• View in gallery

    Drawing depicting the lateral (A) and craniocaudal (B) views of a model created by mounting the sagittal right hemipelvis and limb of a canine cadaver onto a wooden base in a study to evaluate the accuracy and reliability of 3 novel goniometers (2 smartphone-based applications [novel goniometers A and B] and a digital device [novel goniometer C]) versus those of a UG for measurement of stifle joint angles in 8 limbs from 4 canine cadavers. The fur was not clipped but was brushed over the lateral incision sites shown in the inset (A) after the placement of carriage bolts to facilitate mounting. Each limb was fixed at 3 arbitrarily selected angles, and 3 evaluators independently performed measurements in triplicate for each angle with each device. The results were compared with joint angle measurements on standard radiographic views (used as a gold standard for calculation of bias and total error) to assess accuracy; the CV for each device was calculated to assess reliability. (Illustration reprinted with permission from Theresa M. Wendland.)

  • 1. Jaegger G, Marcellin-Little DJ, Levine D. Reliability of goniometry in Labrador Retrievers. Am J Vet Res 2002; 63: 979986.

  • 2. Salamh PA, Kolber M. The reliability, minimal detectable change and concurrent validity of a gravity-based bubble inclinometer and iPhone application for measuring standing lumbar lordosis. Physiother Theory Pract 2014; 30: 6267.

    • Search Google Scholar
    • Export Citation
  • 3. Jandi AS, Schulman AJ. Incidence of motion loss of the stifle joint in dogs with naturally occurring cranial cruciate ligament rupture surgically treated with tibial plateau leveling osteotomy: longitudinal clinical study of 412 cases. Vet Surg 2007; 36: 114121.

    • Search Google Scholar
    • Export Citation
  • 4. Singla D, Veqar Z. Methods of postural assessment used for sports persons. J Clin Diagn Res 2014;8:LE01–LE04.

  • 5. Charlton PC, Mentiplay BF, Pua Y-H, et al. Reliability and concurrent validity of a Smartphone, bubble inclinometer and motion analysis system for measurement of hip joint range of motion. J Sci Med Sport 2015; 18: 262267.

    • Search Google Scholar
    • Export Citation
  • 6. Thomas TM, Marcellin-Little DJ, Roe SC, et al. Comparison of measurements obtained by use of an electrogoniometer and a universal plastic goniometer for the assessment of joint motion in dogs. Am J Vet Res 2006; 67: 19741979.

    • Search Google Scholar
    • Export Citation
  • 7. Milgram J, Slonim E, Kass PH, et al. A radiographic study of joint angles in standing dogs. Vet Comp Orthop Traumatol 2004; 17: 8290.

    • Search Google Scholar
    • Export Citation
  • 8. Brown NP, Bertocci GE, Marcellin-Little DJ. Canine stifle joint biomechanics associated with tibial plateau leveling osteotomy predicted by use of a computer model. Am J Vet Res 2014; 75: 626632.

    • Search Google Scholar
    • Export Citation
  • 9. Kim SE, Jones SC, Lewis DD, et al. In-vivo three-dimensional knee kinematics during daily activities in dogs. J Orthop Res 2015; 33: 16031610.

    • Search Google Scholar
    • Export Citation
  • 10. Jaeger GH, Marcellin-Little DJ, DePuy V, et al. Validity of goniometric joint measurements in cats. Am J Vet Res 2007; 68: 822826.

    • Search Google Scholar
    • Export Citation
  • 11. Powell AC, Landman AB, Bates DW. In search of a few good apps. JAMA 2014; 311: 18511852.

  • 12. Ferriero G, Vercelli S, Sartorio F, et al. Reliability of a smartphone-based goniometer for knee joint goniometry. Int J Rehabil Res 2013; 36: 146151.

    • Search Google Scholar
    • Export Citation
  • 13. Ockendon M, Gilbert RE. Validation of a novel smartphone accelerometer-based knee goniometer. J Knee Surg 2012; 25: 341345.

  • 14. Petazzoni M, Jaeger GH. Atlas of clinical goniometry and radiographic measurements of the canine pelvic limb. 2nd ed. Milano, Italy: Merial, 2008;10–31.

    • Search Google Scholar
    • Export Citation
  • 15. Brosseau L, Tousignant M, Budd J, et al. Intratester and inter-tester reliability and criterion validity of the parallelogram and universal goniometers for active knee flexion in healthy subjects. Physiother Res Int 1997; 2: 150166.

    • Search Google Scholar
    • Export Citation
  • 16. Flatland B, Freeman KP, Friedrichs KR, et al. ASVCP quality assurance guidelines: control of general analytical factors in veterinary laboratories. Vet Clin Pathol 2010; 39: 264277.

    • Search Google Scholar
    • Export Citation
  • 17. Sedgwick P. Limits of agreement (Bland-Altman method). BMJ 2013;346: f1630.

