A portable automatic cough analyser in the ambulatory assessment of cough

Table 3 The validation of the automatic measurement of cough against manual analysis by the MES Cough Analyser.

Nr	T [h]	N_a	N_m	TP	FN	FP	FP rate	TN	SENS [%]	SPEC [%]	AGREE [%]
1	8	1282	1238	1198	40	84	0.0035	23672	96.8	99.6	99.5
2	8	1343	1302	1277	25	66	0.0026	25285	98.1	99.7	99.6
3	8	1154	1120	1105	15	49	0.0019	25552	98.7	99.8	99.7
4	8	1705	1648	1620	28	85	0.0036	23692	98.3	99.6	99.5
5	8	1587	1543	1498	45	89	0.0035	25147	97.1	99.6	99.5
6	8	1782	1715	1657	58	125	0.0053	23680	96.6	99.5	99.3
7	16	2184	2114	2048	66	136	0.0027	50722	96.9	99.7	99.6
8	16	2730	2648	2573	75	157	0.0030	51725	97.2	99.7	99.6
9	16	3155	3041	2921	120	234	0.0045	51279	96.0	99.5	99.3
10	16	3286	3159	3024	135	262	0.0053	48928	95.7	99.5	99.2
11	20	3426	3308	3168	140	258	0.0041	62033	95.7	99.6	99.4
12	20	4311	4162	3976	186	335	0.0054	60805	95.5	99.5	99.2
13	20	4139	3984	3850	134	289	0.0046	61451	96.6	99.5	99.3

Nr - number of measurement period; T - duration of measurement in hours; N_m - number of incidents counted manually by software for the MES Cough Analyser; N_a - number of incidents automatically counted by cough analyser; TP - true positives = number of incidents detected by automatic counting and confirmed by manual analysis; FN - false negatives = number of incidents detected only by manual analysis (not by automatic counting); FP - false positives = number of incidents detected only by automatic analysis (not by manual counting); TN - true negatives = number of potential incidents not registered by automatic or manual counting (as the discrimination time was 1 second, a TN was counted as seconds of measurement without incidents registered either by automatic or manual analysis);

ISSN: 1475-925X