Modelling neuromuscular blockade: a stochastic approach based on clinical data

Figures & data

Figure 1. Block diagram of the NMB model for atracurium, the MPP model.

Figure 2. NMB level of 84 patients undergoing general surgery (_) and NMB for mean parameters (⁃).

Figure 3. Observed signal and its phases.

Figure 4. Mean difference between the reference curve generated from MPP Model with the estimated parameters and and the curves obtained by varying: (a) only ; (b) only ; (c) and in the same amounts and same direction (increase or decrease both) and (d) and in the same amounts but opposite direction: increase () decrease (), for the 84 cases. Vertical dashed lines limit the three phases.

Figure 5. Phase 1. Cases (a) and (b) are acceptable but the case (c) is not acceptable.

Figure 6. Phase 2. Cases (a) and (b) are acceptable but the cases (c) and (d) are not acceptable.

Figure 7. Phase 3. The case (a) is acceptable based on the phase criterion but the case (b) is not.

Figure 8. Bivariate boxplot for .

Table 1. Tests for multivariate normality using the group of tests.

Test	p-Value
Skewness	0.893
Kurtosis	0.492
Generalized Shapiro–Wilk	0.390
Shapiro–Francia	0.250
E-statistic (Energy)	0.389
Based on Mahalanobis distances	0.809

Figure 9. Mean response profiles for the data , , and the response profile for the mean parameters, .

Figure 10. The first 10 minutes of the NMB for 500 simulated cases, (a), and for the 84 real cases, (b).

Table 2. Summary statistics of NMB for the clinical data set and the simulated data set for three different time points.

Time	Data set	Min.	Max.		S	CV (%)
t = 1 min		90	103	98	2.1	2.2
		79	100	99	2.9	2.9
t = 3 min		0.0	99	63	24	38
		7.2	100	64	21	32
t = 9 min		0.0	46	5.7	6.7	118
		0.1	35	8.0	5.9	75

Figure 11. Boxplot for the in three different instants, t = 1, 3, 9 min and for the three data.

Figure 12. Shape parameters used to characterize induced by an initial bolus administration.

Table 3. Shape parameters for the three profiles and differences for .

	Shape parameters
Data set	T50	S	T80
	3.750	−31.68	2.803
	3.930	−24.09	2.685
	2.244	−41.78	1.526
	−0.180	−7.59	0.118
	1.506	10.10	1.277

Figure 13. Shape parameters for data sets , and .

Figure 14. The NMB level record for a patient, (–); the corresponding estimated records obtained from identifying the parameters with the algorithm proposed here, and the algorithm proposed by [13],

Table A1. Summary statistics of patient characteristics.

Patient characteristics	Group	Min.	Max.	Mean	CV (%)
Age (yr)	Total	15	92	58.3	27.7
	M	18	82	61.8	24.5
	F	15	92	54.4	30.4
Weight (kg)	Total	39	94	68.6	17.3
	M	39	94	71.5	17.5
	F	45	90	65.4	15.9

Note: CV (%) = standard deviation/mean 100.

Table A2. Summary statistics of clinical procedure variables.

Clinical procedure variables	Group	Min.	Max.	Mean	CV (%)
Duration	Total	1h04	6h52	3h09	40.0
	M	1h04	6h52	3h24	37.1
	F	1h26	6h02	2h56	41.9
() ()	Total	3.9	20.8	8.0	42.6
	M	4.0	20.8	8.3	46.2
	F	3.9	15.3	7.6	37.4
() (%)	Total	7.2	13.2	9.9	7.4
	M	7.2	13.2	9.9	8.9
	F	8.6	11.7	9.9	6.7

M.M. Silva, Prediction Error Identification of Minimally Parameterized Wiener Models in Anesthesia, Proceedings of 18th IFAC World Congress, Milano, 28 August–2 September 2011.

 Google Scholar