Comparison of Recording Positions of Physical Activity in Patients with Severe COPD Undergoing LTOT

Figures & data

Table 1.  Characteristics of patients

At the start of pulmonary rehabilitation
	All patients (N = 52)	Group A (N = 21)	Group B (N = 31)	P-value
Age (y)	62.2 ± 9.6	66.9 ± 6.6	59.0 ± 9.5	<0.01
Sex (F/M)	22/30	10/11	12/19
BMI (kg/m2)	25.5 ± 7.0	23.4 ± 3.4	26.9 ± 8.3	Ns
FEV1%pred.	38.0 ± 11.8	38.5 ± 11.2	37.7 ± 12.4	Ns
FEV1%VC	44.9 ± 13.0	49.1 ± 16.2	41.9 ± 9.8	Ns
6MWD (m)	265.4 ± 93.7 (N = 37)	218.6 ± 85.2 (N = 19)	314.8 ± 76.6 (N = 18)	<0.01
Borg CR 10	5.2 ± 2.1 (N = 36)	5.8 ± 2.2(N = 18)	4.6 ± 1.9 (N = 18)	Ns
MMRC	3.0 ± 1.0 (N = 40)	3.4 ± 0.8 (N = 20)	2.6 ± 1.1 (N = 20)	<0.05
BODE Score	6.2 ± 1.9 (N = 37)	6.7 ± 1.7 (N = 19)	5.6 ± 1.9 (N = 18)	Ns

Data expressed as mean ± SD. BMI, Body Mass Index; PR, pulmonary rehabilitation; MMRC, modified Medical Research Council (MMRC) dyspnea scale; 6MWD, 6-Minute Walk Distance; Borg CR 10, patients’ self-perceived degree of breathlessness at the end of the 6MWD using the modified Borg CR 10 scale; FEV1, forced expiratory volume in 1 s; VC, vital capacity. Ns, not significant; The p-values refer to the difference between the two subgroups (see text).

Figure 1.  Box plots of the mean physical activity for the entire measurement period assessed at ankle and wrist (I) and hip (II) in patients with (group A) and without (group B) walkers.

Table 2.  Measurements of activity obtained by the GT1M (wrist, ankle) and the RT3 (hip) in patients with and without walker

	Group A: Walker patients Median (N) Interquartile range	Group B: Non-Walker patients Median (N) Interquartile range
Ankle activity counts	215.6 (19)145.7 to 523.0	438.1 (31)333.1 to 655.3
Wrist activity counts	561.4 (20)431.2 to 733.8	834.0 (31)588.9 to 1062.9
Hip activity VMU	72.0 (20)53.9 to 116.9	141.9 (28)96.2 to 192.6

Median values and interquartile ranges are given including the number of cases as N.

Table 3.  Correlations between activity data from different monitor positions in patients with and without walker

	Group A: Walker patients Spearman's rank coefficient r (N)	Group B: Non-Walker patients Spearman's rank coefficient r (N)
Ankle/ Wrist activity	r = 0.13 (18)95% CI –0.42 to 0.6p = 0.63	r = 0.49 (31)95% CI 0.13 to 0.76p < 0.01
Ankle/Hip activity	r = 0.73 (19) 95% CI 0.32 to 0.92p < 0.01	r = 0.65 (28)95% CI 0.38 to 0.81p < 0.01
Wrist/Hip activity	r = 0.33 (19) 95% CI –0.22 to 0.73p = 0.17	r = 0.44 (28)95% CI 0.03 to 0.75p < 0.05

Spearman's rank correlation coefficients are given as r including the number of cases as N, the corresponding 95% confidence intervals (CI) and the respective p-values.

Figure 2.  Relationship between the mean PA values assessed at the ankle and the 6MWD assessed at the beginning and end of the PR program in both patient groups.

Table 4.  Standard multivariate regression model to assess the impact of group and gender and all simultaneously measured PA positions, on the disease severity as described by the BODE score

Coefficients^a
	Unstandardized Coefficients		Standardized Coefficients			Correlations
	B	Std. Error	Beta	t	Sig.	Zero-order	Partial	Part	Exp. Var.
(Constant)	9.237	.971		9.508	.000
Hip activity	.009	.008	.304	1.238	.227	–.459	.240	.174	3%
Ankle activity	–.005	.002	–.648	–2.881	.008	–.638	–.499	–.405	16%
Wrist activity	–.002	.001	–.359	–2.018	.054	–.520	–.374	–.284	8%
Gender	–.785	.621	–.195	–1.264	.218	–.241	–.245	–.178	3%
Group	.044	.598	.011	.073	.942	–.265	.015	.010	0%
a Dependent variable: BODE Score

Model Summary
F	Sig.	R^2	Adjusted R^2	Std. Error of the Estimate
5.096	.002	.505	.406	1.508

