Checking dataset / cleaning procedures
The database was initially explored for consistency and coding information.
Três aspectos são fundamentais em qualquer conjunto de dados, que são: comunicabilidade dos nomes das variáveis, confiabilidade e consistência dos dados. A primeira refere-se à nomeação das variáveis, a segunda refere-se às garantias que o processo de coleta foi adequado e a terceira refere-se à condição de análise dos dados. Empiricamente, é possível verificar mais diretamente à consistência da base a partir de técnicas gráficas, tabulares e analíticas.
Mixed effects
A linear mixed model was computed to check the effect of the factors on the outcome. This model also included an interaction term (time x group) and included random intercepts for each participant.
Random intercept means that each participant was assigned a different intercept value. Therefore, the individual differences of the participants were accounted for in the model. The computational equation is:
\[acq_6 = group + time + (group \ x \ time) + (1 | participant)\]
The table below reports the results. Since the interaction term was significant (F(2, 87.675) = 5.89, p < 0.01), the main effect of time was not interpreted.
|
|
Sum Sq
|
Mean Sq
|
NumDF
|
DenDF
|
F value
|
Pr(>F)
|
|
group
|
0.15
|
0.15
|
1
|
52
|
0.44
|
0.511
|
|
time
|
5.73
|
2.87
|
2
|
89
|
8.20
|
0.001
|
|
group:time
|
4.12
|
2.06
|
2
|
89
|
5.90
|
0.004
|
To check the differences between all pairwise comparions, a post-hoc procedure was carried-out. It is widely know that risk of Type I errors will be increased by performing comparions. Thus, Bonferroni method was modeled to adjust the p-values derived from all comparisons.
Detailing the results, the interaction between “Treatment x Time” was explored to check the group differences over time and the time differences among the two groups. In the aerobic training group (AG), the ACQ6 score decreased significantly from 2.0 (1.6-2.3 95% confidence interval (CI)) at the baseline to 1.3 (0.9-1.6 95% CI) after the intervention and to 1.2 (0.8 to 1.5 95% CI) at 3 months of follow-up (P < 0.001 for both after the intervention and at 3 months of follow-up versus the baseline). In the breathing exercises group (BG), the ACQ6 score showed no significant change when comparing the baseline score of 1.7 (1.2 to 2.0 95% CI) to 1.5 after the intervention (1.0 to 1.8 95% CI; P = 0.71 vs. baseline)) and to 1.7 at 3 months of follow-up (1.3 to 2.1 95% CI; P =1.0 vs. baseline)).