  • 18. Harr KE, Flatland B, Nabity M, et al. ASVCP guidelines: allowable total error guidelines for biochemistry. Vet Clin Pathol 2013; 42: 424436.

    • Search Google Scholar
    • Export Citation
  • 19. Ferriero G, Sartorio F, Foti C, et al. Reliability of a new application for smartphones (DrGoniometer) for elbow angle measurement. PM R 2011; 3: 11531154.

    • Search Google Scholar
    • Export Citation
  • 20. Vohralik SL, Bowen AR, Burns J, et al. Reliability and validity of a smartphone app to measure joint range. Am J Phys Med Rehabil 2015; 94: 325330.

    • Search Google Scholar
    • Export Citation
  • 21. Govoni VM, Rahal SC, Agostinho FS, et al. Goniometric measurements of the forelimb and hindlimb joints in sheep. Vet Comp Orthop Traumatol 2012; 25: 297300.

    • Search Google Scholar
    • Export Citation

Advertisement

Assessment of novel digital and smartphone goniometers for measurement of canine stifle joint angles

View More View Less
  • 1 Department of Clinical Sciences, College of Veterinary Medicine and Biomedical Sciences, Colorado State University, Fort Collins, CO 80523.
  • | 2 Department of Clinical Sciences, College of Veterinary Medicine and Biomedical Sciences, Colorado State University, Fort Collins, CO 80523.
  • | 3 Department of Clinical Sciences, College of Veterinary Medicine and Biomedical Sciences, Colorado State University, Fort Collins, CO 80523.
  • | 4 Department of Clinical Sciences, College of Veterinary Medicine and Biomedical Sciences, Colorado State University, Fort Collins, CO 80523.
  • | 5 Department of Veterinary Microbiology and Preventative Medicine, College of Veterinary Medicine, Iowa State University, Ames, IA 50011.
  • | 6 Department of Clinical Sciences, College of Veterinary Medicine and Biomedical Sciences, Colorado State University, Fort Collins, CO 80523.

Abstract

OBJECTIVE To evaluate accuracy and reliability of 3 novel goniometers for measurement of canine stifle joint angles and compare the results with those obtained with a universal goniometer (UG).

SAMPLE 8 pelvic limbs from 4 canine cadavers.

PROCEDURES Each limb was secured to a wooden platform at 3 arbitrarily selected fixed stifle joint angles. Goniometry was performed with 2 smartphone-based applications (novel goniometers A and B), a digital goniometer (novel goniometer C), and a UG; 3 evaluators performed measurements in triplicate for each angle with each device. Results were compared with stifle joint angle measurements on radiographs (used as a gold standard). Accuracy was determined by calculation of bias and total error, coefficients of variation were calculated to estimate reliability, and strength of linear association between radiographic and goniometer measurements was assessed by calculation of correlation coefficients.

RESULTS Mean coefficient of variation was lowest for the UG (4.88%), followed by novel goniometers B (7.37%), A (7.57%), and C (12.71%). Correlation with radiographic measurements was highest for the UG (r = 0.97), followed by novel goniometers B (0.93), A (0.90), and C (0.78). Constant bias was present for all devices except novel goniometer B. The UG and novel goniometer A had positive constant bias; novel goniometer C had negative constant bias. Total error at 50° and 100° angles was > 5% for all devices.

CONCLUSIONS AND CLINICAL RELEVANCE None of the devices accurately represented radiographically measured stifle joint angles. Additional veterinary studies are indicated prior to the use of novel goniometers in dogs.

Abstract

OBJECTIVE To evaluate accuracy and reliability of 3 novel goniometers for measurement of canine stifle joint angles and compare the results with those obtained with a universal goniometer (UG).

SAMPLE 8 pelvic limbs from 4 canine cadavers.

PROCEDURES Each limb was secured to a wooden platform at 3 arbitrarily selected fixed stifle joint angles. Goniometry was performed with 2 smartphone-based applications (novel goniometers A and B), a digital goniometer (novel goniometer C), and a UG; 3 evaluators performed measurements in triplicate for each angle with each device. Results were compared with stifle joint angle measurements on radiographs (used as a gold standard). Accuracy was determined by calculation of bias and total error, coefficients of variation were calculated to estimate reliability, and strength of linear association between radiographic and goniometer measurements was assessed by calculation of correlation coefficients.

RESULTS Mean coefficient of variation was lowest for the UG (4.88%), followed by novel goniometers B (7.37%), A (7.57%), and C (12.71%). Correlation with radiographic measurements was highest for the UG (r = 0.97), followed by novel goniometers B (0.93), A (0.90), and C (0.78). Constant bias was present for all devices except novel goniometer B. The UG and novel goniometer A had positive constant bias; novel goniometer C had negative constant bias. Total error at 50° and 100° angles was > 5% for all devices.

