- Email: [email protected]
Measurement of active shoulder motion using the Kinect, a commercially available infrared position detection system
J Shoulder Elbow Surg (2015) -, 1-8
www.elsevier.com/locate/ymse
Measurement of active shoulder motion using the Kinect, a commercially available infrared position detection system Frederick A. Matsen III, MD*, Alexander Lauder, MD, Kyle Rector, MS, Peyton Keeling, MD (Candidate 2017), Arien L. Cherones Department of Orthopaedics and Sports Medicine, University of Washington, Seattle, WA, USA Background: The shoulder’s ability to participate in sports and activities of daily living depends on its active range of motion. Clinical goniometry is of limited utility in rigorously assessing limitation of motion and the effectiveness of treatment. We sought to determine (1) whether a validated position-sensing tool, the Kinect, can enable the objective clinical measurement of shoulder motion and (2) the degree to which active range of motion correlates with patient self-assessed shoulder function. Methods: In 10 control subjects, we compared Kinect motion measurements to measurements made on standardized anteroposterior and lateral photographs taken concurrently. In 51 patients, we correlated active motion with the ability to perform the functions of the Simple Shoulder Test (SST). Results: In controls, Kinect measurements strongly agreed with photographic measurements. In patients, the total SST score was strongly correlated with the range of active abduction. The ability to perform each of the individual SST functions was strongly correlated with active motion. The active motion in well-functioning patient shoulders averaged 155 22 abduction, 159 14 flexion, 76 18 external rotation in abduction, 59 25 internal rotation in abduction, and 3.3 3.7 inches of cross-body adduction, values similar to the control shoulders. Use of the Kinect system was practical in clinical examination rooms, requiring <5 minutes to document the 5 motions in both shoulders. Discussion: The Kinect provides a clinically practical method for objectively measuring active shoulder motion. Active motion was an important determinant of patient-assessed shoulder function. Level of evidence: Level III, Diagnostic Study. Ó 2015 Journal of Shoulder and Elbow Surgery Board of Trustees. Keywords: Shoulder; motion; active; measurement
Active range of shoulder motion enables patients to perform activities of daily living and to participate in sports.31,38,43 Restoration of active range of motion is one This prospective study was approved by our Human Subjects Review Committee (Approval #47398). *Reprints requests: Frederick A. Matsen III, MD, Shoulder and Elbow Surgery, Department of Orthopaedics and Sports Medicine, University of Washington Medical Center, 1959 NE Pacific Street, Box 356500, Seattle, WA 98195-6500, USA. E-mail address: [email protected] (F.A. Matsen III).
of the major goals for patients having treatment for shoulder conditions.19,34,42,44,49 Ranges of motion are important elements of commonly used shoulder outcome measures; for example, they account for 40% of the total Constant score.26 However, there is wide variability in the way in which the range of shoulder motion is measured in clinical practice. This includes differences in the experience and training of the person making the measurements; whether and how a hand-held goniometer is used; how the center of rotation is estimated; the plane in which
1058-2746/$ - see front matter Ó 2015 Journal of Shoulder and Elbow Surgery Board of Trustees. http://dx.doi.org/10.1016/j.jse.2015.07.011
2
F.A. Matsen et al.
shoulder motion is measured; whether the patient is lying, sitting, or standing; whether the motion is active or passive; and the control for errors related to trunk lean and rotation.6,8,9,15,22,25,28,32,36,39,45,46,48,50 This difficulty is exemplified by the observation that the interobserver variability in the range of motion measurement in the Constant score can exceed 20%, even when a standardized goniometric method is used.7 The variability in the clinical measurement of shoulder motion makes it difficult to rigorously evaluate the extent of the limitations and the effectiveness of different approaches to the management of common pathologic processes of the shoulder, such as rotator cuff disease, internal rotation deficit,10,13,14,20,27,30,35 and arthritis, as well as to determine the effects of interventions, such as remplissage, on the range of shoulder motion in living patients.11,37 This investigation explored the clinical measurement of the active range of shoulder motion using the Kinect (Microsoft, Redmond, WA, USA), an inexpensive, commercially available infrared system for the detection of body position. The use of the markerless Kinect in the measurement of shoulder positions has been extensively validated by comparison to marker-based optical motion capture systems.1-5,12,18,21,23,24,41,47 We sought to answer 4 questions. First, in normal control subjects, how do the Kinect measurements of abduction, flexion, external rotation in abduction, internal rotation in abduction, and cross-body adduction correlate with measurements of these arm positions made on standardized anteroposterior and lateral photographs taken concurrently? Second, can the Kinect system efficiently and practically measure the range of active shoulder motion of patients in the clinic setting? Third, how does the measured range of active motion correlate with patients’ self-reported shoulder function as documented by the 12 functions of the Simple Shoulder Test (SST)? Fourth, what are the objectively measured ranges of active motion in normally functioning shoulders?
Materials and methods The Kinect is an inexpensive (<$150), commercially available device used in computer games for the observer-independent detection of body position without the attachment of markers or sensors or any physical contact with the subject (Fig. 1). Instead, it projects infrared laser light on the person standing facing it and uses depth-sensing cameras to create a 3-dimensional body map by analyzing the pattern of the light reflected from the subject. The Kinect superimposes known relationships of the human trunk and limbs on this pattern, allowing the determination of the 3-dimensional positions of the forearm, elbow, arm, shoulder, and trunk (Fig. 2). The Kinect output enables the quantification of (1) humerothoracic abduction (the angle between the arm and the axis of the trunk in the coronal plane), (2) humerothoracic flexion (the angle between the arm and the axis of the trunk in the sagittal plane), (3) external rotation in abduction (the angle of the forearm
Figure 1 The Kinect sensor (Microsoft, Redmond, WA, USA) is an inexpensive (<$150), commercially available device with an infrared emitter and depth-sensing cameras enabling it to determine in 3 dimensions the location of the trunk and limbs of a person facing it. The sensor is mounted on a stand so that its height can be adjusted to the level of the subject’s shoulder. The system is controlled by software on an ordinary laptop computer. above the horizontal when the arm is abducted 90 and the elbow is flexed to 90 dgiven a positive sign), (4) internal rotation in abduction (the angle of the forearm below the horizontal when the arm is abducted 90 and the elbow is flexed to 90 dgiven a negative sign), and (5) cross-body adduction (the distance in inches that the point of the elbow can reach across the midline of the trunkdgiven a positive sign if the elbow reaches across the midline). Each of the 5 measurements is made with the subject’s body in the same position facing the sensor.
Kinect measurement of active shoulder motion
3 photographic measurements were made by one of the investigators blinded to the results from the Kinect system. The correlations between the 2 measurement methods were determined from linear regression. The mean differences between the 2 measurement methods were determined for each motion.
Initial clinical experience with Kinect position measurements
Figure 2 The on-screen rendering of the position of the subject’s body parts: thorax (T), scapula (S), arm (A), elbow (E), and forearm (F). The software takes the 3-dimensional positions of these body parts and sequentially captures the angle of abduction (between T and A in the coronal plane), flexion (between T and A in the sagittal plane), external and internal rotation in abduction (the angle between the forearm of the abducted arm, F, and the horizontal), and cross-body adduction (the distance by which the tip of the elbow, E, crosses the midline). Note that while the subject leans to the side, the Kinect makes the correction by measuring the angle between the thorax and the arm.
