In the early 1990s, a team of British researchers set out to determine the rate of successful rehabilitation among individuals with lower-limb amputations of vascular etiologies. Led by a vascular surgeon named A.D. Houghton, MD, the effort first required the objective determination of what qualifies as a successful rehabilitation. The tool used to make this determination has since been termed the Houghton Scale. While it was initially published almost 25 years ago, recent work suggests that Houghton et al.’s assumptions were reasonable and accurate, and may well have implications in the modern effort to quantify functional outcomes among people with lower-limb amputations.
The premise of Houghton et al.’s 1992 publication was not to develop a functional assessment measure. The pressing concern was to determine the percentage of patients with lowerlimb amputations who would regain successful mobility post-amputation. If, for example, wheelchair mobility was the more likely outcome, should alternative surgical pathways be pursued?
But before questions of that nature could be reasonably answered, the research team would need to somehow quantify the functional outcomes of individuals with lower-limb amputations. The administration of a simple, direct survey was used to measure this. Respondents were asked to gauge how much they used their prostheses during waking hours, the environments in which they used their prostheses, and their preferred means of outdoor ambulation. In addition, they were asked to identify whether or not they felt unstable while ambulating on flat terrain, on slopes, and over rough terrain.1 The answers to these three multiple choice questions and three yes-or-no questions led to a collective score ranging from 0-12 (Table 1). The scale was not named in the original publication, nor would it be referred to as the Houghton Scale in print for another 12 years.
Houghton’s early observations were sobering. Of the 440 patients identified through the medical records of eight participating hospitals over a two-year period (1986-88), 75 (17 percent) died without leaving the hospital and 113 (26 percent) were deemed unsuited for prosthetic rehabilitation. By the time the survey was administered several years later, 54 more patients had died, and 19 patients had stopped wearing their prostheses. Thus, within a few years of their amputations, only 40 percent of these vascular patients were eligible to fill out Houghton’s questionnaire. Of these, 102 responses were collected.1
Apparently arbitrarily, the research team declared that a score of 9 or more on the questionnaire was accepted as satisfactory rehabilitation, and a score between 6 and 8 indicated mobility with the prosthesis around the house. Using these thresholds, Houghton et al. reported that, of the original 440 patients, only 52 scored 6 or higher (12 percent) and only 21 scored 9 or higher (5 percent). When confining consideration to those patients who were referred for prosthetic fitting and still alive at the time of inquiry, 29 percent were considered household ambulators (a score of 6 to 8) and only 12 percent were considered satisfactorily rehabilitated (a score of 9 or higher).1
The relative simplicity of the Houghton Scale found important admirers in a 2004 publication of the Archives of Physical Medicine and Rehabilitation. In it, Devlin et al. suggest, “There is a need for a qualitative tool that will quickly and reliably measure prosthetic use by people with amputations. In a busy clinical practice, such a tool must be one that can be administered in a short period of time or it will not be functionally useful. Such a tool would be used for routine clinical follow-up, program evaluation, and research in which prosthetic use is a variable of interest.”2
The candidate tool under their consideration was none other than the Houghton Scale. However, before fully endorsing this measure, the rehabilitation center set out to confirm its test-retest reliability, examine its responsiveness to change, and ensure it was not overly prone to floor and ceiling effects.2 In short, the test performed quite well.
For example, the average score within a cohort of people with lowerlimb amputations as they were discharged from inpatient rehabilitation was 6 out of 12. When the scale was readministered three months later, this average score increased to just under 8. Both scores suggest that the central tendency of the scale is where it should be, in the middle of the possible scores. Further, of the 76 subjects who participated in the scale validation study, only a single subject scored at the floor (0 points) or ceiling (12 points) of the instrument, suggesting an acceptably low floor and ceiling effect.2
As previously mentioned, a good measurement should be able to document changes when they occur. This was the case with the Houghton Scale, with the average score at discharge of 6.14, increasing to 7.7 when the scale was readministered three months later. Test-retest reliability, assessed at oneweek intervals, was also high, with a correlation coefficient of 0.96.
The possible utility of the Houghton Scale in modern practice environments is further suggested by the work of a team of researchers in New York. They began their investigation under the premise that the cut-off scores originally stated for the Houghton Scale parallel Medicare K-level classifications. More specifically, they argue that a score of 9 or higher on the Houghton scale, declared by its developers as “satisfactory rehabilitation,” corresponds to the K3 functional level, or the ability to walk at variable cadence and negotiate most environmental obstacles for prosthesis use beyond simple locomotion. Similarly, they propose that a Houghton score of 6 or higher, originally declared by its developers as “indicating mobility on the prosthesis around the house,” corresponds to the K2 functional level, or the ability to walk outdoors and negotiate low environmental barriers such as curbs and stairs. Houghton Scale scores below 6 are suggested as corresponding to the K1 functional level, or the ability to walk on level, indoor surfaces.3
The verbiage between the Houghton Scale scores and the K-level classification descriptors pose some variations and discrepancies. However, these subtleties are overshadowed by the larger unknown, namely that the cutoff scores originally assigned to the Houghton Scale appear to be arbitrary. The New York-based research team’s work can be seen as an exercise to determine to what extent the cutoff scores originally assigned to the Houghton Scale successfully discriminate between K1 and K2 and between K2 and K3 functional classifications.3
It would have been interesting to see whether the blindly recorded Houghton Scale scores, collected from a convenience sample of 180 patients with lower-limb amputations from eight states and ten outpatient facilities, had correlated to previously assigned, blindly assessed K-level classification values. Unfortunately, if this data was collected, it was not reported. Rather, Houghton scores were compared to a range of additional outcome measures, including self-reported prosthetic function, performance-based metrics, and walking speed criteria.
