Concurrent validity and test-retest reliability for the assessment of body segment lengths using two-dimensional video-recordings in weightlifting

Figures & data

Figure 1. Methodical procedure.

Figure 2. Clean (top series) and jerk (bottom series) from start position to catch position.

Table 1. Body segments and their adjacent anatomical landmarks used for segment length measurements.

Segment	Joint 1	Joint 2
Upper arm	Acromion	Elbow joint center
Forearm	Elbow joint center	Wrist joint center
Trunk	Trochanter major	Acromion
Thigh	Knee joint center	Trochanter major
Shank	Ankle joint center	Knee joint center

Figure 3. Anatomical landmarks and plate diameter assessment in the 2D analysis.

Figure 4. Differences of the segment lengths between 2D and 3D assessment during the clean: red lines = vertical barbell velocity (vertical red lines = left - barbell lift off, right - maximum vertical barbell velocity), black lines in top graph: dotted line = thigh, dashed line = shank, black lines in bottom graph: dotted line = upper arm, dashed line = forearm, solid line =trunk.

Figure 5. Differences of the segment lengths between 2D and 3D assessment during the jerk: red lines = vertical barbell velocity (vertical red lines = left - barbell lift off, right - maximum vertical barbell velocity), black lines in top graph: dotted line = thigh, dashed line = shank, black lines in bottom graph: dotted line = upper arm, dashed line = forearm, solid line =trunk.

Table 2. Statistical measures of concurrent validity of 2D segment length measurements. Bold data represent smallest absolute measurement error for each body segment.

	CCC (95% CL)	Intercept (95% CL)	Slope (95% CL)	SDD	SDD%	lLOA (95% CL)	uLOA (95% CL)
Shank Clean vmax	0.81 (0.66;0.90)	5.76 (−4.1;13.2)	.90 (.72;1.14)	1.61	3.94	−1.51 (−2.4;-0.6)	4.82 (3.9;5.8)
Shank Jerk Start	0.88 (0.76;0.94)	1.39 (−5.3;7.9)	.99 (.84;1.15)	1.42	3.50	−1.79 (−2.6;-1.0)	3.78 (3.0;4.6)
Shank Jerk vmax	0.86 (0.71;0.93)	3.88 (−5.2;9.7)	.92 (.78;1.14)	1.67	4.05	−2.61 (−3.6;-1.7)	3.95 (3.0;4.9)
Thigh Clean vmax	0.79 (0.60;0.90)	1.30 (−11.7;1.6)	1.00 (.76;1.33)	2.39	6.07	−3.42 (−4.8;-2.0)	5.95 (4.6;7.3)
Thigh Jerk Start	0.70 (0.49;0.85)	−.53 (−14.6;11.0)	1.06 (.76;1.41)	2.23	5.58	−2.55 (−3.8;-1.3)	6.19 (4.9;7.5)
Thigh Jerk vmax	0.75 (0.55;0.87)	4.29 (−5.7;12.9)	.94 (.73;1.18)	2.44	5.92	−2.85 (−4.3;-1.5)	6.71 (5.3;8.1)
Trunk Clean Start	0.51 (0.28;0.68)	−4.49 (−33.1;7.8)	1.17 (.93;1.78)	2.89	3.22	−1.79 (−3.5;-0.1)	9.55 (7.9;11.2)
Trunk Clean vmax	0.76 (0.60;0.86)	7.50 (−4.0;15.4)	.92 (.79;1.13)	1.87	3.19	−0.62 (−1.7;-0.5)	6.69 (5.6;7.8)
Trunk Jerk Start	0.78 (0.63;0.87)	−.78 (−6.2;4.9)	1.06 (.96;1.16)	1.45	2.58	−0.33 (−1.2;0.5)	5.36 (4.5;6.2)
Trunk Jerk vmax	0.82 (0.68;0.91)	.40 (−9.5;7.2)	1.03 (.91;1.21)	1.83	3.27	−1.47 (−2.5;-0.4)	5.70 (4.7;6.8)
Upper arm Clean Start	0.64 (0.36;0.82)	−.45 (−1.2;9.3)	.99 (.67;1.30)	1.73	5.53	−4.31 (−5.3;-3.3)	2.46 (1.4;3.5)
Forearm Clean Start	0.82 (0.66;0.91)	5.09 (.6;8.4)*	.81 (.68;.98)**	1.16#	4.19	-2.27# (−2.8;-1.7)#	2.27# (1.7;2.8)#
Forearm Jerk Start	0.57 (0.30;0.75)	5.18 (−6.5;11.4)	.77 (.57;1.14)	2.16	7.24	−6.06 (−7.3;-4.8)	2.42 (1.2;3.7)
Forearm Jerk vmax	0.70 (0.45;0.84)	6.90 (−1.5;13.8)	.74 (.50;1.01)	2.26	7.74	−5.36 (−6.7;-4.0)	3.49 (2.2;4.8)

Notes. CCC = concordance correlation coefficient, SDD = standard deviation of differences (cm), SDD% = standard deviation of differences (%), lLOA = lower limits of agreement (cm), uLOA = upper limits of agreement (cm), CL = confidence limits, * intercept significant different from 0, ** slope significant different from 1, # re-calibrated value.

Figure 6. Deming and Bland-Altman analysis of the agreement of 2D and 3D segment lengths. Only comparison with the best concurrent validity results per segment are shown. Deming regression is shown with fitted linear model (dashed line) and the identity line (2D = 3D, slope = 1; solid line). Bland-Altman plots show the lower and upper limits of agreement (dashed lines) and the systematic bias (solid lines). Bland-Altman plots of shank, thigh, trunk and upper arm show “raw” data, whereas re-calibrated data from Deming regression is shown in the Bland-Altman plot of the forearm, since significant, constant and proportional bias was found.

Table 3. Measures of test-retest reliability in 2D measurements.

	ICC (95% CL)	SEM (in cm) (95% CL)	SEM% (in %) (95% CL)
Shank Jerk Start	0.92 (0.81;0.96)	0.93 (0.73;1.29)	2.31 (1.8;3.2)
Thigh Jerk Start	0.85 (0.73;0.92)	1.37 (1.08;1.90)	3.51 (2.8;4.9)
Trunk Jerk Start	0.96 (0.91;0.98)	0.83 (0.65;1.14)	1.51 (1.2;2.1)
Upper arm Clean Start	0.88 (0.75;0.95)	0.80 (0.63;1.10)	2.52 (2.0;3.5)
Forearm Clean Start	0.96 (0.92;0.98)	0.52 (0.41;0.72)	1.86 (1.5;2.6)

Notes. ICC = intraclass correlation coefficient, SEM = standard error of measurement, SEM% = percentage standard error of measurement.

Data availability statement

The data presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.