A comparative analysis of several crutch-length-estimation techniques.

Correct crutch length is necessary to prevent injury and minimize energy expenditure during crutch ambulation. Patients may obtain crutches from physicians, pharmacists, emergency departments, physical therapists, family members, friends, and even garage sales. Crutch fitting is often done on a trial-and-error basis utilizing many different methods. A study of crutch-length-estimation techniques predictive of ideal crutch length would be clinically relevant. Use of such techniques would ensure proper fit, save valuable clinical time, and, by enabling the health care provider to fit the patient prior to standing, reduce the amount of repositioning required with an acutely injured patient.

For each crutch user, there is "one ... and only one" correct crutch length. [1] This statement is an accepted axiom, but methods for the estimation of crutch length vary widely. Proper crutch length is essential for safe and efficient crutch use. [2] Crutches that are too long may cause brachial plexus injuries [3,4] or axillary artery damage. [5] Crutches that are too long also decrease the patient's ability to handle and use them safely. [6,7] Finally, crutches that are too short cause the patient to assume a poor posture. [2,7,8]

Crutch-length--estimation techniques have been developed to decrease the amount of time the therapist spends determining the correct crutch length for a patient. [9,11] These estimation methods also allow bedridden or acutely injured patients to be fitted for crutches while positioned supine or seated. The crutch-length--estimation procedure decreases patient anxiety because the crutches are fitted before the patient stands for the first time. [12]

Reliability and validity of most esimation techniques are questionable and have not been scientifically evaluated. Clinically, crutch lengths obtained by estimation methods often require readjustments after the patient stands; therefore, determination of the method that best approximates ideal crutch fit would be the most useful. [10]

Crutch-length--estimation methods vary widely. In 1920, a large number of crutch dealers recommended that crutch length be equal to 77% of the patient's height. [9] Other methods that have been suggested over the years include the following: (1) measuring the distance from the patient's axilla to the heel and adding 4 in (10.2 cm), [7,13] (2) measuring from the axilla to the floor and adding 2 in (5.1 cm), [1,14] (3)0 measuring from the anterior axillary fold to a point 4 to 8 in (10.2-20.3 cm) lateral to the heel. [1, 7,15,16] (4) subtracting 16 to 18 in (40.6-45.7 cm) from the patient's height, [13,17] (5) measuring from the axilla to the heel of the shoe with the patient positioned supine, [7] and (6) measuring from the olecranon process of one ulna to the tip of the middle finger of the opposite hand, as described by Najdeski. [12]

A frequently used criterion for correct crutch length, with the patient in the relaxed upright posture, is that there be 1.5 to 2 in (3.8-5.1 cm) between the anterior axillary fold and the axillary pad when the crutch tips are 4 to 6 in (10.2-15.2 cm) anterior and lateral to the toes. [2] Our clinical experience suggests that an increase in axillary distance to 2.5 in (6.4 cm) is better tolerated.

In 1965, Beckwith [10] compared the "77%-of-height" method and methods 2 through 5 listed earlier. Fifteen subjects participated in Beckwith's study, and each subject measured each of the other 14 subjects using all five methods to obtain a total of 210 measurements for each method. These measurements were compared with the clinically determined ideal crutch length, and the mean differences were calculated. Beckwith concluded that height minus 16 in (40.6 cm) yielded the least difference from ideal height and that the 77%-of-height method was the next most accurate method. The research design used in Beckwith's [10] study had several weaknesses. First, although 210 measurements were obtained for each method, the sample consisted of only 15 subjects. Second, several of the measurers made calculation errors when computing 77% of height and height minus 16 in (40.6 cm). These errors were discussed, but they were left uncorrected because of the exploratory nature of the study.

A quick and easy method to assess the ideal crutch length for any patient is needed. To date, virtually all methods have been based on empirical suggestions from the field of practitioners and have not been confirmed statistically. The purpose of this study, therefore, was to determine which of several crutch-fitting techniques best predicts the appropriate crutch length.