|
time
|
group
|
emmean
|
SE
|
df
|
lower.CL
|
upper.CL
|
|
Baseline
|
AG
|
2.0
|
0.17
|
89
|
1.65
|
2.3
|
|
Post-intervention
|
AG
|
1.3
|
0.17
|
96
|
0.94
|
1.6
|
|
3-month follow-up
|
AG
|
1.2
|
0.18
|
110
|
0.80
|
1.5
|
|
Baseline
|
BG
|
1.7
|
0.18
|
89
|
1.31
|
2.0
|
|
Post-intervention
|
BG
|
1.5
|
0.18
|
92
|
1.10
|
1.8
|
|
3-month follow-up
|
BG
|
1.7
|
0.20
|
111
|
1.33
|
2.1
|
|
|
contrast
|
group
|
estimate
|
SE
|
df
|
t.ratio
|
p.value
|
|
Baseline - Post-intervention
|
AG
|
0.69
|
0.16
|
88
|
4.26
|
0.00
|
|
Baseline - 3-month follow-up
|
AG
|
0.81
|
0.18
|
90
|
4.61
|
0.00
|
|
Post-intervention - 3-month follow-up
|
AG
|
0.12
|
0.18
|
88
|
0.67
|
1.00
|
|
Baseline - Post-intervention
|
BG
|
0.21
|
0.17
|
86
|
1.22
|
0.68
|
|
Baseline - 3-month follow-up
|
BG
|
-0.06
|
0.19
|
89
|
-0.32
|
1.00
|
|
Post-intervention - 3-month follow-up
|
BG
|
-0.27
|
0.19
|
88
|
-1.41
|
0.49
|
|
Overall, the main difference between baseline and post-intervention (Δ = .448, p = 0.001), and between the baseline and the follow-up (3 months after) (Δ = .373, p = 0.014) were significant. The comparison between the post-intervention and the 3 months follow-up was not significant (Δ = - 0.075, p = 1).
|
time
|
emmean
|
SE
|
df
|
lower.CL
|
upper.CL
|
|
Baseline
|
1.8
|
0.12
|
89
|
1.6
|
2.1
|
|
Post-intervention
|
1.4
|
0.12
|
94
|
1.1
|
1.6
|
|
3-month follow-up
|
1.4
|
0.14
|
110
|
1.2
|
1.7
|
|
|
contrast
|
estimate
|
SE
|
df
|
t.ratio
|
p.value
|
|
Baseline - Post-intervention
|
0.45
|
0.12
|
87
|
3.82
|
0.00
|
|
Baseline - 3-month follow-up
|
0.37
|
0.13
|
90
|
2.91
|
0.01
|
|
Post-intervention - 3-month follow-up
|
-0.07
|
0.13
|
88
|
-0.58
|
1.00
|
|
|
group
|
time
|
emmean
|
SE
|
df
|
lower.CL
|
upper.CL
|
|
AG
|
Baseline
|
2.0
|
0.17
|
89
|
1.65
|
2.3
|
|
BG
|
Baseline
|
1.7
|
0.18
|
89
|
1.31
|
2.0
|
|
AG
|
Post-intervention
|
1.3
|
0.17
|
96
|
0.94
|
1.6
|
|
BG
|
Post-intervention
|
1.5
|
0.18
|
92
|
1.10
|
1.8
|
|
AG
|
3-month follow-up
|
1.2
|
0.18
|
110
|
0.80
|
1.5
|
|
BG
|
3-month follow-up
|
1.7
|
0.20
|
111
|
1.33
|
2.1
|
|
|
contrast
|
time
|
estimate
|
SE
|
df
|
t.ratio
|
p.value
|
|
AG - BG
|
Baseline
|
0.31
|
0.24
|
89
|
1.26
|
0.209
|
|
AG - BG
|
Post-intervention
|
-0.18
|
0.25
|
94
|
-0.71
|
0.481
|
|
AG - BG
|
3-month follow-up
|
-0.56
|
0.27
|
110
|
-2.07
|
0.041
|
|