Comparison of Kinect and photographic ranges of motion We compared the Kinect measurements with measurements made on standardized anteroposterior and lateral photographs taken concurrently. This comparison was carried out for the right shoulders of 10 normal adult volunteer subjects (5 men, 5 women; age, 23-33 years) from our residency and staff; none of these shoulders had a history of shoulder disease or current shoulder symptoms. Each subject actively positioned the arm sequentially in 5 different positions each of abduction, flexion, rotation, and cross-body adduction while the Kinect measurements and photographs were made. To measure abduction on the photographs, the angle between the arm and the axis of the trunk was measured with a protractor on an anteroposterior view of the subject. For flexion, the angle between the arm and the axis of the trunk was measured with a protractor on a lateral photograph. For rotation of the abducted arm, the angle between the forearm of the flexed elbow and the horizontal was measured with a protractor on a lateral photograph. For cross-body adduction, the distance the elbow reached across the midline was scaled to a ruler placed on the subject’s chest during the anteroposterior photograph. The
We reviewed the medical records of the first 51 shoulders in 32 patients with a variety of diagnoses (cuff disease, instability, arthritis) seen in our shoulder clinic having both (1) the Kinect measurement of maximal active range of abduction, flexion, external rotation in abduction, internal rotation in abduction, and cross-body adduction and (2) the SST patient self-assessment of shoulder comfort and function as a part of their clinical evaluation. The average range of each motion for shoulders that could and could not perform each of the 12 SST functions was compared by the unpaired t test assuming unequal variance. Statistical significance was set at P < .05. Next, a linear regression was carried out for each of the 5 motions against the total number of SST functions the shoulder was able to perform; the associated correlation coefficients were calculated. Finally, the 21 shoulders with a high level of function as evidenced by a positive answer to the SST question ‘‘Can you throw overhand?’’ were identified and their ranges of motion compared with those of the 10 control subjects measured with the Kinect system.
Results In the 10 control subjects, the results of the Kinect and photographic measurements were highly consistent with each other. The scatter plots (Figs. 3-6) for the different motions were essentially linear, with Pearson correlation coefficients (r) of 0.997, 0.992, 0.982, and 0.995 for abduction, flexion, cross-body adduction, and rotation in 90 of abduction, respectively. The average differences between the 2 types of measurement were 1.4 4.9 for abduction, 2.4 7.6 for flexion, 0.15 1.52 inches for cross-body adduction, and 3.5 6.2 for rotation in abduction. The maximal active ranges of motion in the 10 control subjects averaged 163 9 of abduction, 160 6 of flexion, 80 7 of external rotation in abduction, 67 4 of internal rotation in abduction, and 0 1 inches of cross-body adduction. The Kinect measurements were easily obtained in our clinic without the need to have reflectors or sensors attached to the subject’s body or limbs. The average time to document the 5 motions in both shoulders was 4.8 1.0 minutes (range, 3.4-7.5). The Kinect measurement system was portable, allowing easy movement from one clinic examination room to another. For the 51 shoulders of clinic patients, the Kinect measurements of active motion correlated with the patient’s assessment of the shoulder’s ability to perform the SST. The Pearson correlation coefficients (r) for the total number of SST functions
4
Figure 3 Agreement in measurements of abduction between the Kinect and protractor measurements made on standardized anteroposterior photographs. Pearson correlation coefficient (r) ¼ 0.997.
Figure 4 Agreement in measurements of flexion between the Kinect and protractor measurements made on standardized lateral photographs. Pearson correlation coefficient (r) ¼ 0.992.
F.A. Matsen et al.
Figure 6 Agreement in measurements of rotation with the arm in 90 of abduction between the Kinect and protractor measurements made on standardized lateral photographs. Pearson correlation coefficient (r) ¼ 0.995.
Figure 7 Linear regression of the total number of the 12 SST functions performable against the maximum range of active abduction. The linear regression equation is shown with r ¼ 0.79.
measured by the Kinect system were significantly different between those shoulders that could and those that could not perform each of the 12 individual functions of the SST (Table I). The 21 shoulders for which the patient indicated the ability to throw overhand had maximal ranges of motion very similar to those of the 10 normal control shoulders; the variability was greater among the shoulders of patients than among the controls (Table II).
Discussion Figure 5 Agreement in measurements of cross-body adduction between the Kinect and ruler measurements made on standardized anteroposterior photographs. Pearson correlation coefficient (r) ¼ 0.982.
a shoulder could perform and the active range of motion were 0.79 for abduction (Fig. 7), 0.67 for flexion, 0.56 for external rotation, 0.50 for internal rotation, and 0.33 for cross-body adduction. The maximal ranges of motion as
This investigation is unique in that it (1) presents a clinically practical system for the documentation of the active range of shoulder motion using the Kinect, a previously validated, inexpensive, commercially available body position sensing device; (2) demonstrates the consistency of the Kinect measurements with those made from standardized photographs; (3) shows the practicality of Kinect measurements of range of motion in the clinic setting; (4)
Kinect measurement of active shoulder motion
5
Table I Maximal ranges of abduction, flexion, external rotation in abduction, internal rotation in abduction, and cross-body adduction for shoulders able and unable to perform each of the 12 functions of the SST SST function) No. of patients Abduction Flexion External rotation Internal rotation Cross-body adduction able or unable (degrees) (degrees) in abduction in abduction (distance that elbow to perform (degrees) (degrees) reaches across midline in inches) Comfortable by side Unable P-value Sleep comfortably Unable P-value Tuck in shirt Unable P-value Place hand behind head Unable P-value Place coin on shelf Unable P-value Lift 1 pound to shoulder level Unable P-value Lift 8 pounds to top of head Unable P-value Carry 8 pounds by side Unable P-value Toss underhand Unable P-value Throw overhand Unable P-value Wash back of opposite shoulder Unable P-value Do usual work Unable P-value
37 14 21 27 31 20 31 20 33 16 31 19 21 27 29 21 29 21 18 30 25 25 20 21
139 89 .001 149 107 <.001 143 98 <.001 147 91 <.001 145 87 <.001 148 90 <.001 155 101 <.001 148 94 <.001 145 98 <.001 159 106 <.001 153 99 <.001 154 99 <.001
35 150 41 102 .003 31 155 43 122 .002 36 153 38 111 <.001 30 154 38 110 <.001 32 153 38 106 <.001 32 154 35 107 <.001 22 159 41 116 <.001 32 149 38 120 .013 31 151 43 117 .007 13 158 43 124 <.001 22 160 42 113 <.001 23 156 44 118 .002
26 73 17 46 42 35 .006 21 75 15 45 55 32 .006 30 75 18 38 46 32 .001 22 75 17 44 46 32 .002 24 76 14 43 45 33 .003 25 72 20 40 49 33 .013 14 76 18 42 52 30 .002 31 73 18 43 50 33 .011 24 70 23 47 54 32 .059 15 72 20 45 58 31 .068 15 76 18 42 50 30 <.001 17 75 17 46 53 33 .011
53 41 .131 58 45 .099 59 34 <.001 54 42 .090 58 34 .003 58 35 .003 59 40 .009 58 39 .007 54 45 .172 63 45 .010 64 34 <.001 58 41 .023
25 24 26 24 23 22 25 24 23 23 22 25 25 23 22 24 24 25 20 24 18 23 19 25
3.3 5.0 .061 3.1 4.2 .303 3.0 4.9 .044 3.6 4.1 .589 2.8 5.0 .012 2.9 5.1 .020 3.3 4.4 .268 3.1 4.8 .062 2.9 5.0 .030 2.9 4.4 .190 2.7 4.8 .037 2.5 4.2 .106
3.7 2.4 4.1 2.8 3.7 2.7 3.8 2.7 3.6 2.1 3.5 2.9 3.7 3.0 3.8 2.6 3.7 2.7 4.1 3.0 3.8 2.8 3.5 2.8
Values are means standard deviation. P values are for the unpaired t test comparing those shoulders able and unable to perform the function. ) Note that not all patients answered every question.
demonstrates the relationship between the measured ranges of motion and patient self-assessed shoulder function; and (5) provides objective values for the ranges of motion in normally functioning shouldersddata that can serve as the basis for future comparisons with shoulders compromised by conditions such as rotator cuff disease and arthritis. The importance of objective measurement of the active range of shoulder motion rests in the fact that improvement in the functional range of motion is one of the principal goals of patients seeking treatment for shoulder problems. For this reason, estimates of range of motion are a part of almost
every clinical outcome report and are an intrinsic component of many shoulder scoring systems. The problem is that the clinical measurement of the range of shoulder motion is not standardized among observers: the position of the patient, the control of the thoracic reference, the estimation of the center of shoulder motion, the plane of shoulder motion, whether the motion is active or passive, the manner in which the hand-held measuring device is applied, and even whether a measuring device is used at all differ widely among reports of clinical outcomes.8,9,15,17,22,25,28,29,32,33,39,40,45,46,48,50 The rigorous investigation of different procedures, performed by
6
F.A. Matsen et al. Table II Mean (SD) maximal ranges of 5 shoulder motions for the right shoulders of 10 control subjects and 21 shoulders of patients who were able to throw overhand Motion measured
Normal control shoulders (n ¼ 10)
Abduction (degrees) Flexion (degrees) External rotation in abduction (degrees) Internal rotation in abduction (degrees) Cross-body adduction (inches cross the midline)
163 160 80 67 0
9 6 7 4 1
Patients’ shoulders that were able to throw (n ¼ 21)
P
155 22 159 14 76 18 59 25 3 4
.173 .945 .383 .178 <.001
The right column shows the P values for the unpaired t test comparing these 2 groups of shoulders. Note that whereas the shoulders of patients had slightly less motion than the controls, the difference was significant only for cross-body adduction. Note also that the standard deviations were larger for the patients than for the controls.