In this study, we estimated subjects' ideal crutch length using each of the following methods: (1) 77% of height, (2) height minus 16 in (40.6 cm), (3) olecranon of one arm to tip of middle finger of the opposite hand, (4) olecranon of one arm to tip of little finger of the opposite hand, (5) 77% of arm span, (6) arm span minus 16 in (40.6 cm), and (7) axillary fold to heel. Calculations for estimation methods involving height were performed twice: once with heights reported by the subjects and once with actual heights measured by the investigators. The standard against which the various methods were compared was the ideal crutch length as determined by an experienced orthopedic physical therapist (CJB).

Method

Subjects

One-hundred seven active-duty military volunteers (57 male, 50 female) from Fort Sam Houston, Tex, participated in this study (Tab. 1). The subjects' ages ranged from 19 to 44 years. Male subjects had a mean age of 25.4 years, with a standard deviation of 4.5 years. Female subjects had a mean age of 28.4 years, with a standard deviation of 6.3 years. Heights ranged from 64 in (162.6 cm) to 76 in (193.1 cm), with an overall standard deviation of 7.4 in (18.7 cm) for self-reported height and for actual measured height. All subjects were informed of the experimental protocol, and each subject gave written consent before participating in the study. Individuals were excluded from the study if they did not meet the height requirements for the adult's standard wooden crutch (5 ft 4 in-6 ft 4 in [1.5 m 10.2 cm-1.8 m 10.2 cm]).

Equipment

All subjects were fitted with a pair of adult's adjustable, varnished, wooden axillary crutches, (*) which were equipped with rubber crutch tips and axillary pads. A standard physical examination scale was used to measure height. Other measurements were taken using cloth measuring tapes. Linear measurements were taken in inches, rather than centimeters, because our standard axillary crutches continue to be manufactued with holes in 1-in increments.

Procedure

During the study, the subjects passed through four stations. At the first station, the subjects were asked their height and then their actual height was measured. Height was measured with the subjects wearing footgear, usually military boots or tennis shoes, on a standard height-weight scale to the nearest 0.5 in (1.3 cm). At the second station, subjects were positioned against a tape measure running horizontally across the wall. The lengths from the subjects' olecranon to the middle finger and from the olecranon to the fifth finger were recorded (Fig. 1). While remaining in this position, the subject's arm span was measured (ie, from the tip of the middle finger of the subject's right hand to the tip of the middle finger of the other hand) (Fig. 2). At the third station, the subjects assumed a supine position and the distances from the axillary fold to the heel of the shoe were measured (Fig. 3). At the last station, an ideal crutch length value for each individual in the stationary position was determined by an orthopedic physical therapist with 12 years of clinical experience. Subjects were asked to stand with their feet a comfortable distance apart and their shoulders relaxed. The crutch tips were placed 6 in (15.2 cm) laterally and 6 in (15.2 in) anteriorly from the fifth toe using a measured template, and the top of the foam-rubber axillary pad was placed 2.5 in (6.4 cm) below the axillary fold (Fig. 4).

Subjects were permitted to leave at this point, and values for the remaining techniques were determined. The subjects' self-reported and actual heights recorded at station 1 and the

[TABULAR DATA OMITTED]

arm span recorded at station 2 were used in the calculations. Each measurement was then multiplied by 77%, and 16 in (40.6 cm) was subtracted from each measurement.

Data Analysis

Descriptive statistics in the form of means and standard deviations were obtained for all crutch measurement variables. Next, Pearson Product-Moment Correlation Coefficients (r) among the ideal crutch length and the other measures were calculated to determine the degree to which each of the variables was associated with the ideal length criterion. Linear regression equations were developed for the best predictive variables. All crutch-length--estimation techniques and the regression equations were then assessed for accuracy against the ideal crutch length. Mean squared error (MSE) indices were computed for all techniques as the average of the squared deviations, or errors, between the technique measurements and the ideal crutch lengths. [18] The MSE method was chosen over a simple difference method because in squaring errors, more weight was given to large errors.