Within the realm of self-reported outcomes, subjects were asked to complete the 12-question mobility subscale of the Prosthetic Evaluation Questionnaire (PEQ) and the 16-question Activities-specific Balance Confidence (ABC) Scale. Performance-based metrics included the Timed Up and Go (TUG) test and three tasks from the Berg Balance Scale (BBS) (standing with eyes closed, looking behind over one’s shoulder, and turning a full 360 degrees). Walking speed was calculated based on the observed performance on the two-minute walk test (2MWT).
Houghton Scale scores stratified the convenience sample as follows: 23.5 percent were in the lowest category (suggestive of a K1 classification level); 30.6 percent were scored in the middle category (suggested as a K2 classification level); and 45.9 percent scored in the highest category (posited as the equivalent of a K3 classification level). The expected relationships between Houghton scores and the chosen outcome measures were observed (Table 2).
Viewed collectively, this data confirms what we would reasonably expect—that patients who report higher scores on the Houghton Scale tend to have better mobility, improved balance confidence, better balance, faster transfers, and higher walking velocities than those who have lower scores. If we accept the authors’ supposition that Houghton scores correspond to K-level classifications, then the data becomes more meaningful, suggesting the performance values across a range of tests that equate to a given K-level classification.
The Significance of Correlations…or Lack Thereof
Additional insights are gained when the correlations observed between Houghton scores and the other outcome measures are further considered. The highest correlation was with the ABC Scale, reported at 0.76. This was followed by correlations with PEQ Mobility Subscale and 2MWT values, both reported at 0.73. A lower correlation was observed with the BBS (0.67) with a negative correlation of identical value reported with the TUG test (-0.67).
These values all appear reasonable, but the authors themselves declared at the outset of their study that coefficients were interpreted as moderate (0.5-0.75), good (0.75-0.9), and excellent (0.9 or higher). Against this standard, the correlations between the Houghton Scale and these other outcomes can only be described as moderate.
But if we accept the premise that Houghton cut-off scores correspond reasonably with K-level assignments, should we expect to observe things any differently? This is, after all, the trouble with K-level classification. Many within the rehabilitation and reimbursement communities want this classification to be more objective and quantifiable, but it continues to defy such attempts. Some K3-level walkers who routinely navigate environmental barriers do so with greater confidence and speed than others. Among K2 walkers, some are limited in their ambulation by their walking speed, while others are limited by their balance, and others are limited by their confidence. Similar comments could be made with regard to K1-level walkers. Attempts to assign K-level based on a single objective data point fail to appreciate the collective considerations that ultimately determine the ability and potential of a given patient to ambulate in the community across variable speeds.
Ultimately, objective K-level assignment remains elusive. Given its high clinical utility by virtue of its brevity and simplicity, and its moderate relationships across a host of functional domains, the Houghton Scale may ultimately contribute to the assignment and support of K-level classifications. As Wong et al. state, “Further examination of the parallels between the Houghton Scale and K-level categories may yield a combination of performance-based and self-report measure for evidence-based determinations of K-level….”3 Should this prove to be the case, Mr. Houghton’s opus from 25 years ago will have proven to be a welcome addition to prosthetic rehabilitation.
Phil Stevens, MEd, CPO, FAAOP, is in clinical practice with Hanger Clinic, Salt Lake City. He can be reached at .
- Houghton, A. D., P. R. Taylor, S. Thurlow, E. Rootes, and I. McColl. 1992. Success rates for rehabilitation of vascular amputees: Implications and preoperative assessment and amputation level. The British Journal of Surgery 79 (8):753-5.
- Devlin, M., T. Pauley, K. Head, and S. Garfinkel. 2004. Houghton Scale of prosthetic use in people with lower extremity amputations: Reliability, validity, and responsiveness to change. Archives of Physical Medicine and Rehabilitation 85 (8):1339-44.
- Wong, C. K., W. Gibbs, and E. S. Chen. 2016. Use of the Houghton scale to classify community and household walking ability in people with lower-limb amputation: Criterion-related validity. Archives of Physical Medicine and Rehabilitation 97 (7):1130-6.