different surgeons for different diagnoses in different patient populations, requires a practical, efficient, inexpensive, and objective method for the clinical assessment of active range of motion. The markerless Kinect has been extensively validated by comparing its measurements to those made with marker-based optical sensing systems.1-5,12,18,21,23,24,41,47 These articles have documented the test-retest reliability of the Kinect system.3,4,18 They point to the fact that optical stereophotogrammetry and electromagnetic sensing systems are expensive, are time-consuming, and would occupy an impractical amount of space in the clinic.1,2 Furthermore, these articles point out the advantages of the Kinect markerless approach over conventional markerbased motion capture systems: less time-consuming, avoidance of errors in marker placement, ease of use, lower cost, greater transportability, and the use of advanced pattern recognition to detect skeletal elements.3,4,16,18 The limitations of our study are that we did not repeat the previously published validation and reproducibility studies with the Kinect and that we have yet to analyze the Kinect measurements before and after the treatment of different shoulder conditions.
Conclusion This study demonstrates that the Kinect infrared sensor technology provides a practical and inexpensive method for the measurement of the active range of shoulder motion in a clinical research environment. Our results showed the strong relationship between the Kinect measurements of active range of motion and the patients’ ability to perform 12 standard defined shoulder functions. The Kinect measurements in well-functioning shoulders yielded normative data against which shoulders with different pathologic processes can be compared. The use of such an objective system for the measurement of active shoulder motion holds promise for clarifying the
indications for and the results of therapeutic interventions designed to improve shoulder function.
Disclaimer The authors, their immediate families, and any research foundation with which they are affiliated have not received any financial payments or other benefits from any commercial entity related to the subject of this article.
References 1. Aizawa J, Masuda T, Hyodo K, Jinno T, Yagishita K, Nakamaru K, et al. Ranges of active joint motion for the shoulder, elbow, and wrist in healthy adults. Disabil Rehabil 2013;35:1342-9. http://dx.doi.org/ 10.3109/09638288.2012.731133 2. Alta TD, de Toledo JM, Veeger HE, Janssen TW, Willems WJ. The active and passive kinematic difference between primary reverse and total shoulder prostheses. J Shoulder Elbow Surg 2014;23:1395-402. http://dx.doi.org/10.1016/j.jse.2014.01.040 3. Bonnechere B, Jansen B, Salvia P, Bouzahouene H, Omelina L, Moiseev F, et al. Validity and reliability of the Kinect within functional assessment activities: comparison with standard stereophotogrammetry. Gait Posture 2014;39:593-8. http://dx.doi.org/10. 1016/j.gaitpost.2013.09.018 4. Bonnechere B, Sholukha V, Jansen B, Omelina L, Rooze M, Van Sint Jan S. Determination of repeatability of Kinect sensor. Telemed J E Health 2014;20:451-3. http://dx.doi.org/10.1089/tmj.2013.0247 5. Cippitelli E, Gasparrini S, Gambi E. Kinect as a tool for gait analysis: validation of real-time joint extraction algorithm working in side view. Sensors 2015;15:1417-34. http://dx.doi.org/10.3390/s150101417 6. Collin P, Matsumura N, Ladermann A, Denard PJ, Walch G. Relationship between massive chronic rotator cuff tear pattern and loss of active shoulder range of motion. J Shoulder Elbow Surg 2014;23: 1195-202. http://dx.doi.org/10.1016/j.jse.2013.11.019 7. Conboy VB, Morris RW, Kiss J, Carr AJ. An evaluation of the ConstantMurley shoulder assessment. J Bone Joint Surg Br 1996;78:229-32. 8. Cools AM, De Wilde L, Van Tongel A, Ceyssens C, Ryckewaert R, Cambier DC. Measuring shoulder external and internal rotation
Kinect measurement of active shoulder motion
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
strength and range of motion: comprehensive intra-rater and inter-rater reliability study of several testing protocols. J Shoulder Elbow Surg 2014;23:1454-61. http://dx.doi.org/10.1016/j.jse.2014.01.006 Cuff D, Pupello D, Virani N, Levy J, Frankle M. Reverse shoulder arthroplasty for the treatment of rotator cuff deficiency. J Bone Joint Surg Am 2008;90:1244-51. http://dx.doi.org/10.2106/JBJS.G.00775 Dines JS, Frank JB, Akerman M, Yocum LA. Glenohumeral internal rotation deficits in baseball players with ulnar collateral ligament insufficiency. Am J Sports Med 2009;37:566-70. http://dx.doi.org/10. 1177/0363546508326712 Elkinson I, Giles JW, Faber KJ, Boons HW, Ferreira LM, Johnson JA, et al. The effect of the remplissage procedure on shoulder stability and range of motion: an in vitro biomechanical assessment. J Bone Joint Surg Am 2012;94:1003-12. http://dx.doi.org/10.2106/JBJS.J.01956 Fernandez-Baena A, Susın A, Lligadas X. Biomechanical validation of upper-body and lower-body joint movements of Kinect motion capture data for rehabilitation treatments. 2012. In. Fourth International Conference on Intelligent Networking and Collaborative Systems: Available at: http://upcommons.upc.edu/e-prints/bitstream/2117/1711 7/1/incos2012_656_661.pdf. (accessed January 5, 2015) Garrison JC, Cole MA, Conway JE, Macko MJ, Thigpen C, Shanley E. Shoulder range of motion deficits in baseball players with an ulnar collateral ligament tear. Am J Sports Med 2012;40:2597-603. http:// dx.doi.org/10.1177/0363546512459175 Gates JJ, Gupta A, McGarry MH, Tibone JE, Lee TQ. The effect of glenohumeral internal rotation deficit due to posterior capsular contracture on passive glenohumeral joint motion. Am J Sports Med 2012;41:2794-800. http://dx.doi.org/10.1177/0363546512462012 Giphart JE, Burnkhorst JP, Horn NH, Shelburne KB, Torry MR, Millett PJ. Effect of plane of arm elevation on glenohumeral kinematics: a normative biplane fluoroscopy study. J Bone Joint Surg Am 2013;95:238-45. http://dx.doi.org/10.2106/JBJS.J.01875 Haddad CA, Saad M, Perez Mdel C, Miranda Junior F. Assessment of posture and joint movements of the upper limbs of patients after mastectomy and lymphadenectomy. Einstein (Sao Paulo) 2013;11: 426-34. http://dx.doi.org/10.1590/S1679-45082013000400004 Hall JM, Azar FM, Miller RH 3rd, Smith R, Throckmorton TW. Accuracy and reliability testing of two methods to measure internal rotation of the glenohumeral joint. J Shoulder Elbow Surg 2014;23: 1296-300. http://dx.doi.org/10.1016/j.jse.2013.12.015 Hawi N, Liodakis E, Musolli D, Suero EM, Stuebig T, Claassen L, et al. Range of motion assessment of the shoulder and elbow joints using a motion sensing input device: a pilot study. Technol Health Care 2014;22:289-95. http://dx.doi.org/10.3233/THC-140831 Henn RF 3rd, Ghomrawi H, Rutledge JR, Mazumdar M, Mancuso CA, Marx RG. Preoperative patient expectations of total shoulder arthroplasty. J Bone Joint Surg Am 2011;93:2110-5. http://dx.doi.org/10. 2106/JBJS.J.01114 Hibberd EE, Oyama S, Myers JB. Increase in humeral retrotorsion accounts for age-related increase in glenohumeral internal rotation deficit in youth and adolescent baseball players. Am J Sports Med 2014;42:851-8. http://dx.doi.org/10.1177/0363546513519325 Hondori HM, Khademi M. A review on technical and clinical impact of Microsoft Kinect on physical therapy and rehabilitation. J Med Eng 2014;2014. http://dx.doi.org/10.1155/2014/846514 Kim SJ, Kim SH, Lee SK, Seo JW, Chun YM. Arthroscopic repair of massive contracted rotator cuff tears: aggressive release with anterior and posterior interval slides do not improve cuff healing and integrity. J Bone Joint Surg Am 2013;95:1482-8. http://dx.doi.org/10.2106/ JBJS.L.01193 Kizony R, Weiss PL, Elion O, Harel S, Baum-Cohen I, Krasovsky T, et al. Development and validation of tele-health system for stroke rehabilitation. Int J Disabil Hum Dev 2014;13:361-8. http://dx.doi.org/ 10.1515/ijdhd-2014-0329 Kurillo G, Han JJ, Obdrzalek S, Yan P, Abresch RT, Nicorici A, et al. Upper extremity reachable workspace evaluation with Kinect. Stud Health Technol Inform 2013;184:247-53.
