The validity of patient and physician estimates of walking distance

Abstract

Objective: To establish the validity of patient and physician estimates of maximum walking distance versus actual measured maximum walking distance.

Design: Assessment of concurrent validity (patient and physician estimates were compared with a gold standard measure at the same time).

Setting: University-affiliated rehabilitation department in a tertiary care hospital.

Subjects: A sequential sample of 31 patients over the age of 17 referred to the physical medicine and rehabilitation outpatients department between 1 May 2000 and 20 July 2001, who had at least some degree of walking difficulty.

Interventions: Patients and their physicians were asked to provide estimates of walking distance independently after a regular appointment prior to the patient being escorted along a pre-measured walking course.

Main outcome measures: Actual distance walked was compared with estimates using Pearson correlation coefficients.

Results: Pearson correlation coefficients for patient estimate versus actual was 0.789 (p ≤ 0.0001), and mean estimate of patient and physician estimates versus actual was 0.812 (p ≤ 0.0001). Physician estimate versus actual and patient estimate versus physician estimate were 0.349 and -0.139 (neither significant). Sixty-seven per cent (20/30) of patients overestimated how far they thought they could walk versus 23% (7.30) of physicians who overestimated. Neither group were found to be ‘good’ estimators of maximal walking distance.

Conclusions: Neither patients nor physicians provide valid estimates of maximal walking distance. Patients consistently overestimate their maximal walking distance, whereas physicians tend to underestimate. Interestingly, patients' estimates (although over inflated) do correlate well with actual walking distance, while physician estimates are not at all correlated. This study suggests that reliance on self-reported or physician-estimated maximum walking distances (whether for clinical, research or other reasons) is potentially flawed.