuries; 13%, visual and auditory injuries; 7%, other; 92%, white; mean age +/- standard deviation, 49.8 +/- 12.9y; mean length of disability, 36.9 +/- 14.9y).

Interventions: Not applicable.

Main outcome measures: Physical activity was assessed with the Physical Activity Scale for Individuals with Physical Disabilities (PASIPD). The PASIPD requests the number of days a week and hours daily (categories) of participation in recreational, household, and occupational activities over the past 7 days. Total scores were calculated as the average hours daily times a metabolic equivalent value and summed over items.

Results: Pearson correlations between each survey item and the total PASIPD score were all statistically significant (P < .05) and >or= .20 (range, .20- .67). Factor analysis with principal component extraction and varimax orthogonal rotations revealed 5 latent factors (eigenvalues >or= 1, factor loadings >or= .40): home repair and lawn and garden, housework, vigorous sport and recreation, light sport and recreation, and occupation and transportation. These 5 factors accounted for 63% of the total variance. Cronbach alpha coefficients ranged from.37 to.65, indicating low-to-moderate internal consistency within factors. Those who reported being "active/highly active" had higher total and subcategory scores compared with those "not active at all." Those in "excellent" health had higher total, vigorous sport and recreation, and occupation and transportation subcategory scores compared with those who rated their health "fair/poor" (all P < .05).

Conclusion: These results provide preliminary support for the construct validity of the PASIPD. Additional validation studies using an external criterion and in more generalizable samples are warranted.