7 25. Levine WN, Fischer CR, Nguyen D, Flatow EL, Ahmad CS, Bigliani LU. Long-term follow-up of shoulder hemiarthroplasty for glenohumeral osteoarthritis. J Bone Joint Surg Am 2012;94:e164. http://dx.doi.org/10.2106/JBJS.K.00603 26. Levy O, Haddo O, Massoud S, Mullett H, Atoun E. A patient-derived Constant-Murley score is comparable to a clinician-derived score. Clin Orthop Relat Res 2014;472:294-303. http://dx.doi.org/10.1007/ s11999-013-3249-3 27. Litner D, Mayol M, Uzodinma O, Jones R, Labossiere D. Glenohumeral internal rotation deficits in professional pitchers enrolled in an internal rotation stretching program. Am J Sports Med 2007;35: 617-21. http://dx.doi.org/10.1177/0363546506296736 28. Ludewig PM, Phadke V, Braman JP, Hassett DR, Cieminski CJ, LaPrade RF. Motion of the shoulder complex during multiplanar humeral elevation. J Bone Joint Surg Am 2009;91:378-89. http://dx.doi. org/10.2106/JBJS.G.01483 29. MacDermid JC, Chesworth BM, Patterson S, Roth JH. Intratester and intertester reliability of goniometric measurement of passive lateral shoulder rotation. J Hand Ther 1999;12:187-92. 30. Maenhout A, Van Eessel V, Van Dyck L, Vanraes A, Cools A. Quantifying acromiohumeral distance in overhead athletes with glenohumeral internal rotation loss and the influence of a stretching program. Am J Sports Med 2012;40:2105-12. http://dx.doi.org/10. 1177/0363546512454530 31. Matsen FA 3rd, Lippitt SB, Sidles JA, Harryman DT 2nd. Practical evaluation and management of the shoulder. Philadelphia: WB Saunders; 1994. 32. Mizuno N, Denard PJ, Raiss P, Walch G. Reverse total shoulder arthroplasty for primary glenohumeral osteoarthritis in patients with a biconcave glenoid. J Bone Joint Surg Am 2013;95:1297-304. http://dx. doi.org/10.2106/JBJS.L.00820 33. Muir SW, Corea CL, Beaupre L. Evaluating change in clinical status: reliability and measures of agreement for the assessment of glenohumeral range of motion. N Am J Sports Phys Ther 2010;5: 98-110. 34. Myers JB, Laudner KG, Pasquale MR, Bradley JP, Lephart SM. Glenohumeral range of motion deficits and posterior shoulder tightness in throwers with pathologic internal impingement. Am J Sports Med 2006;34:385-91. http://dx.doi.org/10.1177/03635465052 81804 35. Myers JB, Oyama S, Wassinger CA, Ricci RD, Abt JP, Conley KM, et al. Reliability, precision, accuracy, and validity of posterior shoulder tightness assessment in overhead athletes. Am J Sports Med 2007;35: 1922-30. http://dx.doi.org/10.1177/0363546507304142 36. O’Neill BJ, O’Briain D, Hirpara KM, Shaughnesy M, Yeatman EA, Kaar TK. Digital photography for assessment of shoulder range of motion: a novel clinical and research tool. Int J Shoulder Surg 2013;7: 23-7. http://dx.doi.org/10.4103/0973-6042.109888 37. Omi R, Hooke AW, Zhao KD, Matsuhashi T, Goto A, Yamamoto N, et al. The effect of the remplissage procedure on shoulder range of motion: a cadaveric study. Arthroscopy 2014;30:178-87. http://dx.doi. org/10.1016/j.arthro.2013.11.003 38. Pearl ML, Harris SL, Lippitt SB, Sidles JA, Harryman DT 2nd, Matsen FA 3rd. A system for describing positions of the humerus relative to the thorax and its use in the presentation of several functionally important arm positions. J Shoulder Elbow Surg 1992;1: 113-8. 39. Raiss P, Schmitt M, Bruckner T, Kasten P, Pape G, Loew M, et al. Results of cemented total shoulder replacement with a minimum follow-up of ten years. J Bone Joint Surg Am 2012;94. http://dx.doi. org/10.2106/JBJS.K.00580. e1711-10. 40. Riddle DL, Rothstein JM, Lamb RL. Goniometric reliability in a clinical setting. Shoulder measurements. Phys Ther 1987;67:668-73. 41. Schouten B, Fedtke S, Bekker T, Schijven M, Gekker A. Games for health. Proceedings of the 3rd European conference on gaming and playful interaction in health care. Wiesbaden: Springer Vieweg; 2013. p. 1-293.