CONCLUSIONS AND CLINICAL RELEVANCE None of the devices accurately represented radiographically measured stifle joint angles. Additional veterinary studies are indicated prior to the use of novel goniometers in dogs.

Goniometry is frequently used as an objective outcome measure during rehabilitation to assess clinical changes in orthopedic patients in veterinary and human medicine.1,2 These measurements also allow for quantitative patient evaluation to aid in diagnosis as well as prognostication.3,4 Goniometry is considered a simple, reliable, low-cost, and time-efficient outcome measurement readily available for practical clinical application.5

Currently used tools for measurement of joint range of motion include radiography, the commonly used plastic UG, and more specialized technologies such as electrogoniometers, bubble inclinometers, fluoroscopy, and kinematic gait analysis.5–9 Regardless of the device used, these methods require time to obtain measurements and some require additional equipment that may not be widely available to clinicians.5,10 The challenges inherent to veterinary practice require devices and methods that allow observers to accurately, reliably, and efficiently measure joint angles in a broadly diverse set of patients.6

The medical use of handheld digital devices, particularly smartphone-based applications, has steadily increased since their advent, with a recently reported estimate of > 40,000 health, fitness, and medical smartphone-based applications currently available to consumers.11 Smartphone-based applications provide clinicians with alternative goniometers that are readily available, are easily transportable, can allow for sharing of patient data, and can be less expensive than some other specialized devices used for range of motion measurement. Such applications are increasingly considered acceptable alternatives for various indications in human medicine.5,12,13 However, to our knowledge, the use of smartphone-based applications has not yet been evaluated in dogs in the peer-reviewed literature. In addition, specialized handheld digital goniometers have recently been introduced to the medical market. If any device was found to be more accurate and reliable for measurement of joint angles than a traditional UG, a higher cost might be justified for its use in clinical and research applications.

The purpose of the study reported here was to evaluate the reliability and accuracy of 3 novel goniometers (2 smartphone-based applications and a novel digital goniometer) for angle measurements of the stifle joint in dogs and compare the results with those obtained by use of a UG, with radiographic measurements used as the gold standard. We hypothesized that there would be no differences in these variables among the 4 goniometry devices.

Materials and Methods

Sample

A convenience sample of 8 canine cadaver hind limbs from 4 mixed breed medium-sized dogs (euthanized for reasons unrelated to the study) were used. Medium size was defined such that the length of the hind limb was similar to, but did not exceed, 17 inches from the hip joint to the tarsocrural joint with the limb in extension to allow for radiographic evaluation without requiring image manipulation. Limbs were excluded if palpable orthopedic abnormalities of the hind limbs were present. The limbs were prepared to include the hemipelvis with the hip joint left intact. The limbs were individually mounted on wooden platforms, simulating lateral recumbency, and each was arranged in 3 arbitrarily selected stifle joint angles within a previously described expected range for live dogs of a similar size,1 resulting in a total of 24 angles that ranged from 57.6° to 122.2°.

Model construction

Lateral and medial incisions were made over the diaphyses of the femur and tibia without clipping the fur over the site. Soft tissues were sharply and bluntly dissected away from the bone, exposing the periosteum. A 4.75-mm drill bit was used to create 2 or 3 evenly spaced holes (depending on the length of the bone) in the femoral diaphysis, parallel to the transverse plane and perpendicular to the sagittal plane. Two holes, one each at the proximal and distal ends of the diaphysis, were drilled into the tibia in similar manner. Two or 3 holes, corresponding to the number and placement of holes in the femur, were drilled into a 14 × 17-inch wooden platform. Two holes, corresponding with the holes in the tibia, were drilled into the same platform in 3 different positions to produce the 3 selected stifle joint angles. Carriage bolts were driven through the holes in the bones and secured with washers and nuts, and then the free ends of the bolts were driven through the holes in the wooden platform and again secured with washers and nuts. This design resulted in the limb being raised off the wooden platform in a lateral position to allow for circumferential palpation of the limb as would be done in a clinical setting (Figure 1). The bolts used for affixing the limb were covered with overlaying musculature and were positioned away from the landmarks used by palpation. The lateral incisions were cosmetically closed by suturing in a simple continuous subcuticular pattern. The fur was brushed over each lateral incision site to conceal it.

Figure 1—
Figure 1—

Drawing depicting the lateral (A) and craniocaudal (B) views of a model created by mounting the sagittal right hemipelvis and limb of a canine cadaver onto a wooden base in a study to evaluate the accuracy and reliability of 3 novel goniometers (2 smartphone-based applications [novel goniometers A and B] and a digital device [novel goniometer C]) versus those of a UG for measurement of stifle joint angles in 8 limbs from 4 canine cadavers. The fur was not clipped but was brushed over the lateral incision sites shown in the inset (A) after the placement of carriage bolts to facilitate mounting. Each limb was fixed at 3 arbitrarily selected angles, and 3 evaluators independently performed measurements in triplicate for each angle with each device. The results were compared with joint angle measurements on standard radiographic views (used as a gold standard for calculation of bias and total error) to assess accuracy; the CV for each device was calculated to assess reliability. (Illustration reprinted with permission from Theresa M. Wendland.)