Results

Descriptive statistics for subjects and measurements are reported in Table 1. The average ideal crutch length was determined to be 51.9 in (131.7 cm). There was almost a 0.5-in (1.3-cm) difference between the average actual height (69.3 in [176.0 cm]) and the average self-reported height (68.8 in [174.8 cm]). Means and standard deviations for the arm-span, axilla, and olecranon-to-opposite-finger measures are also presented in Table 1. Treating actual height and the other four physical measures as items, the internal consistency index for measurement reliability had an alpha coefficient of .96 Table 2 contains the resultant Pearson correlation coefficients and linear regression values. Actual height had the strongest relationship with ideal crutch length (r=.955, P,.01). Self-reported height also had a high degree of association wit, ideal crutch length (r=.9 P<.01). Both actual and self-reported height variations, therefore, each accounted for about 90% (100 X [r.sup.2]) of the variance in the ideal crutch length values. The remainder of the linear regression correlations, although statistically significant, were weaker and ranged from .60 to .84.

Two simple linear regression equations were computed by regression of the ideal crutch length values on each of the height variables. The resultant slope and y-intercept for actual height were .719 and 2 in (5.2 cm), respectively. Slope and y-intercept for self-reported height were .680 and 4.8 in (12.3 cm), respectively.

To compare regression equation predictions with the other techniques, an MSE analysis was conducted. Comparative results are listed in Table 3. As expected, height equation results produced the least amount of error because of the nature of the least-squares solutions. Comparisons among the other techniques showed that the MSE value obtained by taking 77% of the self-reported height had the smallest amount of error, but was still 1.40 units greater than the self-reported height equation estimates and 1.47 units greater than the actual height equation estimates. Of the experimental crutch-length--estimation techniques, 77% of self-reported height had the least MSE (1.90) and

Table 2. Pearson Product-Moment Correlation Coefficients (r) and Linear Regression Correlation Coefficients ([r.sup.2]) Correlated with Ideal Crutch Length (a)

                             r         [r.sup.2] (b)
Actual height                .955      .91
Self-reported height         .947      .90
Axilla to heel               .919      .84
Arm span                     .821      .67
Elbow to tip of 3rd finger   .777      .60
Elbow to tip of 5th finger   .774      .60
  (a)  Ideal crutch length is defined as the length of
the crutch, including accessories, obtained
during stance when the crutch tip is 6 in (15.2
cm) lateral and 6 in (15.2 cm) anterior to the
fifth toe and the axillary pad is 2.5 in (6.4 cm)
below the axillary fold.
  (b)  All correlations were statistically significant
(P<.01).

[TABULAR DATA OMITTED]

axillary fold to heel had the greatest MSE (14.09).

Discussion

The purpose of this study was to determine which of several crutch-length--estimation techniques most accurately predicts ideal crutch length. That is, is there an acceptable alternative to trial-and-error crutch fitting? In reviewing the literature, it was evident that crutch-length-estimation techniques vary widely. Virtually all techniques derived by clinical experience have not been scientifically validated; this study provides statistical insight into crutch-length-fitting techniques. Results supported our alternate hypothesis that one or more of the estimation techniques would be significantly predictive of ideal crutch length. Specifically, fairly strong evidence for use of the height measures emerged, whereas evidence for the use of elbow, arm-span, and axilla measures was far weaker.

Our findings appear to be consistent with work conducted previously. In 1920, a crutch dealer found 77% of the patient's height to be the best predictor for crutch length. [9] Beckwith, [10] in 1965, concluded that the height-minus-16 in (40.6 cm) method yielded the least difference from ideal crutch length and that the 77%-of-height method was the next most accurate method. Beckwith's study also revealed the most accurate methods to be those involving a mathematical manipulation of height, specifically, self-reported height among this population.

In addition, through regression analysis of the data, equations were derived for crutch-length estimation. The equation having the least error multiplies the subject's actual height by 72% and then adds 2 in (5.2 cm). The second most statistically accurate method, also calculated by regression analysis, multiplies the subject's self-reported height b7 68% and adds 4.8 in (12.3 cm). These two equations are the most accurate, because, with the use of linear regression analysis, one cannot obtain a better-fitting model than a least-squares solution. These equations can allow a health care provider to fit a patient for crutches before

Table 4. Calculated Ideal Crutch Lengths (in Inches) (a)