8 42. Torres RR, Gomes JL. Measurement of glenohumeral internal rotation in asymptomatic tennis players and swimmers. Am J Sports Med 2009;37:1017-23. http://dx.doi.org/10.1177/0363546508329544 43. Triffitt PD. The relationship between motion of the shoulder and the stated ability to perform activities of daily living. J Bone Joint Surg Am 1998;80:41-6. 44. Tyler TF, Nicholas SJ, Lee SJ, Mullaney M, McHugh MP. Correction of posterior shoulder tightness is associated with symptom resolution in patients with internal impingement. Am J Sports Med 2010;38: 114-9. http://dx.doi.org/10.1177/0363546509346050 45. Vairo GL, Duffey ML, Owens BD, Cameron KL. Clinical descriptive measures of shoulder range of motion for a healthy, young and physically active cohort. Sports Med Arthrosc Rehabil Ther Technol 2012;4:33. http://dx.doi.org/10.1186/1758-2555-4-33 46. van de Pol RJ, van Trijffel E, Lucas C. Inter-rater reliability for measurement of passive physiological range of motion of upper extremity joints is better if instruments are used: a systematic review. J Physiother 2010;56:7-17. http://dx.doi.org/10.1016/S1836-9553(10) 70049-7
F.A. Matsen et al. 47. Webster D, Celik O. Experimental evaluation of Microsoft Kinect’s accuracy and capture rate for stroke rehabilitaton applications. Houston, TX: Institute of Electrical and Electronics Engineers (IEEE); 2014. p. 455-60. 48. Werner BC, Holzgrefe RE, Griffin JW, Lyons ML, Cosgrove CT, Hart JM, et al. Validation of an innovative method of shoulder rangeof-motion measurement using a smartphone clinometer application. J Shoulder Elbow Surg 2014;23:e275-82. http://dx.doi.org/10.1016/j.jse. 2014.02.030 49. Wilk KE, Macrina LC, Fleisig GS, Porterfield R, Simpson CD, Harker P, et al. Correlation of glenohumeral internal rotation deficit and total rotational motion to shoulder injuries in professional baseball pitchers. Am J Sports Med 2011;39:329-35. http://dx.doi.org/10.1177/ 0363546510384223 50. Young AA, Walch G, Pape G, Gohike F, Favard L. Secondary rotator cuff dysfunction following total shoulder arthroplasty for primary glenohumeral osteoarthritis: results of a multicenter study with more than five years of follow-up. J Bone Joint Surg Am 2012;94:685-93. http://dx.doi.org/10.2106/JBJS.J.00727
www.elsevier.com/locate/ymse
Measurement of active shoulder motion using the Kinect, a commercially available infrared position detection system Frederick A. Matsen III, MD*, Alexander Lauder, MD, Kyle Rector, MS, Peyton Keeling, MD (Candidate 2017), Arien L. Cherones Department of Orthopaedics and Sports Medicine, University of Washington, Seattle, WA, USA Background: The shoulder’s ability to participate in sports and activities of daily living depends on its active range of motion. Clinical goniometry is of limited utility in rigorously assessing limitation of motion and the effectiveness of treatment. We sought to determine (1) whether a validated position-sensing tool, the Kinect, can enable the objective clinical measurement of shoulder motion and (2) the degree to which active range of motion correlates with patient self-assessed shoulder function. Methods: In 10 control subjects, we compared Kinect motion measurements to measurements made on standardized anteroposterior and lateral photographs taken concurrently. In 51 patients, we correlated active motion with the ability to perform the functions of the Simple Shoulder Test (SST). Results: In controls, Kinect measurements strongly agreed with photographic measurements. In patients, the total SST score was strongly correlated with the range of active abduction. The ability to perform each of the individual SST functions was strongly correlated with active motion. The active motion in well-functioning patient shoulders averaged 155 22 abduction, 159 14 flexion, 76 18 external rotation in abduction, 59 25 internal rotation in abduction, and 3.3 3.7 inches of cross-body adduction, values similar to the control shoulders. Use of the Kinect system was practical in clinical examination rooms, requiring <5 minutes to document the 5 motions in both shoulders. Discussion: The Kinect provides a clinically practical method for objectively measuring active shoulder motion. Active motion was an important determinant of patient-assessed shoulder function. Level of evidence: Level III, Diagnostic Study. Ó 2015 Journal of Shoulder and Elbow Surgery Board of Trustees. Keywords: Shoulder; motion; active; measurement
Active range of shoulder motion enables patients to perform activities of daily living and to participate in sports.31,38,43 Restoration of active range of motion is one This prospective study was approved by our Human Subjects Review Committee (Approval #47398). *Reprints requests: Frederick A. Matsen III, MD, Shoulder and Elbow Surgery, Department of Orthopaedics and Sports Medicine, University of Washington Medical Center, 1959 NE Pacific Street, Box 356500, Seattle, WA 98195-6500, USA. E-mail address: [email protected] (F.A. Matsen III).
of the major goals for patients having treatment for shoulder conditions.19,34,42,44,49 Ranges of motion are important elements of commonly used shoulder outcome measures; for example, they account for 40% of the total Constant score.26 However, there is wide variability in the way in which the range of shoulder motion is measured in clinical practice. This includes differences in the experience and training of the person making the measurements; whether and how a hand-held goniometer is used; how the center of rotation is estimated; the plane in which
1058-2746/$ - see front matter Ó 2015 Journal of Shoulder and Elbow Surgery Board of Trustees. http://dx.doi.org/10.1016/j.jse.2015.07.011
2
F.A. Matsen et al.
shoulder motion is measured; whether the patient is lying, sitting, or standing; whether the motion is active or passive; and the control for errors related to trunk lean and rotation.6,8,9,15,22,25,28,32,36,39,45,46,48,50 This difficulty is exemplified by the observation that the interobserver variability in the range of motion measurement in the Constant score can exceed 20%, even when a standardized goniometric method is used.7 The variability in the clinical measurement of shoulder motion makes it difficult to rigorously evaluate the extent of the limitations and the effectiveness of different approaches to the management of common pathologic processes of the shoulder, such as rotator cuff disease, internal rotation deficit,10,13,14,20,27,30,35 and arthritis, as well as to determine the effects of interventions, such as remplissage, on the range of shoulder motion in living patients.11,37 This investigation explored the clinical measurement of the active range of shoulder motion using the Kinect (Microsoft, Redmond, WA, USA), an inexpensive, commercially available infrared system for the detection of body position. The use of the markerless Kinect in the measurement of shoulder positions has been extensively validated by comparison to marker-based optical motion capture systems.1-5,12,18,21,23,24,41,47 We sought to answer 4 questions. First, in normal control subjects, how do the Kinect measurements of abduction, flexion, external rotation in abduction, internal rotation in abduction, and cross-body adduction correlate with measurements of these arm positions made on standardized anteroposterior and lateral photographs taken concurrently? Second, can the Kinect system efficiently and practically measure the range of active shoulder motion of patients in the clinic setting? Third, how does the measured range of active motion correlate with patients’ self-reported shoulder function as documented by the 12 functions of the Simple Shoulder Test (SST)? Fourth, what are the objectively measured ranges of active motion in normally functioning shoulders?
Materials and methods The Kinect is an inexpensive (<$150), commercially available device used in computer games for the observer-independent detection of body position without the attachment of markers or sensors or any physical contact with the subject (Fig. 1). Instead, it projects infrared laser light on the person standing facing it and uses depth-sensing cameras to create a 3-dimensional body map by analyzing the pattern of the light reflected from the subject. The Kinect superimposes known relationships of the human trunk and limbs on this pattern, allowing the determination of the 3-dimensional positions of the forearm, elbow, arm, shoulder, and trunk (Fig. 2). The Kinect output enables the quantification of (1) humerothoracic abduction (the angle between the arm and the axis of the trunk in the coronal plane), (2) humerothoracic flexion (the angle between the arm and the axis of the trunk in the sagittal plane), (3) external rotation in abduction (the angle of the forearm
Figure 1 The Kinect sensor (Microsoft, Redmond, WA, USA) is an inexpensive (<$150), commercially available device with an infrared emitter and depth-sensing cameras enabling it to determine in 3 dimensions the location of the trunk and limbs of a person facing it. The sensor is mounted on a stand so that its height can be adjusted to the level of the subject’s shoulder. The system is controlled by software on an ordinary laptop computer. above the horizontal when the arm is abducted 90 and the elbow is flexed to 90 dgiven a positive sign), (4) internal rotation in abduction (the angle of the forearm below the horizontal when the arm is abducted 90 and the elbow is flexed to 90 dgiven a negative sign), and (5) cross-body adduction (the distance in inches that the point of the elbow can reach across the midline of the trunkdgiven a positive sign if the elbow reaches across the midline). Each of the 5 measurements is made with the subject’s body in the same position facing the sensor.
Kinect measurement of active shoulder motion
3 photographic measurements were made by one of the investigators blinded to the results from the Kinect system. The correlations between the 2 measurement methods were determined from linear regression. The mean differences between the 2 measurement methods were determined for each motion.
Initial clinical experience with Kinect position measurements
Figure 2 The on-screen rendering of the position of the subject’s body parts: thorax (T), scapula (S), arm (A), elbow (E), and forearm (F). The software takes the 3-dimensional positions of these body parts and sequentially captures the angle of abduction (between T and A in the coronal plane), flexion (between T and A in the sagittal plane), external and internal rotation in abduction (the angle between the forearm of the abducted arm, F, and the horizontal), and cross-body adduction (the distance by which the tip of the elbow, E, crosses the midline). Note that while the subject leans to the side, the Kinect makes the correction by measuring the angle between the thorax and the arm.
Comparison of Kinect and photographic ranges of motion We compared the Kinect measurements with measurements made on standardized anteroposterior and lateral photographs taken concurrently. This comparison was carried out for the right shoulders of 10 normal adult volunteer subjects (5 men, 5 women; age, 23-33 years) from our residency and staff; none of these shoulders had a history of shoulder disease or current shoulder symptoms. Each subject actively positioned the arm sequentially in 5 different positions each of abduction, flexion, rotation, and cross-body adduction while the Kinect measurements and photographs were made. To measure abduction on the photographs, the angle between the arm and the axis of the trunk was measured with a protractor on an anteroposterior view of the subject. For flexion, the angle between the arm and the axis of the trunk was measured with a protractor on a lateral photograph. For rotation of the abducted arm, the angle between the forearm of the flexed elbow and the horizontal was measured with a protractor on a lateral photograph. For cross-body adduction, the distance the elbow reached across the midline was scaled to a ruler placed on the subject’s chest during the anteroposterior photograph. The
We reviewed the medical records of the first 51 shoulders in 32 patients with a variety of diagnoses (cuff disease, instability, arthritis) seen in our shoulder clinic having both (1) the Kinect measurement of maximal active range of abduction, flexion, external rotation in abduction, internal rotation in abduction, and cross-body adduction and (2) the SST patient self-assessment of shoulder comfort and function as a part of their clinical evaluation. The average range of each motion for shoulders that could and could not perform each of the 12 SST functions was compared by the unpaired t test assuming unequal variance. Statistical significance was set at P < .05. Next, a linear regression was carried out for each of the 5 motions against the total number of SST functions the shoulder was able to perform; the associated correlation coefficients were calculated. Finally, the 21 shoulders with a high level of function as evidenced by a positive answer to the SST question ‘‘Can you throw overhand?’’ were identified and their ranges of motion compared with those of the 10 control subjects measured with the Kinect system.