Citation: American Journal of Veterinary Research 77, 7; 10.2460/ajvr.77.7.749

Each limb remained affixed at the first selected angle until radiography was completed and all observers had completed their measurements. The nuts were then loosened from the bolts affixing the tibia to the wooden platform; the tibia was moved to create another selected angle, and the nuts were replaced and tightened. The limbs were then radiographed, and goniometric measurements were obtained for the new position. The sequence was then repeated for the final position.

Goniometers

The goniometers used in this investigation included 2 smartphone-based applications (novel goniometers Aa and Bb; cost for the applications alone, $11.99 and $1.99, respectively), a specialized digital handheld goniometerc (novel goniometer C; $245.00), and a 12-inch plastic protractor UGd ($10.00). Application and equipment costs are reported in US dollars at the start of the study (May 2013). The applications were installed on 1 smartphonee that was used throughout the duration of the study.

Data collection

The angle of the stifle joint for all measurements was defined as the angle formed by the longitudinal axis of the femur and the longitudinal axis of the tibia as previously described.1,14 The longitudinal axis of the femur was defined as the line connecting the greater trochanter to the center of the lateral condyle, in the epicondylar region of the femur. The longitudinal axis of the tibia was defined as the line connecting the lateral malleolus to the craniocaudal midpoint of the proximal aspect of the tibia at the level of the tibial crest.1,14

Radiographs of each limb at each selected angle were obtained prior to any goniometric measurements. A single lateromedial radiographic projection was obtained of each limb with the platform in contact with the radiographic table. Focal distance from the x-ray tube to the table was set at 40 inches. A Digital Communications in Medicine (DICOM) viewer was used to make all stifle joint angle measurements by use of the described anatomic landmarks.f One evaluator (NRK) performed all radiographic measurements in triplicate, and the mean of the 3 values was used in analyses.

Three evaluators (NRK, JLH, and SAF), who were trained in the use of each novel goniometer and experienced in use of the UG in daily practice, performed all measurements. The evaluators were given written and verbal instructions on the use of the methods and allowed the opportunity to practice with the devices prior to beginning study measurements. The order in which the evaluators used each goniometer was predetermined such that no device was used repeatedly for measurement of subsequent angles. After completing a set of measurements with one device and prior to measuring with the next, the presentation of the limbs was arbitrarily rearranged by 1 author who did not perform any measurements (KAF). This was done out of sight of the evaluators in an attempt to eliminate any recall of previously measured angles. Thus, 96 goniometric measurements (24 angles measured with 4 devices) were independently performed by 3 evaluators in triplicate fashion, resulting in a total of 864 measurements.

The UG was used in a standard manner according to a published protocol.1 Briefly, the transparent arms of the UG were aligned with the described anatomic landmarks on the limb, with the movement arm aligned distally, and the resulting angle was recorded. As noted previously,1 the axis of rotation did not always overlie the center of the joint.

After opening the smartphone application for novel goniometer A, the device was used in protractor mode with the midpoint of the base of the protractor placed over the center of rotation for the stifle joint, the reference arm positioned along the longitudinal axis of the femur, and the movement arm positioned along the longitudinal axis of the tibia.14 The angle was recorded from the continuous digital display on the screen.

In the application for novel goniometer B, the right or left knee setting (corresponding to the stifle joint being assessed) was selected, and measurements were performed in accordance with the manufacturer instructions. Within the application, the camera was positioned parallel to the plane of the limb to obtain a lateral view. An inclinometer was used to determine correct alignment. The smartphone was moved vertically away from the limb until it was positioned to encompass both the proximal and distal landmarks. On the application interface, the inclinometer was aligned and a photograph was obtained. Then the center landmark indicator was placed over the center of rotation for the stifle joint on the image, and the proximal and distal landmark indicators were positioned over the greater trochanter and the lateral malleolus, respectively.

Novel goniometer C was used according to directions provided by the manufacturer. Briefly, the goniometer was aligned over the mid-diaphysis of the femur, the lasers were activated and used to establish a line between the described bony landmarks of the femur, and an operating button was pressed to create a reference point. The goniometer was then moved in position over the mid-diaphysis of the tibia with the lasers aligned along the longitudinal axis of the bone; excessive rotation of the device was avoided during movement. The operating button was pressed a second time, and the angle displayed on the screen was recorded.

