Reliability of distance estimation by doctors and patients: cross sectional studyBMJ 1997; 315 doi: https://doi.org/10.1136/bmj.315.7123.1652 (Published 20 December 1997) Cite this as: BMJ 1997;315:1652
- a Department of Neurology, United Medical and Dental Schools of Guy's and St Thomas's Hospital, Guy's Hospital, London SE1 9RT
- Correspondence to: Professor Hughes
Objective: To assess the reliability and accuracy of distance estimated by doctors and patients.
Design: Comparison between estimated and measured distances of six familiar sites around Guy's Hospital, London.
Subjects: 100 hospital consultants and 100 patients.
Main outcome measures: Median (range) of estimated distances, and mean (SD) of the difference between estimated and measured distances.
Results: Both doctors and patients gave a wide range of estimates of distance. The estimates differed by up to 14.6-fold from the measured distances, and the difference between minimum and maximum estimates was up to 62.5-fold.
Conclusion: Doctors and patients were inaccurate at estimating distances, which implies that estimates of distances walked are not reliable indicators of a person's health.
Doctors and patients are inaccurate at estimating distances
Estimates of distance can differ by up to 14.6-fold from measured distances, and the difference between the minimum and maximum estimates can be up to 62.5-fold
When estimates are expressed as a percentage of the measured distances, estimates for shorter distances are more inaccurate than those of longer distances
Decisions regarding health based on estimates of distance are unreliable
The economic implication of estimating distance is considerable
The assessment of a patient's walking ability is a simple and practical method of evaluating the state of respiratory, cardiovascular, peripheral vascular, and neurological disease.1 2 Such assessment correlates well with more sophisticated assessments of cardiorespiratory function or muscle strength3 and is important in assigning scores in many clinical disability rating scales—for example, Kurtzke's expanded disability status scale for multiple sclerosis.4
The two most common methods for patient assessment are the maximum distance a patient can walk or the distance they can walk until the onset of symptoms. These distances are infrequently measured in clinical practice. Doctors have traditionally relied on their own or patients' estimates of the distance walked around familiar places. One study assessing the accuracy of trained and untrained artillery observers in estimating target distances ranging from 600 m to 1550 m showed wide variability.5 To our knowledge, there are no published studies assessing the accuracy of distance estimates made by doctors and patients.
Subjects and methods
We sent a questionnaire to all 198 consultants in our hospital asking them to estimate (in yards or metres) the dimensions of a hospital ward and the distances between five familiar sites at the hospital. A category for don't know was provided to prevent guessing. One hundred and five (53%) questionnaires were returned, of which 100 were completed. The same questionnaire was given to 100 consecutive adult patients from a general medical and neurological ward and a neurology outpatient clinic. None of the consultants or patients had an overt psychiatric disorder or cognitive dysfunction. All study sites were later measured with an architect's tape measure in metres.
The consultants were more familiar with the hospital sites than the patients. The number of consultants giving estimates for the six distances varied between 45 and 97 and the number of patients between 10 and 62 (1). Both consultants and patients inaccurately estimated the distances. Their mean estimates correlated moderately with the measured distances (r=0.73 and 0.56 respectively), and the range of estimates was wide and generally greater for consultants. The estimates for the whole group differed by up to 14.6-fold from the measured distances, and the difference between minimum and maximum estimates was up to 62.5-fold. This variability was partly due to the presence of a few outliers (1) since the differences between the measured distances and the median estimates of both groups were small.
When estimates were expressed as a percentage of the measured distances the estimates for shorter distances were more inaccurate than those of the longer distances. The patients' mean estimate of a ward 6.6 m wide was 17.4 m, an error of 163.9%, while the consultants' mean estimate of the same ward was 10.1 m, an error of 52.8%. On the other hand, the patients' mean estimate of a 319.1 m walk to the local station was 452.6 m, a 41.8% error, while the consultants' mean estimate of the same distance was 339.2 m, a 6.3% error. However the differences between the mean estimates of both groups were not significant, with the exception of the 319.1 m distance (1).
This study suggests that people are inaccurate at estimating distances and that medical education is no safeguard. The range of estimates was wide suggesting that any decisions about health based on estimating distance are unreliable. Although estimates were proportional to the distance measured—indicating that consultants and patients were capable of comparing distances and therefore possibly able to estimate changes—the potential value of this observation needs to be evaluated against the test-retest variability of these estimates. The comparable inaccuracy of estimates for both groups suggests that selecting mainly patients with neurological disease (90%) did not bias results or limit their generalisability to other patient groups. Participants are also unlikely to have deliberately provided spurious estimates since the outlier values were from participants who had given more reasonable estimates. As both groups comprised adults the results cannot be generalised to other ages.
Clinical assessments and therapeutic decisions are often based on estimates of distance—for example, the severity of angina, claudication, and chronic respiratory failure—and the effect of treatment on these conditions is usually assessed by the distance a patient can walk before the onset of symptoms. The results of some of the most hailed clinical trials in multiple sclerosis have been based on clinical disability scales related to ambulation, including the expanded disability status scale.6 7 A 1.0 step change on this 20 grade scale is regarded as an important change, although the difference between grades 5.5, 5.0, 4.5, and 4.0 is the ability to walk 100, 200, 300, or 500 metres respectively.2 Such distances are usually estimated and infrequently measured in hospital wards and outpatient clinics. The economic implication of distance estimation is considerable as over 1.2 million people in the United Kingdom are in receipt of the higher rate component of the disability living allowance (currently £33.90 ($54.20) a week) (Department of Social Security, personal communication). This allowance is paid to five categories of patients, including people who have difficulty in walking.8 A person is eligible for this allowance on the basis of a self assessment questionnaire.9
We thank the patients and consultants who participated, and Noor and Sana Sharrack for their help in measuring some of the study distances.
Conflict of interest: None.