Results In the 10 control subjects, the results of the Kinect and photographic measurements were highly consistent with each other. The scatter plots (Figs. 3-6) for the different motions were essentially linear, with Pearson correlation coefficients (r) of 0.997, 0.992, 0.982, and 0.995 for abduction, flexion, cross-body adduction, and rotation in 90 of abduction, respectively. The average differences between the 2 types of measurement were 1.4 4.9 for abduction, 2.4 7.6 for flexion, 0.15 1.52 inches for cross-body adduction, and 3.5 6.2 for rotation in abduction. The maximal active ranges of motion in the 10 control subjects averaged 163 9 of abduction, 160 6 of flexion, 80 7 of external rotation in abduction, 67 4 of internal rotation in abduction, and 0 1 inches of cross-body adduction. The Kinect measurements were easily obtained in our clinic without the need to have reflectors or sensors attached to the subject’s body or limbs. The average time to document the 5 motions in both shoulders was 4.8 1.0 minutes (range, 3.4-7.5). The Kinect measurement system was portable, allowing easy movement from one clinic examination room to another. For the 51 shoulders of clinic patients, the Kinect measurements of active motion correlated with the patient’s assessment of the shoulder’s ability to perform the SST. The Pearson correlation coefficients (r) for the total number of SST functions
4
Figure 3 Agreement in measurements of abduction between the Kinect and protractor measurements made on standardized anteroposterior photographs. Pearson correlation coefficient (r) ¼ 0.997.
Figure 4 Agreement in measurements of flexion between the Kinect and protractor measurements made on standardized lateral photographs. Pearson correlation coefficient (r) ¼ 0.992.
F.A. Matsen et al.
Figure 6 Agreement in measurements of rotation with the arm in 90 of abduction between the Kinect and protractor measurements made on standardized lateral photographs. Pearson correlation coefficient (r) ¼ 0.995.
Figure 7 Linear regression of the total number of the 12 SST functions performable against the maximum range of active abduction. The linear regression equation is shown with r ¼ 0.79.
measured by the Kinect system were significantly different between those shoulders that could and those that could not perform each of the 12 individual functions of the SST (Table I). The 21 shoulders for which the patient indicated the ability to throw overhand had maximal ranges of motion very similar to those of the 10 normal control shoulders; the variability was greater among the shoulders of patients than among the controls (Table II).
Discussion Figure 5 Agreement in measurements of cross-body adduction between the Kinect and ruler measurements made on standardized anteroposterior photographs. Pearson correlation coefficient (r) ¼ 0.982.
a shoulder could perform and the active range of motion were 0.79 for abduction (Fig. 7), 0.67 for flexion, 0.56 for external rotation, 0.50 for internal rotation, and 0.33 for cross-body adduction. The maximal ranges of motion as
This investigation is unique in that it (1) presents a clinically practical system for the documentation of the active range of shoulder motion using the Kinect, a previously validated, inexpensive, commercially available body position sensing device; (2) demonstrates the consistency of the Kinect measurements with those made from standardized photographs; (3) shows the practicality of Kinect measurements of range of motion in the clinic setting; (4)
Kinect measurement of active shoulder motion
5
Table I Maximal ranges of abduction, flexion, external rotation in abduction, internal rotation in abduction, and cross-body adduction for shoulders able and unable to perform each of the 12 functions of the SST SST function) No. of patients Abduction Flexion External rotation Internal rotation Cross-body adduction able or unable (degrees) (degrees) in abduction in abduction (distance that elbow to perform (degrees) (degrees) reaches across midline in inches) Comfortable by side Unable P-value Sleep comfortably Unable P-value Tuck in shirt Unable P-value Place hand behind head Unable P-value Place coin on shelf Unable P-value Lift 1 pound to shoulder level Unable P-value Lift 8 pounds to top of head Unable P-value Carry 8 pounds by side Unable P-value Toss underhand Unable P-value Throw overhand Unable P-value Wash back of opposite shoulder Unable P-value Do usual work Unable P-value
37 14 21 27 31 20 31 20 33 16 31 19 21 27 29 21 29 21 18 30 25 25 20 21
139 89 .001 149 107 <.001 143 98 <.001 147 91 <.001 145 87 <.001 148 90 <.001 155 101 <.001 148 94 <.001 145 98 <.001 159 106 <.001 153 99 <.001 154 99 <.001
35 150 41 102 .003 31 155 43 122 .002 36 153 38 111 <.001 30 154 38 110 <.001 32 153 38 106 <.001 32 154 35 107 <.001 22 159 41 116 <.001 32 149 38 120 .013 31 151 43 117 .007 13 158 43 124 <.001 22 160 42 113 <.001 23 156 44 118 .002
26 73 17 46 42 35 .006 21 75 15 45 55 32 .006 30 75 18 38 46 32 .001 22 75 17 44 46 32 .002 24 76 14 43 45 33 .003 25 72 20 40 49 33 .013 14 76 18 42 52 30 .002 31 73 18 43 50 33 .011 24 70 23 47 54 32 .059 15 72 20 45 58 31 .068 15 76 18 42 50 30 <.001 17 75 17 46 53 33 .011
53 41 .131 58 45 .099 59 34 <.001 54 42 .090 58 34 .003 58 35 .003 59 40 .009 58 39 .007 54 45 .172 63 45 .010 64 34 <.001 58 41 .023
25 24 26 24 23 22 25 24 23 23 22 25 25 23 22 24 24 25 20 24 18 23 19 25
3.3 5.0 .061 3.1 4.2 .303 3.0 4.9 .044 3.6 4.1 .589 2.8 5.0 .012 2.9 5.1 .020 3.3 4.4 .268 3.1 4.8 .062 2.9 5.0 .030 2.9 4.4 .190 2.7 4.8 .037 2.5 4.2 .106
3.7 2.4 4.1 2.8 3.7 2.7 3.8 2.7 3.6 2.1 3.5 2.9 3.7 3.0 3.8 2.6 3.7 2.7 4.1 3.0 3.8 2.8 3.5 2.8
Values are means standard deviation. P values are for the unpaired t test comparing those shoulders able and unable to perform the function. ) Note that not all patients answered every question.
demonstrates the relationship between the measured ranges of motion and patient self-assessed shoulder function; and (5) provides objective values for the ranges of motion in normally functioning shouldersddata that can serve as the basis for future comparisons with shoulders compromised by conditions such as rotator cuff disease and arthritis. The importance of objective measurement of the active range of shoulder motion rests in the fact that improvement in the functional range of motion is one of the principal goals of patients seeking treatment for shoulder problems. For this reason, estimates of range of motion are a part of almost
every clinical outcome report and are an intrinsic component of many shoulder scoring systems. The problem is that the clinical measurement of the range of shoulder motion is not standardized among observers: the position of the patient, the control of the thoracic reference, the estimation of the center of shoulder motion, the plane of shoulder motion, whether the motion is active or passive, the manner in which the hand-held measuring device is applied, and even whether a measuring device is used at all differ widely among reports of clinical outcomes.8,9,15,17,22,25,28,29,32,33,39,40,45,46,48,50 The rigorous investigation of different procedures, performed by
6
F.A. Matsen et al. Table II Mean (SD) maximal ranges of 5 shoulder motions for the right shoulders of 10 control subjects and 21 shoulders of patients who were able to throw overhand Motion measured
Normal control shoulders (n ¼ 10)
Abduction (degrees) Flexion (degrees) External rotation in abduction (degrees) Internal rotation in abduction (degrees) Cross-body adduction (inches cross the midline)
163 160 80 67 0
9 6 7 4 1
Patients’ shoulders that were able to throw (n ¼ 21)
P
155 22 159 14 76 18 59 25 3 4
.173 .945 .383 .178 <.001
The right column shows the P values for the unpaired t test comparing these 2 groups of shoulders. Note that whereas the shoulders of patients had slightly less motion than the controls, the difference was significant only for cross-body adduction. Note also that the standard deviations were larger for the patients than for the controls.