Adjustment of recorded angle measurements

If the supplementary angle (determined as 180° minus the angle of interest) was displayed by a device or recorded by the evaluator, the data were amended to reflect the appropriate angle. Novel goniometer A displayed both angles concurrently on the application screen, and an adjustment of the measured angle was applied if the supplementary angle was recorded. Novel goniometer B (an application designed for use with human patients) was designed to display the supplementary angle to that which is standard in veterinary medicine, and each angle measured with this device was adjusted accordingly. Novel goniometer C would display either the standard angle or the supplementary angle, depending on the direction the device was rotated. The angle for all devices was then evaluated for a deviation of > 30° from the radiographically measured angle. For such deviations, if the supplementary angle was within 30° of the radiographically measured angle, this value was amended to eliminate obvious recording errors. The value of > 30° was chosen because it is well outside of previously reported measurement variability ranges.13,15

Statistical analysis

Statistical analyses were performed by use of statisticalg and electronic spreadsheeth software; a modification of the recommended approach to laboratory method validation16 was used. The CV for each angle was calculated by dividing the SD by the mean of the 3 measurements obtained by the 3 evaluators (ie, 9 replicates). A mean CV was then calculated with data from all the angles for each device and expressed as a percentage.

For the assessment of bias, the mean of the 3 replicates of the radiographically measured stifle joint angle was treated as the true value, and the mean of the 3 replicates of each angle measured by each observer was treated as the measured value for each tested device. To determine the appropriate statistical method for estimating bias, the relationship between device-measured and radiographic values was assessed for each goniometer by means of simple linear regression. The correlation coefficient (r) was < 0.99 for all goniometers; therefore, Deming regression was used to assess the relationship between the angle measured by each device and the radiographic angle. Constant bias and proportional bias were considered present if the confidence interval for the y intercept or the slope of the regression line did not include 0 or overlap 1, respectively. The regression equation (measured value = [slope X radiographic value] + y-intercept) was used to predict the measured angle for each device when the radiographic value was 50° or 100°. Percentage bias at these angles was then calculated as (measured value – radiographic value)/radiographic value X 100.

The total observed error, at an acute angle and an obtuse angle that were both representative of the data sample as well as canine joint stifle range of motion (50° and 100°), was calculated as 2 X CV (%) + bias (%) and subjectively compared with a total allowable error set at 5% on the basis of previously reported differences in clinical outcomes for dogs after tibial plateau leveling osteotomy.3 Estimates of biases and observed error resulted in single values for each device at a given angle.

Results

Ten of the 216 stifle joint angle measurements were obtained with novel goniometer B, which records the supplementary angle to that used in veterinary medicine; therefore those measurements were automatically adjusted and were subsequently reverted to the originally recorded angle, because it was made evident to the investigator recording the values (KAF) that the evaluator had already adjusted the measurement during data collection to indicate the angle of interest. For the remaining devices, the total number of measurements adjusted because it was determined that the supplementary angle had been recorded was 34 (30/216 obtained with novel goniometer C and 4/216 obtained with the UG). Following these adjustments, 33 angles differed from the radiographically measured angle by > 30°. These included 27 measurements obtained with novel goniometer C, 5 obtained with novel goniometer B, and 1 obtained with novel goniometer A.

Reliability and correlation with radiographic measurements

The UG had the greatest reliability for stifle joint angle measurements, with a mean CV of 4.88%, and novel goniometer C had the least reliability, with a mean CV of 12.71%. Mean CVs for novel goniometers A (7.57%) and B (7.37%) fell between these values (Table 1).

Table 1—

Results of Deming regression and simple line regression analysis to assess bias and correlation between canine stifle joint angles measured with each of 4 devices (a UG and novel goniometers A, B, and C) and radiographically measured angles for the same joints, with CVs calculated as precision estimates for each device.

    CV (%)
DeviceSlope (95% CI)y-intercept (95% CI)Correlation coefficient (r)MeanRange
UG0.9058 (0.8524 to 0.9593)17.17 (12.17 to 22.16)0.974.882.24 to 8.34
Novel goniometers
 A0.8536 (0.7543 to 0.9528)22.88 (13.61 to 32.16)0.907.573.15 to 14.46
 B1.002 (0.9092 to 1.095)4.874 (−3.825 to 13.57)0.937.372.44 to 20.40
 C1.440 (1.169 to 1.710)−32.52 (−57.84 to −7.207)0.7812.713.11 to 24.63

Eight hind limbs from 4 canine cadavers, each mounted onto a wooden platform and secured with carriage bolts, were used in the experiments. Each limb was assessed independently at 3 different stifle joint angles by 3 investigators; each investigator used each device to measure angles in triplicate, and results were compared with radiographic measurement for the same stifle joint in the same position. Proportional bias was considered present when the confidence interval (CI) for the slope of the regression line did not overlap 1. Constant bias was considered present when the CI for the y-intercept did not overlap 0. Positive and negative biases are indicated by positive and negative y-intercept values, respectively.