Fifty-eight percentage of participants from the AG and 28% of BG participants showed a clinically significant improvement in the ACQ6 score (≥0.5 points)
|
group
|
mean(imp_acq6)
|
|
AG
|
0.59
|
|
BG
|
0.28
|
|
group
|
mean(imp_acq6)
|
|
AG
|
0.38
|
|
BG
|
0.08
|
Asma day-free [GEE]

## Analysis of 'Wald statistic' Table
## Model: poisson, link: log
## Response: dias_sintomas
## Terms added sequentially (first to last)
##
## Df X2 P(>|Chi|)
## grupo_15 1 0.15 0.6960
## tempo_16 3 15.81 0.0012 **
## grupo_15:tempo_16 3 4.10 0.2509
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
|
term
|
estimate
|
std.error
|
statistic
|
p.value
|
conf.low
|
conf.high
|
|
(Intercept)
|
2.28
|
0.21
|
122.32
|
0.00
|
1.87
|
2.68
|
|
grupo_15
|
-0.25
|
0.31
|
0.63
|
0.43
|
-0.86
|
0.36
|
|
tempo_162meses
|
0.21
|
0.11
|
3.75
|
0.05
|
0.00
|
0.42
|
|
tempo_163meses
|
0.20
|
0.08
|
5.87
|
0.02
|
0.04
|
0.36
|
|
tempo_16PRE
|
-0.44
|
0.18
|
5.77
|
0.02
|
-0.81
|
-0.08
|
|
grupo_15:tempo_162meses
|
0.05
|
0.19
|
0.07
|
0.80
|
-0.32
|
0.42
|
|
grupo_15:tempo_163meses
|
0.13
|
0.19
|
0.47
|
0.49
|
-0.24
|
0.51
|
|
grupo_15:tempo_16PRE
|
0.46
|
0.23
|
3.97
|
0.05
|
0.01
|
0.92
|
Activity
|
|
Sum Sq
|
Mean Sq
|
NumDF
|
DenDF
|
F value
|
Pr(>F)
|
|
group
|
0.28
|
0.28
|
1
|
50
|
0.43
|
0.51
|
|
time
|
10.45
|
5.23
|
2
|
88
|
8.01
|
0.00
|
|
group:time
|
4.48
|
2.24
|
2
|
88
|
3.44
|
0.04
|
Score total
|
|
Sum Sq
|
Mean Sq
|
NumDF
|
DenDF
|
F value
|
Pr(>F)
|
|
group
|
0.03
|
0.03
|
1
|
49
|
0.05
|
0.82
|
|
time
|
11.14
|
5.57
|
2
|
88
|
8.52
|
0.00
|
|
group:time
|
1.58
|
0.79
|
2
|
88
|
1.21
|
0.30
|
|
time
|
group
|
emmean
|
SE
|
df
|
lower.CL
|
upper.CL
|
|
Baseline
|
AG
|
3.9
|
0.21
|
98
|
3.4
|
4.3
|
|
Post-intervention
|
AG
|
4.6
|
0.21
|
98
|
4.2
|
5.1
|
|
3-month follow-up
|
AG
|
4.6
|
0.24
|
115
|
4.2
|
5.1
|
|
Baseline
|
BG
|
4.2
|
0.22
|
98
|
3.7
|
4.6
|
|
Post-intervention
|
BG
|
4.4
|
0.22
|
100
|
4.0
|
4.9
|
|
3-month follow-up
|
BG
|
4.7
|
0.24
|
115
|
4.2
|
5.2
|
|
|
contrast
|
group
|
estimate
|
SE
|
df
|
t.ratio
|
p.value
|
|
Baseline - Post-intervention
|
AG
|
-0.78
|
0.22
|
85
|
-3.46
|
0.00
|
|
Baseline - 3-month follow-up
|
AG
|
-0.79
|
0.25
|
89
|
-3.21
|
0.01
|
|
Post-intervention - 3-month follow-up
|
AG
|
-0.01
|
0.25
|
89
|
-0.05
|
1.00
|
|
Baseline - Post-intervention
|
BG
|
-0.27
|
0.23
|
86
|
-1.18
|
0.72
|
|
Baseline - 3-month follow-up
|
BG
|
-0.53
|
0.25
|
90
|
-2.09
|
0.12
|
|
Post-intervention - 3-month follow-up
|
BG
|
-0.25
|
0.25
|
89
|
-1.00
|
0.96
|
|

Shuttel Distance
|
|
Sum Sq
|
Mean Sq
|
NumDF
|
DenDF
|
F value
|
Pr(>F)
|
|
group_shuttle
|
4.1e-01
|
4.1e-01
|
1
|
52
|
0.0
|
0.99
|
|
time_shuttle
|
1.2e+05
|
1.2e+05
|
1
|
48
|
71.1
|
0.00
|
|
group_shuttle:time_shuttle
|
7.4e+03
|
7.4e+03
|
1
|
48
|
4.4
|
0.04
|
The results show the interaction effect was significant (F(1, 47.89) = 4.37, p = 0.042), and also the main effect of time (F(1, 47.98) = 59.02, p < 0.01). Group was not significant (F(1, 52.27) = 0.01, p = 0.98). For participants in the AG, the mean ISWT distance improved from baseline 342 (from 302 to 382 95% CI) to 429 meters (389 to 470 95% CI (P < 0.01, post hoc pairwised comparision after vs baseline). In the breathing exercise group (BG), from 360 to 412 (369 to 456 95% CI) after the intervention (P < 0.01, post hoc pairwised comparision after vs baseline).