                                Self-Reported
Actual Height   Crutch Length   Height          Crutch Length
64              48.0            64              48.5
65              48.5            65              49.0
66              49.5            66            49.5
67              50.0            67              50.5
68              51.0            68              51.0
69              51.5            69              51.5
70              52.5            70              52.5
71              53.0            71              53.0
72              54.0            72              54.0
73              54.5            73              54.5
74              55.0            74              55.0
75              56.0            75              56.0
76              56.5            76              56.5
  (a)  Crutch lenght includes crutch tip and axillary
pad.  Formulas: 0.719 X actual height + 2 in;
0.683 X self-reported height +4.8 in.

the patient has to stand, and without the need to take any measurements when self-reported height is used. To simplify application, the equations were applied to the range of possible patient height requirements for the adult's standard wooden crutch. Table 4 shows the crutch lengths that correspond to various heights using the two equations. The values presented in Tables 4 and 5 are easily applied in the clinical setting.

Of the methods tested, it is interesting that the two self-reported height techniques were more accurate than the two actual height measures. Clinically, a patient's height is most often determined by verbal report, rather than a measurement on a scale, as was performed in our study. The height is usually obtained by verbal report because of time constraints in a clinical setting. Patients are also often in bed or disabled by an injury. In these cases, an actual height measurement with a patient standing on a scale would be impractical or impossible. It is important to note that in a military population, most subjects have their height taken at least biannually for physical training tests; thus, their self-reported heights may be closer to their actual heights than one might find in the general population.

When determining actual height, subjects were measured while wearing footgear (ie, boots or tennis shoes). Their heights were taken in this manner because, pragmatically, patients will be wearing footgear while using crutches.

The axillary-fold-to-heel method was found to be the least accurate crutch-length-estimation technique. This method is often recommended for a patient lying in bed. As we have shown, a more accurate measurement could be obtained by simple application of the calculated ideal crutch length (Tab. 4).

Because there is little published literature comparing crutch-length-estimation methods with ideal crutch length, this study suggests the need for further investigation. One of the variables in this study that needs to be addressed by further research is the determination of ideal crutch length. The ideal crutch length was determined by one experienced physical therapist (CJB). This procedure was based on the assumption that every patient has an ideal crutch length. By using a physical therapist with many years of experience at fitting crutches, we assumed that assigned ideal crutch lengths would be expected to correspond to other clinicians' determinations of ideal crutch length. If this study were performed in a clinical setting, in which four or five clinicians separately determined ideal crutch lengths and the resultant measurements were then averaged, a more reliable and stable estimate of ideal crutch length would be achieved.

Possibilities for further research include comparing crutch length determined by the derived equations with the height of present metal crutches and comparing ideal crutch lengths as related to age, sex, race, specific disabilities, or other demographic variables. Clinical surveys as to how crutches are currently fitted and surveys as to how crutch fitting is taught in various physical therapy and other health care schools would also be pertinent.

Conclusion

This study produced two equations for accurate estimation of ideal crutch length. Previous estimation techniques were ranked from the most accurate to the least accurate as follows:

(1) reported heightx0.77, (2) self-reported height --16 in (40.6 cm), (3) actual heightx0.77, (4) actual height--16 in (40.6 cm), (5) olecranon to tip of fifth finger, (6) olecranon to tip of third finger, (7) arm spanx0.77, (8) arm span--16 in (40.6 cm), and (9) axillary fold to heel. The most accurate crutch-measurement techniques have been determined to ensure proper fit, to reduce the time a therapist spends determining ideal crutch length, to allow a bedridden or acutely injured patient to be fitted for crutches while seated or positioned supine, and to prevent further disabilities that could result from improperly fitted crutches.

Table 5. Calculated Ideal Crutch Lengths (in Centimeters) (a)