different surgeons for different diagnoses in different patient populations, requires a practical, efficient, inexpensive, and objective method for the clinical assessment of active range of motion. The markerless Kinect has been extensively validated by comparing its measurements to those made with marker-based optical sensing systems.1-5,12,18,21,23,24,41,47 These articles have documented the test-retest reliability of the Kinect system.3,4,18 They point to the fact that optical stereophotogrammetry and electromagnetic sensing systems are expensive, are time-consuming, and would occupy an impractical amount of space in the clinic.1,2 Furthermore, these articles point out the advantages of the Kinect markerless approach over conventional markerbased motion capture systems: less time-consuming, avoidance of errors in marker placement, ease of use, lower cost, greater transportability, and the use of advanced pattern recognition to detect skeletal elements.3,4,16,18 The limitations of our study are that we did not repeat the previously published validation and reproducibility studies with the Kinect and that we have yet to analyze the Kinect measurements before and after the treatment of different shoulder conditions.
Conclusion This study demonstrates that the Kinect infrared sensor technology provides a practical and inexpensive method for the measurement of the active range of shoulder motion in a clinical research environment. Our results showed the strong relationship between the Kinect measurements of active range of motion and the patients’ ability to perform 12 standard defined shoulder functions. The Kinect measurements in well-functioning shoulders yielded normative data against which shoulders with different pathologic processes can be compared. The use of such an objective system for the measurement of active shoulder motion holds promise for clarifying the
indications for and the results of therapeutic interventions designed to improve shoulder function.
Disclaimer The authors, their immediate families, and any research foundation with which they are affiliated have not received any financial payments or other benefits from any commercial entity related to the subject of this article.
References 1. Aizawa J, Masuda T, Hyodo K, Jinno T, Yagishita K, Nakamaru K, et al. Ranges of active joint motion for the shoulder, elbow, and wrist in healthy adults. Disabil Rehabil 2013;35:1342-9. http://dx.doi.org/ 10.3109/09638288.2012.731133 2. Alta TD, de Toledo JM, Veeger HE, Janssen TW, Willems WJ. The active and passive kinematic difference between primary reverse and total shoulder prostheses. J Shoulder Elbow Surg 2014;23:1395-402. http://dx.doi.org/10.1016/j.jse.2014.01.040 3. Bonnechere B, Jansen B, Salvia P, Bouzahouene H, Omelina L, Moiseev F, et al. Validity and reliability of the Kinect within functional assessment activities: comparison with standard stereophotogrammetry. Gait Posture 2014;39:593-8. http://dx.doi.org/10. 1016/j.gaitpost.2013.09.018 4. Bonnechere B, Sholukha V, Jansen B, Omelina L, Rooze M, Van Sint Jan S. Determination of repeatability of Kinect sensor. Telemed J E Health 2014;20:451-3. http://dx.doi.org/10.1089/tmj.2013.0247 5. Cippitelli E, Gasparrini S, Gambi E. Kinect as a tool for gait analysis: validation of real-time joint extraction algorithm working in side view. Sensors 2015;15:1417-34. http://dx.doi.org/10.3390/s150101417 6. Collin P, Matsumura N, Ladermann A, Denard PJ, Walch G. Relationship between massive chronic rotator cuff tear pattern and loss of active shoulder range of motion. J Shoulder Elbow Surg 2014;23: 1195-202. http://dx.doi.org/10.1016/j.jse.2013.11.019 7. Conboy VB, Morris RW, Kiss J, Carr AJ. An evaluation of the ConstantMurley shoulder assessment. J Bone Joint Surg Br 1996;78:229-32. 8. Cools AM, De Wilde L, Van Tongel A, Ceyssens C, Ryckewaert R, Cambier DC. Measuring shoulder external and internal rotation
Kinect measurement of active shoulder motion
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
strength and range of motion: comprehensive intra-rater and inter-rater reliability study of several testing protocols. J Shoulder Elbow Surg 2014;23:1454-61. http://dx.doi.org/10.1016/j.jse.2014.01.006 Cuff D, Pupello D, Virani N, Levy J, Frankle M. Reverse shoulder arthroplasty for the treatment of rotator cuff deficiency. J Bone Joint Surg Am 2008;90:1244-51. http://dx.doi.org/10.2106/JBJS.G.00775 Dines JS, Frank JB, Akerman M, Yocum LA. Glenohumeral internal rotation deficits in baseball players with ulnar collateral ligament insufficiency. Am J Sports Med 2009;37:566-70. http://dx.doi.org/10. 1177/0363546508326712 Elkinson I, Giles JW, Faber KJ, Boons HW, Ferreira LM, Johnson JA, et al. The effect of the remplissage procedure on shoulder stability and range of motion: an in vitro biomechanical assessment. J Bone Joint Surg Am 2012;94:1003-12. http://dx.doi.org/10.2106/JBJS.J.01956 Fernandez-Baena A, Susın A, Lligadas X. Biomechanical validation of upper-body and lower-body joint movements of Kinect motion capture data for rehabilitation treatments. 2012. In. Fourth International Conference on Intelligent Networking and Collaborative Systems: Available at: http://upcommons.upc.edu/e-prints/bitstream/2117/1711 7/1/incos2012_656_661.pdf. (accessed January 5, 2015) Garrison JC, Cole MA, Conway JE, Macko MJ, Thigpen C, Shanley E. Shoulder range of motion deficits in baseball players with an ulnar collateral ligament tear. Am J Sports Med 2012;40:2597-603. http:// dx.doi.org/10.1177/0363546512459175 Gates JJ, Gupta A, McGarry MH, Tibone JE, Lee TQ. The effect of glenohumeral internal rotation deficit due to posterior capsular contracture on passive glenohumeral joint motion. Am J Sports Med 2012;41:2794-800. http://dx.doi.org/10.1177/0363546512462012 Giphart JE, Burnkhorst JP, Horn NH, Shelburne KB, Torry MR, Millett PJ. Effect of plane of arm elevation on glenohumeral kinematics: a normative biplane fluoroscopy study. J Bone Joint Surg Am 2013;95:238-45. http://dx.doi.org/10.2106/JBJS.J.01875 Haddad CA, Saad M, Perez Mdel C, Miranda Junior F. Assessment of posture and joint movements of the upper limbs of patients after mastectomy and lymphadenectomy. Einstein (Sao Paulo) 2013;11: 426-34. http://dx.doi.org/10.1590/S1679-45082013000400004 Hall JM, Azar FM, Miller RH 3rd, Smith R, Throckmorton TW. Accuracy and reliability testing of two methods to measure internal rotation of the glenohumeral joint. J Shoulder Elbow Surg 2014;23: 1296-300. http://dx.doi.org/10.1016/j.jse.2013.12.015 Hawi N, Liodakis E, Musolli D, Suero EM, Stuebig T, Claassen L, et al. Range of motion assessment of the shoulder and elbow joints using a motion sensing input device: a pilot study. Technol Health Care 2014;22:289-95. http://dx.doi.org/10.3233/THC-140831 Henn RF 3rd, Ghomrawi H, Rutledge JR, Mazumdar M, Mancuso CA, Marx RG. Preoperative patient expectations of total shoulder arthroplasty. J Bone Joint Surg Am 2011;93:2110-5. http://dx.doi.org/10. 2106/JBJS.J.01114 Hibberd EE, Oyama S, Myers JB. Increase in humeral retrotorsion accounts for age-related increase in glenohumeral internal rotation deficit in youth and adolescent baseball players. Am J Sports Med 2014;42:851-8. http://dx.doi.org/10.1177/0363546513519325 Hondori HM, Khademi M. A review on technical and clinical impact of Microsoft Kinect on physical therapy and rehabilitation. J Med Eng 2014;2014. http://dx.doi.org/10.1155/2014/846514 Kim SJ, Kim SH, Lee SK, Seo JW, Chun YM. Arthroscopic repair of massive contracted rotator cuff tears: aggressive release with anterior and posterior interval slides do not improve cuff healing and integrity. J Bone Joint Surg Am 2013;95:1482-8. http://dx.doi.org/10.2106/ JBJS.L.01193 Kizony R, Weiss PL, Elion O, Harel S, Baum-Cohen I, Krasovsky T, et al. Development and validation of tele-health system for stroke rehabilitation. Int J Disabil Hum Dev 2014;13:361-8. http://dx.doi.org/ 10.1515/ijdhd-2014-0329 Kurillo G, Han JJ, Obdrzalek S, Yan P, Abresch RT, Nicorici A, et al. Upper extremity reachable workspace evaluation with Kinect. Stud Health Technol Inform 2013;184:247-53.