The adjusted mean difference in the walking distance between the AG and the BG was 17.8 meters at the baseline and 17 after intervention.
|
time_shuttle
|
group_shuttle
|
emmean
|
SE
|
df
|
lower.CL
|
upper.CL
|
|
Baseline
|
AG
|
342
|
20
|
59
|
302
|
382
|
|
Post-intervention
|
AG
|
429
|
20
|
62
|
389
|
470
|
|
Baseline
|
BG
|
360
|
22
|
59
|
317
|
403
|
|
Post-intervention
|
BG
|
412
|
22
|
62
|
369
|
456
|
|
|
contrast
|
group_shuttle
|
estimate
|
SE
|
df
|
t.ratio
|
p.value
|
|
Baseline - Post-intervention
|
AG
|
-88
|
11
|
48
|
-7.7
|
0
|
|
Baseline - Post-intervention
|
BG
|
-53
|
12
|
48
|
-4.3
|
0
|
|
|
group_shuttle
|
time_shuttle
|
emmean
|
SE
|
df
|
lower.CL
|
upper.CL
|
|
AG
|
Baseline
|
342
|
20
|
59
|
302
|
382
|
|
BG
|
Baseline
|
360
|
22
|
59
|
317
|
403
|
|
AG
|
Post-intervention
|
429
|
20
|
62
|
389
|
470
|
|
BG
|
Post-intervention
|
412
|
22
|
62
|
369
|
456
|
|
|
contrast
|
time_shuttle
|
estimate
|
SE
|
df
|
t.ratio
|
p.value
|
|
AG - BG
|
Baseline
|
-18
|
29
|
59
|
-0.61
|
0.55
|
|
AG - BG
|
Post-intervention
|
17
|
30
|
62
|
0.57
|
0.57
|
|
Hospital Anxiety and Depression Scale (HADS):
|
|
Sum Sq
|
Mean Sq
|
NumDF
|
DenDF
|
F value
|
Pr(>F)
|
|
group
|
1.0
|
1
|
1
|
52
|
0.18
|
0.67
|
|
time
|
28.9
|
14
|
2
|
81
|
2.57
|
0.08
|
|
group:time
|
7.9
|
4
|
2
|
81
|
0.70
|
0.50
|
There were no interactions nor difference between the groups in both domains (Table 2). Pairwise comparision over time revealed a significant decrease after intervention and follow up for both groups with regard to depression.
|
group
|
time
|
emmean
|
SE
|
df
|
lower.CL
|
upper.CL
|
|
AG
|
Baseline
|
8.7
|
0.76
|
75
|
7.2
|
10.2
|
|
BG
|
Baseline
|
8.6
|
0.83
|
75
|
6.9
|
10.2
|
|
AG
|
Post-intervention
|
7.5
|
0.79
|
84
|
5.9
|
9.0
|
|
BG
|
Post-intervention
|
7.7
|
0.85
|
80
|
6.0
|
9.4
|
|
AG
|
3-month follow-up
|
7.8
|
0.87
|
102
|
6.1
|
9.5
|
|
BG
|
3-month follow-up
|
9.0
|
0.93
|
98
|
7.2
|
10.8
|
|
|
contrast
|
time
|
estimate
|
SE
|
df
|
t.ratio
|
p.value
|
|
AG - BG
|
Baseline
|
0.11
|
1.1
|
75
|
0.09
|
0.92
|
|
AG - BG
|
Post-intervention
|
-0.23
|
1.2
|
82
|
-0.20
|
0.84
|
|
AG - BG
|
3-month follow-up
|
-1.18
|
1.3
|
100
|
-0.93
|
0.36
|
|
|
time
|
group
|
emmean
|
SE
|
df
|
lower.CL
|
upper.CL
|
|
Baseline
|
AG
|
8.7
|
0.76
|
75
|
7.2
|
10.2
|
|
Post-intervention
|
AG
|
7.5
|
0.79
|
84
|
5.9
|
9.0
|
|
3-month follow-up
|
AG
|
7.8
|
0.87
|
102
|
6.1
|
9.5
|
|
Baseline
|
BG
|
8.6
|
0.83
|
75
|
6.9
|
10.2
|
|
Post-intervention
|
BG
|
7.7
|
0.85
|
80
|
6.0
|
9.4
|
|
3-month follow-up
|
BG
|
9.0
|
0.93
|
98
|
7.2
|
10.8
|
|
|
contrast
|
group
|
estimate
|
SE
|
df
|
t.ratio
|
p.value
|
|
Baseline - Post-intervention
|
AG
|
1.23
|
0.66
|
80
|
1.85
|
0.20
|
|
Baseline - 3-month follow-up
|
AG
|
0.86
|
0.75
|
82
|
1.15
|
0.76
|
|
Post-intervention - 3-month follow-up
|
AG
|
-0.37
|
0.76
|
80
|
-0.49
|
1.00
|
|
Baseline - Post-intervention
|
BG
|
0.89
|
0.71
|
79
|
1.26
|
0.64
|
|
Baseline - 3-month follow-up
|
BG
|
-0.42
|
0.80
|
81
|
-0.53
|
1.00
|
|
Post-intervention - 3-month follow-up
|
BG
|
-1.31
|
0.80
|
80
|
-1.64
|
0.32
|
|