Actual   Crutch   Self-Reported   Crutch
Height   Length   Height          Length
163      122.4    163             123.6
164      123.1    164             124.3
165      123.8    165             125.0
166      124.5    166             125.7
167      125.3    167             126.4
168      126.0    168             127.0
169      126.7    169             127.7
170      127.4    170             128.4
171      128.1    171             129.1
172      128.9    172             129.8
173      129.6    173             130.4
174      130.3    174             131.1
175      131.0    175             131.8
176      131.7    176             132.5
177      132.5    177             133.2
178      133.2    178             133.9
179      133.9    179             134.5
180      134.6    180             135.2
181      135.3    181             135.9
182      136.0    182             136.6
183      136.8    183             137.3
184      137.5    184             138.0
185      138.2    185             138.6
186      138.9    186             139.3
187      139.6    187             140.0
188      140.4    188             140.7
189      141.1    189             141.4
190      141.8    190             142.1
191      142.5    191             142.7
192      143.2    192             143.4
193      144.0    193             144.1
  (a)  Crutch length includes crutch tip and axillary pad.  Metric formulas:
0.719 X actual height + 5.2 cm;
0.683 X self-reported height + 12.3 cm.

(*) LAMCO, 269 Ludlow St,Stamford, CT 06902.

References

[1] Cicenia EF, Holberman M. Crutches and crutch management. Am J Phys Med. 1957;36:359-383.

[2] Lowman EW, Rush HA. Self-help devices, crutch prescription: measurement. Postgrad Med. 1962;31:303-305.

[3] Rudin LN, Levine L. Bilateral compression of the radial nerve. Phys Ther Rev. 1951;31:229-231.

[4] Shabes D, Scheiber M. Suprascapular neuropathy related to the use of crutches. Am J Phys Med. 1986;65:298-299.

[5] Dounis E. Rose GK, Wilson RS, et al. A comparison of efficiency of three types of crutches using oxygen consumption. Rheumatol Rehabil. 1980;19:252-255.

[6] Cohen S, Viellion G. Teaching a patient how to use crutches. Am J Nurs. 1979;79:1111-1126.

[7] Deaver CG. What every physician should know about the teaching of crutch walking. JAMA. 1950;142:470-472.

[8] Knocke L. Crutch walking. Am J Nurs. 1961;61:70-73.

[9] Determining the length of crutches. Mod Hosp. 1920;15:332.

[10] Beckwith JM. Analysis of methods of teaching axillary crutch measurement. Phys Ther. 1965;45:1060-1065.

[11] Davenport J. Improved method for fitting crutches and canes. Phys Ther Rev. 1960;40:591.

[12] Najdeski P. Crutch measurement from the sitting position. Phys Ther. 1977;57:826-827.

[13] Mahoney HT. The after-care of poliomyelitis: teaching coordination and balance. Am J Nurs. 1932;32:13-16.

[14] Nelson D. Crutch walking. Am J Nurs. 1939;39:1088-1093.

[15] Olmstead L. Crutch walking. Am J Nurs. 1945;45:28-35.

[16] Childs TF. An analysis of swing-through crutch gait. Phys Ther. 1964;44:804-807.

[17] Stuart L. Method of measurement for walking appliances. Canadian Journal of Occupational Therapy. 1965;32:87-88.

[18] Levin RI, Rubin DS, Stinson JP. Quantitative Approaches to Management. 6th ed. New York, NY: McGraw-Hill Book Co; 1986:112.

D Bauer, MPT, is Staff Physical Therapist, McDonald Army Hospital, For Eustis, VA 23604-5548.

D Finch, MPT, is Staff Physical Therapist, Winn Army Community Hospital, Fort Stewart, GA 31314-5300.

K McGough, MPT, is Staff Physical Therapist, Keller Army Hospital, US Military Academy, West Point, NY 10996-1197.

C Benson, MPT, is Assistant Professor, US Army-Baylor University Graduate Program in Physical Therapy, Academy of Health Sciences, US Army, Medicine and Surgery Division-Physical Therapy Branch, Fort Sam Houston, TX 78234-6100 (USA). Address all correspondence to MAJ Benson.

K Finstuen, PhD, is Associate Professor, US Army-Baylor University Graduate Program in Physical Therapy, Academy of Health Sciences, US Army, Fort Sam Houston.

S Allison, MPT, is Instructor, Academy of Health Sciences, US Army, Medicine and Surgery Division Physical Therapy Branch, Fort Sam Houston.

This study was approved by the Clinical Investigations and Human Use Committees at Brooke Army Medical Center, San Antonio, TX.

The opinions or assertions contained herein are the private views of the authors and are not to be construed as official or as reflecting the views of the US Department of the Army or the US Department of Defense.