The UG measurements were most closely correlated with radiographic (true) measurements, with a correlation coefficient (r) of 0.97, followed by novel goniometers B (r = 0.97) and A (r = 0.93). Novel goniometer C had the lowest degree of correlation with the radiographic measurements (r = 0.78; Table 1). The correlation coefficient r was used to determine the most appropriate regression method for bias estimation.

Bias

Deming regression analysis revealed evidence of proportional and constant bias for measurement of stifle joint angles by all devices except novel goniometer B (Table 1). The UG and novel goniometer A had positive constant bias, and novel goniometer C had negative constant bias, reflecting overestimation and underestimation of angles, respectively. Estimation of bias in measurements for each device for radiographically measured stifle joint angles of 50° and 100° was summarized (Table 2). Subjectively, novel goniometer B had the lowest degree of bias at both of these angles; the greatest degree of bias at 100° was found for novel goniometer C, and the greatest degree of bias at 50° was found for novel goniometer A.

Table 2—

Estimates of bias and total observed error for measurement of obtuse (100°) and acute (50°) stifle joint angles with the 4 devices in Table 1.

 100°50°
DeviceMeasured angleBias (%)Observed error (%)Measured angleBias (%)Observed error (%)
UG108.758.7518.5162.4624.9234.68
Novel goniometer
 A108.248.2423.3865.5631.1246.26
 B105.075.0719.8154.979.9524.69
 C111.4811.4836.9039.4821.0446.46

The measured angle for each device when the radiographic value was 50° or 100° was predicted by the regression equation (measured value = [slope X radiographic value] + y-intercept). Percentage bias at these angles was then calculated as (measured value – radiographic value)/radiographic value X 100. The total observed error at 50° and 100° was calculated as 2 X CV (%) + bias (%) and subjectively compared with a total allowable error set at 5%.

Performance assessment

Estimates of the total observed error for each device at radiographically measured angles of 50° and 100° were calculated (Table 2). All 4 devices exceeded the total allowable error of 5% and therefore failed to meet the study requirements for acceptable accuracy to represent radiographic measurements. The UG had the lowest observed error rate at a 100° angle (total observed error, 18.51%), and novel goniometer B had the lowest observed error rate at a 50° angle (total observed error, 24.69%). Novel goniometer C had the highest observed error rate of all devices at both 100° and 50° (36.90% and 46.46%, respectively).

Discussion

When evaluating whether a device is useful to determine a clinical measurement, several approaches can be considered. Reliability is frequently evaluated by calculation of the CV of measurements for a given device. Accuracy of a device can be assessed in many different ways, including the assessment of bias when compared with a gold standard. Calculation of the observed total error rate integrates both CV and bias, and comparison of this value with the total allowable error rate determines whether the device meets the needs of clinicians or researchers. The correlation coefficient assesses the strength of a linear association between devices and therefore is not a measure of accuracy.17 Bias, an expression of inaccuracy, is an assessment of the closeness of a measurement with that of the true value.18 The evaluation of bias is beneficial in that it can be used to predict the extent of inaccuracy for different magnitudes of measurement. The presence of proportional bias indicates that the degree of overestimation or underestimation varies for large angles versus small angles. Constant bias, where the degree of error remains independent of the measured true value, would indicate that a device consistently under- or overestimates the true value.16,18

Taken together, the results of the present study indicated that no tested novel goniometer performed better overall than the UG. Although results for all devices were somewhat variable, the UG had a low mean CV and low degrees of proportional and constant bias for stifle joint angle measurements when compared with radiographic values; when evaluated at acute and obtuse angles, the estimates of percentage bias and total observed error were also considered low. For clinical evaluations of progression after treatment is instituted, the relative lack of variability is likely the most important factor to consider. Among the devices tested, measurements made with the UG were also most closely correlated with the true angle as measured on radiographs.

Goniometry is most frequently used to assess clinical progression, such as evaluating the success of a treatment by serial comparison of stifle joint range of motion measurements.3 Given this purpose, the comparison of a goniometric measurement with the radiographic measurement is likely not the most essential outcome measure for device assessment. Rather, how much variability exists for a given patient in measurements from a single device or between devices has a greater potential effect on outcome measures. Therefore, although novel goniometer B had the lowest percentage bias estimate for acute and obtuse angles and the lowest observed error rate estimate for acute angles when compared with radiographic measurements, this does not suggest that the device would be the best alternative to the UG. Novel goniometer A had a proportional bias similar to that of the UG, whereas novel goniometer C had the largest degree of proportional bias of all devices tested. Therefore, the use of novel goniometer A, rather than novel goniometer B, would produce similar results for the same angle as measured by the UG. Furthermore, novel goniometer C would result in different amounts of variation according to how acute or obtuse the measured angle was. While the UG remains the preferred goniometer, on the basis of these findings, novel goniometer A would likely produce the most similar results to the UG in clinical use.