|
|
Sum Sq
|
Mean Sq
|
NumDF
|
DenDF
|
F value
|
Pr(>F)
|
|
group
|
4.945
|
4.945
|
1
|
52
|
1.100
|
0.299
|
|
time
|
60.160
|
30.080
|
2
|
80
|
6.692
|
0.002
|
|
group:time
|
0.019
|
0.009
|
2
|
80
|
0.002
|
0.998
|
|
group
|
time
|
emmean
|
SE
|
df
|
lower.CL
|
upper.CL
|
|
AG
|
Baseline
|
7.2
|
0.77
|
69
|
5.6
|
8.7
|
|
BG
|
Baseline
|
8.2
|
0.84
|
69
|
6.6
|
9.9
|
|
AG
|
Post-intervention
|
5.6
|
0.80
|
76
|
4.0
|
7.1
|
|
BG
|
Post-intervention
|
6.7
|
0.86
|
73
|
5.0
|
8.4
|
|
AG
|
3-month follow-up
|
6.3
|
0.86
|
93
|
4.6
|
8.0
|
|
BG
|
3-month follow-up
|
7.4
|
0.92
|
89
|
5.6
|
9.2
|
|
|
contrast
|
time
|
estimate
|
SE
|
df
|
t.ratio
|
p.value
|
|
AG - BG
|
Baseline
|
-1.1
|
1.1
|
69
|
-0.94
|
0.35
|
|
AG - BG
|
Post-intervention
|
-1.1
|
1.2
|
74
|
-0.96
|
0.34
|
|
AG - BG
|
3-month follow-up
|
-1.1
|
1.3
|
90
|
-0.89
|
0.37
|
|
|
time
|
group
|
emmean
|
SE
|
df
|
lower.CL
|
upper.CL
|
|
Baseline
|
AG
|
7.2
|
0.77
|
69
|
5.6
|
8.7
|
|
Post-intervention
|
AG
|
5.6
|
0.80
|
76
|
4.0
|
7.1
|
|
3-month follow-up
|
AG
|
6.3
|
0.86
|
93
|
4.6
|
8.0
|
|
Baseline
|
BG
|
8.2
|
0.84
|
69
|
6.6
|
9.9
|
|
Post-intervention
|
BG
|
6.7
|
0.86
|
73
|
5.0
|
8.4
|
|
3-month follow-up
|
BG
|
7.4
|
0.92
|
89
|
5.6
|
9.2
|
|
|
contrast
|
group
|
estimate
|
SE
|
df
|
t.ratio
|
p.value
|
|
Baseline - Post-intervention
|
AG
|
1.62
|
0.59
|
80
|
2.7
|
0.02
|
|
Baseline - 3-month follow-up
|
AG
|
0.89
|
0.67
|
81
|
1.3
|
0.57
|
|
Post-intervention - 3-month follow-up
|
AG
|
-0.72
|
0.68
|
79
|
-1.1
|
0.87
|
|
Baseline - Post-intervention
|
BG
|
1.56
|
0.64
|
79
|
2.5
|
0.05
|
|
Baseline - 3-month follow-up
|
BG
|
0.84
|
0.72
|
80
|
1.2
|
0.73
|
|
Post-intervention - 3-month follow-up
|
BG
|
-0.72
|
0.72
|
79
|
-1.0
|
0.96
|
|
Modeling data with imputate data
This study is being conceived with 53 participants and the following graph display the proportion of missing cases.
## # A tibble: 1 x 1
## `sum(is.na(.))`
## <int>
## 1 0