Table

Imputation Results
Imputation Method		Fully Conditional Specification
Fully Conditional Specification Method Iterations		10
Dependent Variables	Imputed	@6MWDbegin,@6MWDende,Leg_Activity,Hip_Activity,Arm_Activity
	Not Imputed(Too Many Missing Values)
	Not Imputed(No Missing Values)	Sex,Age,FEV1PercentPred,Group
Imputation Sequence		Sex,Age,FEV1PercentPred,Group,Arm_Activity,Leg_Activity,Hip_Activity,@6MWDbegin,@6MWDende

Table

Correlations for Non-Walker Patients after Imputation (Group B)
Imputation Number				6MWD start of PR	6MWD end of PR	Arm activity	Leg activity	Hip activity
Pooled	Spearman's rho	6MWD start of PR	Correlation Coefficient	1.000	.752	.278	.528	.412
			Sig. (2-tailed)		.000	.199	.013	.041
			N	31	31	31	31	31
		6MWD end of PR	Correlation Coefficient	.752	1.000	.271	.469	.347
			Sig. (2-tailed)	.000		.267	.047	.132
			N	31	31	31	31	31
		Arm activity	Correlation Coefficient	.278	.271	1.000	.490	.433
			Sig. (2-tailed)	.199	.267		.005	.015
			N	31	31	31	31	31
		Leg activity	Correlation Coefficient	.528	.469	.490	1.000	.670
			Sig. (2-tailed)	.013	.047	.005		.000
			N	31	31	31	31	31
		Hip activity	Correlation Coefficient	.412	.347	.433	.670	1.000
			Sig. (2-tailed)	.041	.132	.015	.000
			N	31	31	31	31	31

Table

Imputation Number				6MWD start of PR	6MWD end of PR	Arm activity	Leg activity	Hip activity
Original data	Spearman's rho	6MWD start of PR	Correlation Coefficient	1.000	.683	.346	.603	.406
			Sig. (2-tailed)	.	.002	.159	.008	.133
			N	18	18	18	18	15
		6MWD end of PR	Correlation Coefficient	.683	1.000	.202	.531	.339
			Sig. (2-tailed)	.002	.	.421	.023	.216
			N	18	18	18	18	15
		Arm activity	Correlation Coefficient	.346	.202	1.000	.490	.444
			Sig. (2-tailed)	.159	.421	.	.005	.018
			N	18	18	31	31	28
		Leg activity	Correlation Coefficient	.603	.531	.490	1.000	.653
			Sig. (2-tailed)	.008	.023	.005	.	.000
			N	18	18	31	31	28
		Hip activity	Correlation Coefficient	.406	.339	.444	.653	1.000
			Sig. (2-tailed)	.133	.216	.018	.000	.
			N	15	15	28	28	28

Table

Imputation Number				6MWD start of PR	6MWD end of PR	Arm activity	Leg activity	Hip activity
Pooled	Spearman's rho	6MWD start of PR	Correlation Coefficient	1.000	.904	.353	.606	.301
			Sig. (2-tailed)		.000	.130	.008	.260
			N	21	21	21	21	21
		6MWD end of PR	Correlation Coefficient	.904	1.000	.454	.629	.296
			Sig. (2-tailed)	.000		.044	.005	.249
			N	21	21	21	21	21
		Arm activity	Correlation Coefficient	.353	.454	1.000	.255	.303
			Sig. (2-tailed)	.130	.044		.279	.190
			N	21	21	21	21	21
		Leg activity	Correlation Coefficient	.606	.629	.255	1.000	.727
			Sig. (2-tailed)	.008	.005	.279		.000
			N	21	21	21	21	21
		Hip activity	Correlation Coefficient	.301	.296	.303	.727	1.000
			Sig. (2-tailed)	.260	.249	.190	.000
			N	21	21	21	21	21

Table

Imputation Number				6MWD start of PR	6MWD end of PR	Arm activity	Leg activity	Hip activity
Original data	Spearman's rho	6MWD start of PR	Correlation Coefficient	1.000	.917	.375	.545	.229
			Sig. (2-tailed)	.	.000	.126	.024	.361
			N	19	19	18	17	18
		6MWD end of PR	Correlation Coefficient	.917	1.000	.463	.625	.278
			Sig. (2-tailed)	.000	.	.053	.007	.264
			N	19	19	18	17	18
		Arm activity	Correlation Coefficient	.375	.463	1.000	.125	.332
			Sig. (2-tailed)	.126	.053	.	.622	.166
			N	18	18	20	18	19
		Leg activity	Correlation Coefficient	.545	.625	.125	1.000	.726
			Sig. (2-tailed)	.024	.007	.622	.	.000
			N	17	17	18	19	19
		Hip activity	Correlation Coefficient	.229	.278	.332	.726	1.000
			Sig. (2-tailed)	.361	.264	.166	.000	.
			N	18	18	19	19	20

Results: No great changes in the correlation coefficient r could be observed between the original data with reduced sample size due to missing values and the imputed table output in both groups.