Authors of 1 study3 found that dogs with a ≥ 10° loss in range of motion in the stifle joint (as measured with a UG) after tibial plateau leveling osteotomy had significantly higher clinical lameness scores than dogs with no loss in range of motion, and reported that this degree of change should be considered clinically relevant. If a patient's stifle joint range of motion is 121° (the difference in median extension and flexion values reported in a study1 of clinically normal Labrador Retrievers), then 5% device error would result in a 6° change in a given measured stifle joint angle or a potential 12° change in measured range of motion. This total allowable error was subjectively compared with estimates of total observed error for each device at an obtuse (100°) and an acute (50°) angle. No device, including the UG, had acceptable performance at either angle as judged by this criterion.

A variety of goniometric applications are currently available for download to personal electronic devices such as smartphones. However, an application specifically designed for use in veterinary medicine does not yet exist. The applications used in the present study were selected to represent the different types of applications available, including photograph-based measurements and real-time measurements (used by novel goniometers B and A, respectively). The availability of validation studies for goniometry devices in the human literature was also considered.2,12,19,20

Several limitations of the tested applications were discovered during the study. Subjectively, it was found that the darker or thicker the fur, the more difficult and time-consuming it was for the evaluators to correctly identify the bony landmarks in the photograph-based application. For novel goniometer A, the relatively short goniometer arm length, limited by the size of the phone, could make alignment between the bony landmarks on larger patients more challenging. Novel goniometer C, which is readily portable and has an easily visible readout, was suggested to be a reliable and reproducible tool on the basis of intraclass correlation coefficient and inter- and intrarater reliability in non–peer-reviewed literaturei; the reason for the poor performance of novel goniometer C in our study was unclear, and seemed to be related to the device itself, as results were similar for all observers.

Limitations of the present study included the ex vivo design, which could possibly have skewed the results toward being more favorable for all devices. Since the limb was removed and fixed in position, any movement or influence of muscle tone was eliminated. The design of the study was advantageous in that it limited many variables that can influence measurement data, including limb position, variability in perceived range of motion limits, signs of pain elicited in a patient limiting the measurable range of motion, and patient cooperation, as well as positioning of joints proximal and distal to the stifle joint. Other patient factors including size, age, body conformation, and pathological physiomechanical changes have also been suggested to contribute to variation in goniometric measurements.9,14,21 Thus, we expect that the performance of the novel goniometers tested in the present study will not be improved when used for clinical evaluation of live patients. Validation of such devices for use with live dogs of various sizes and evaluation of their utility in a clinical setting should be pursued in future studies, and it is important that clinicians critically evaluate such studies prior to incorporating the use of these devices in daily practice.

Acknowledgments

Supported by the Young Investigator Grant Program in the Center for Companion Animal Studies, Colorado State University.

The authors declare that there were no conflicts of interest.

Presented in abstract form at the 16th Annual Colorado State University College of Veterinary Medicine and Biomedical Sciences Research Day, Fort Collins, Colo, January 2015; and at the 42nd Annual Veterinary Orthopedic Society Conference, Sun Valley, Idaho, March 2015.

ABBREVIATIONS

CV

Coefficient of variation

UG

Universal goniometer

Footnotes

a.

Smartphone-based application iHandy Carpenter, version 2.2.2, iHandySoft Inc, New York, NY.

b.

Smartphone-based application DrGoniometer, version 2.9, CDM Srl, Milano, Italy.

c.

Handheld digital goniometer HALO, HALO Medical Devices, Subiaco, WA, Australia.

d.

Elite Medical Instruments, Fullerton, Calif.

e.

iPhone 4, Apple Inc, Cupertino, Calif.

f.

Orthoplan Elite, Sound-Eklin, Carlsbad, Calif.

g.

Prism, version 5.01, 2007, Graphpad Software Inc, La Jolla, Calif.

h.

Microsoft Excel, version 14, 2010, Microsoft Corp, Redmond, Wash.

i.

Halo Medical Devices. Validation and collaborative research. Available at: halomedicaldevices.com/research/. Accessed Jan 8, 2015.

References

  • 1. Jaegger G, Marcellin-Little DJ, Levine D. Reliability of goniometry in Labrador Retrievers. Am J Vet Res 2002; 63: 979986.

  • 2. Salamh PA, Kolber M. The reliability, minimal detectable change and concurrent validity of a gravity-based bubble inclinometer and iPhone application for measuring standing lumbar lordosis. Physiother Theory Pract 2014; 30: 6267.