7 25. Levine WN, Fischer CR, Nguyen D, Flatow EL, Ahmad CS, Bigliani LU. Long-term follow-up of shoulder hemiarthroplasty for glenohumeral osteoarthritis. J Bone Joint Surg Am 2012;94:e164. http://dx.doi.org/10.2106/JBJS.K.00603 26. Levy O, Haddo O, Massoud S, Mullett H, Atoun E. A patient-derived Constant-Murley score is comparable to a clinician-derived score. Clin Orthop Relat Res 2014;472:294-303. http://dx.doi.org/10.1007/ s11999-013-3249-3 27. Litner D, Mayol M, Uzodinma O, Jones R, Labossiere D. Glenohumeral internal rotation deficits in professional pitchers enrolled in an internal rotation stretching program. Am J Sports Med 2007;35: 617-21. http://dx.doi.org/10.1177/0363546506296736 28. Ludewig PM, Phadke V, Braman JP, Hassett DR, Cieminski CJ, LaPrade RF. Motion of the shoulder complex during multiplanar humeral elevation. J Bone Joint Surg Am 2009;91:378-89. http://dx.doi. org/10.2106/JBJS.G.01483 29. MacDermid JC, Chesworth BM, Patterson S, Roth JH. Intratester and intertester reliability of goniometric measurement of passive lateral shoulder rotation. J Hand Ther 1999;12:187-92. 30. Maenhout A, Van Eessel V, Van Dyck L, Vanraes A, Cools A. Quantifying acromiohumeral distance in overhead athletes with glenohumeral internal rotation loss and the influence of a stretching program. Am J Sports Med 2012;40:2105-12. http://dx.doi.org/10. 1177/0363546512454530 31. Matsen FA 3rd, Lippitt SB, Sidles JA, Harryman DT 2nd. Practical evaluation and management of the shoulder. Philadelphia: WB Saunders; 1994. 32. Mizuno N, Denard PJ, Raiss P, Walch G. Reverse total shoulder arthroplasty for primary glenohumeral osteoarthritis in patients with a biconcave glenoid. J Bone Joint Surg Am 2013;95:1297-304. http://dx. doi.org/10.2106/JBJS.L.00820 33. Muir SW, Corea CL, Beaupre L. Evaluating change in clinical status: reliability and measures of agreement for the assessment of glenohumeral range of motion. N Am J Sports Phys Ther 2010;5: 98-110. 34. Myers JB, Laudner KG, Pasquale MR, Bradley JP, Lephart SM. Glenohumeral range of motion deficits and posterior shoulder tightness in throwers with pathologic internal impingement. Am J Sports Med 2006;34:385-91. http://dx.doi.org/10.1177/03635465052 81804 35. Myers JB, Oyama S, Wassinger CA, Ricci RD, Abt JP, Conley KM, et al. Reliability, precision, accuracy, and validity of posterior shoulder tightness assessment in overhead athletes. Am J Sports Med 2007;35: 1922-30. http://dx.doi.org/10.1177/0363546507304142 36. O’Neill BJ, O’Briain D, Hirpara KM, Shaughnesy M, Yeatman EA, Kaar TK. Digital photography for assessment of shoulder range of motion: a novel clinical and research tool. Int J Shoulder Surg 2013;7: 23-7. http://dx.doi.org/10.4103/0973-6042.109888 37. Omi R, Hooke AW, Zhao KD, Matsuhashi T, Goto A, Yamamoto N, et al. The effect of the remplissage procedure on shoulder range of motion: a cadaveric study. Arthroscopy 2014;30:178-87. http://dx.doi. org/10.1016/j.arthro.2013.11.003 38. Pearl ML, Harris SL, Lippitt SB, Sidles JA, Harryman DT 2nd, Matsen FA 3rd. A system for describing positions of the humerus relative to the thorax and its use in the presentation of several functionally important arm positions. J Shoulder Elbow Surg 1992;1: 113-8. 39. Raiss P, Schmitt M, Bruckner T, Kasten P, Pape G, Loew M, et al. Results of cemented total shoulder replacement with a minimum follow-up of ten years. J Bone Joint Surg Am 2012;94. http://dx.doi. org/10.2106/JBJS.K.00580. e1711-10. 40. Riddle DL, Rothstein JM, Lamb RL. Goniometric reliability in a clinical setting. Shoulder measurements. Phys Ther 1987;67:668-73. 41. Schouten B, Fedtke S, Bekker T, Schijven M, Gekker A. Games for health. Proceedings of the 3rd European conference on gaming and playful interaction in health care. Wiesbaden: Springer Vieweg; 2013. p. 1-293.
8 42. Torres RR, Gomes JL. Measurement of glenohumeral internal rotation in asymptomatic tennis players and swimmers. Am J Sports Med 2009;37:1017-23. http://dx.doi.org/10.1177/0363546508329544 43. Triffitt PD. The relationship between motion of the shoulder and the stated ability to perform activities of daily living. J Bone Joint Surg Am 1998;80:41-6. 44. Tyler TF, Nicholas SJ, Lee SJ, Mullaney M, McHugh MP. Correction of posterior shoulder tightness is associated with symptom resolution in patients with internal impingement. Am J Sports Med 2010;38: 114-9. http://dx.doi.org/10.1177/0363546509346050 45. Vairo GL, Duffey ML, Owens BD, Cameron KL. Clinical descriptive measures of shoulder range of motion for a healthy, young and physically active cohort. Sports Med Arthrosc Rehabil Ther Technol 2012;4:33. http://dx.doi.org/10.1186/1758-2555-4-33 46. van de Pol RJ, van Trijffel E, Lucas C. Inter-rater reliability for measurement of passive physiological range of motion of upper extremity joints is better if instruments are used: a systematic review. J Physiother 2010;56:7-17. http://dx.doi.org/10.1016/S1836-9553(10) 70049-7
F.A. Matsen et al. 47. Webster D, Celik O. Experimental evaluation of Microsoft Kinect’s accuracy and capture rate for stroke rehabilitaton applications. Houston, TX: Institute of Electrical and Electronics Engineers (IEEE); 2014. p. 455-60. 48. Werner BC, Holzgrefe RE, Griffin JW, Lyons ML, Cosgrove CT, Hart JM, et al. Validation of an innovative method of shoulder rangeof-motion measurement using a smartphone clinometer application. J Shoulder Elbow Surg 2014;23:e275-82. http://dx.doi.org/10.1016/j.jse. 2014.02.030 49. Wilk KE, Macrina LC, Fleisig GS, Porterfield R, Simpson CD, Harker P, et al. Correlation of glenohumeral internal rotation deficit and total rotational motion to shoulder injuries in professional baseball pitchers. Am J Sports Med 2011;39:329-35. http://dx.doi.org/10.1177/ 0363546510384223 50. Young AA, Walch G, Pape G, Gohike F, Favard L. Secondary rotator cuff dysfunction following total shoulder arthroplasty for primary glenohumeral osteoarthritis: results of a multicenter study with more than five years of follow-up. J Bone Joint Surg Am 2012;94:685-93. http://dx.doi.org/10.2106/JBJS.J.00727