LMM with complete dataset (ater MI procedures)
ACQ6
LMM using complete dataset
|
|
Sum Sq
|
Mean Sq
|
NumDF
|
DenDF
|
F value
|
Pr(>F)
|
|
group
|
0.031
|
0.031
|
1
|
52
|
0.073
|
0.788
|
|
time
|
5.902
|
2.951
|
2
|
104
|
6.916
|
0.002
|
|
group:time
|
3.186
|
1.593
|
2
|
104
|
3.734
|
0.027
|
Symptoms

Total Score

LMM using complete dataset
|
|
Sum Sq
|
Mean Sq
|
NumDF
|
DenDF
|
F value
|
Pr(>F)
|
|
group
|
0.01
|
0.01
|
1
|
52
|
0.01
|
0.92
|
|
time
|
11.05
|
5.53
|
2
|
104
|
7.12
|
0.00
|
|
group:time
|
1.19
|
0.59
|
2
|
104
|
0.76
|
0.47
|
HADS

LMM using complete dataset
|
|
Sum Sq
|
Mean Sq
|
NumDF
|
DenDF
|
F value
|
Pr(>F)
|
|
group
|
0.07
|
0.07
|
1
|
52
|
0.01
|
0.93
|
|
time
|
51.33
|
25.67
|
2
|
104
|
3.05
|
0.05
|
|
group:time
|
13.78
|
6.89
|
2
|
104
|
0.82
|
0.44
|

LMM using complete dataset
|
|
Df
|
X2
|
P(>|Chi|)
|
|
grupo_15
|
1
|
0.13
|
0.72
|
|
tempo_16
|
3
|
14.21
|
0.00
|
|
grupo_15:tempo_16
|
3
|
3.89
|
0.27
|
|
|
Sum Sq
|
Mean Sq
|
NumDF
|
DenDF
|
F value
|
Pr(>F)
|
|
grupo_20
|
0.24
|
0.24
|
1
|
52
|
0.0
|
0.99
|
|
tempo_21
|
91985.79
|
91985.79
|
1
|
52
|
48.1
|
0.00
|
|
grupo_20:tempo_21
|
8234.79
|
8234.79
|
1
|
52
|
4.3
|
0.04
|
Lasts variables

|
|
Sum Sq
|
Mean Sq
|
NumDF
|
DenDF
|
F value
|
Pr(>F)
|
|
grupo1
|
2606574
|
2606574
|
1
|
52
|
0.32
|
0.58
|
|
tempo1
|
13817149
|
13817149
|
1
|
52
|
1.68
|
0.20
|
|
grupo1:tempo1
|
201
|
201
|
1
|
52
|
0.00
|
1.00
|
|
|
Sum Sq
|
Mean Sq
|
NumDF
|
DenDF
|
F value
|
Pr(>F)
|
|
grupo
|
2807340
|
2807340
|
1
|
73
|
0.29
|
0.59
|
|
tempo
|
38637068
|
38637068
|
1
|
73
|
4.03
|
0.05
|
|
grupo:tempo
|
1743
|
1743
|
1
|
73
|
0.00
|
0.99
|

|
|
Sum Sq
|
Mean Sq
|
NumDF
|
DenDF
|
F value
|
Pr(>F)
|
|
grupo1
|
0.60
|
0.60
|
1
|
52
|
0.12
|
0.73
|
|
tempo1
|
2.05
|
2.05
|
1
|
52
|
0.40
|
0.53
|
|
grupo1:tempo1
|
0.95
|
0.95
|
1
|
52
|
0.18
|
0.67
|
|
|
Sum Sq
|
Mean Sq
|
NumDF
|
DenDF
|
F value
|
Pr(>F)
|
|
grupo
|
0.03
|
0.03
|
1
|
43
|
0.01
|
0.92
|
|
tempo
|
18.05
|
18.05
|
1
|
38
|
5.32
|
0.03
|
|
grupo:tempo
|
10.58
|
10.58
|
1
|
37
|
3.12
|
0.09
|