    • Search Google Scholar
    • Export Citation
  • 3. Jandi AS, Schulman AJ. Incidence of motion loss of the stifle joint in dogs with naturally occurring cranial cruciate ligament rupture surgically treated with tibial plateau leveling osteotomy: longitudinal clinical study of 412 cases. Vet Surg 2007; 36: 114121.

    • Search Google Scholar
    • Export Citation
  • 4. Singla D, Veqar Z. Methods of postural assessment used for sports persons. J Clin Diagn Res 2014;8:LE01–LE04.

  • 5. Charlton PC, Mentiplay BF, Pua Y-H, et al. Reliability and concurrent validity of a Smartphone, bubble inclinometer and motion analysis system for measurement of hip joint range of motion. J Sci Med Sport 2015; 18: 262267.

    • Search Google Scholar
    • Export Citation
  • 6. Thomas TM, Marcellin-Little DJ, Roe SC, et al. Comparison of measurements obtained by use of an electrogoniometer and a universal plastic goniometer for the assessment of joint motion in dogs. Am J Vet Res 2006; 67: 19741979.

    • Search Google Scholar
    • Export Citation
  • 7. Milgram J, Slonim E, Kass PH, et al. A radiographic study of joint angles in standing dogs. Vet Comp Orthop Traumatol 2004; 17: 8290.

    • Search Google Scholar
    • Export Citation
  • 8. Brown NP, Bertocci GE, Marcellin-Little DJ. Canine stifle joint biomechanics associated with tibial plateau leveling osteotomy predicted by use of a computer model. Am J Vet Res 2014; 75: 626632.

    • Search Google Scholar
    • Export Citation
  • 9. Kim SE, Jones SC, Lewis DD, et al. In-vivo three-dimensional knee kinematics during daily activities in dogs. J Orthop Res 2015; 33: 16031610.

    • Search Google Scholar
    • Export Citation
  • 10. Jaeger GH, Marcellin-Little DJ, DePuy V, et al. Validity of goniometric joint measurements in cats. Am J Vet Res 2007; 68: 822826.

    • Search Google Scholar
    • Export Citation
  • 11. Powell AC, Landman AB, Bates DW. In search of a few good apps. JAMA 2014; 311: 18511852.

  • 12. Ferriero G, Vercelli S, Sartorio F, et al. Reliability of a smartphone-based goniometer for knee joint goniometry. Int J Rehabil Res 2013; 36: 146151.

    • Search Google Scholar
    • Export Citation
  • 13. Ockendon M, Gilbert RE. Validation of a novel smartphone accelerometer-based knee goniometer. J Knee Surg 2012; 25: 341345.

  • 14. Petazzoni M, Jaeger GH. Atlas of clinical goniometry and radiographic measurements of the canine pelvic limb. 2nd ed. Milano, Italy: Merial, 2008;10–31.

    • Search Google Scholar
    • Export Citation
  • 15. Brosseau L, Tousignant M, Budd J, et al. Intratester and inter-tester reliability and criterion validity of the parallelogram and universal goniometers for active knee flexion in healthy subjects. Physiother Res Int 1997; 2: 150166.

    • Search Google Scholar
    • Export Citation
  • 16. Flatland B, Freeman KP, Friedrichs KR, et al. ASVCP quality assurance guidelines: control of general analytical factors in veterinary laboratories. Vet Clin Pathol 2010; 39: 264277.

    • Search Google Scholar
    • Export Citation
  • 17. Sedgwick P. Limits of agreement (Bland-Altman method). BMJ 2013;346: f1630.

  • 18. Harr KE, Flatland B, Nabity M, et al. ASVCP guidelines: allowable total error guidelines for biochemistry. Vet Clin Pathol 2013; 42: 424436.

    • Search Google Scholar
    • Export Citation
  • 19. Ferriero G, Sartorio F, Foti C, et al. Reliability of a new application for smartphones (DrGoniometer) for elbow angle measurement. PM R 2011; 3: 11531154.

    • Search Google Scholar
    • Export Citation
  • 20. Vohralik SL, Bowen AR, Burns J, et al. Reliability and validity of a smartphone app to measure joint range. Am J Phys Med Rehabil 2015; 94: 325330.

    • Search Google Scholar
    • Export Citation
  • 21. Govoni VM, Rahal SC, Agostinho FS, et al. Goniometric measurements of the forelimb and hindlimb joints in sheep. Vet Comp Orthop Traumatol 2012; 25: 297300.

    • Search Google Scholar
    • Export Citation

Contributor Notes

Address correspondence to Dr. Duerr (Felix.Duerr@colostate.edu).

Dr. Freund's present address is Department of Clinical Sciences, College of Veterinary Medicine, Cornell University, Ithaca, NY 14853.

Dr. Kieves' present address is Department of Veterinary Clinical Sciences, College of Veterinary Medicine, The Ohio State University, Columbus, OH 43210.