Laboratory evaluation of an optimised internet-based speech-in-noise test for occupational high-frequency hearing loss screening: Occupational Earcheck

Figures & data

Figure 1. Audiometric thresholds for NH and HFHL subjects (for test ear). Error bars represent SDs.

Table 1. Participant characteristics.

Characteristics	NH (N = 18)	HFHL (N = 15)
Male	7 (38.9%)	15 (100%)
Mean age (years)	27.3 (SD = 12.7)	56.3 (SD = 7.0)
Profession
Student	14 (77.8%)	0
Construction-related	1 (5.6%)	9 (60%)
Music-related	1 (5.6%)	2 (13.3%)
Other	2 (11.2%)	4 (26.7%)
Occupational noise exposure	2 (11.1%)	14 (93.3%)
Leisure noise exposure	15 (83.3%)	8 (53.3%)
Use of hearing protection	9 (50%)	15 (100%)

Table 2. Characteristics of the test noises.

Noise version	Filtering	Cut-off frequency	Noise floor
LTASS	–	–	–
LP 1	LP	1.4 kHz	−12 dB
LP 2	LP	1.4 kHz	−15 dB
LP 3	LP	1.6 kHz	−12 dB
LP 4	LP	1.6 kHz	−15 dB

Figure 2. Participant flowchart.

Figure 3. Mean SRT in dB SNR, for NH and HFHL subjects, for all test versions (OEC test). Error bars represent 95% confidence intervals.

Table 3. Mean differences in SRT (dB SNR) (SD) for NH and HFHL subjects.

Test version	NHmean SRT	HFHLmean SRT	ΔNH-HFHL	95% CI
LTASS	−11.7 (1.3)	−9.7 (1.7)	−1.9*	−3	−0.9
LP 1	−19.6 (1.9)	−14.2 (3.0)	−5.4*	−7.1	−3.7
LP 2	−20.6 (2.6)	−14.2 (4.0)	−6.4*	−8.8	−4
LP 3	−19.1 (1.7)	−12.7 (2.5)	−6.3*	−7.8	−4.9
LP 4	−20.3 (2.3)	−13.3 (3.5)	−7.0*	−9.1	−5

*Differences are significant at p < 0.001. All p values are corrected using Bonferroni correction for multiple comparisons.

Table 4. ROC area, sensitivity and specificity, and cut-off value for pass/fail, for all test versions.

Test version	AUC (95% CI)	Sensitivity (%)	Specificity (%)	Cut-off valueSRT (dB SNR)
LTASS	0.85 (0.70–1.00)	80	78	−10.7
LP 1	0.94 (0.85–1.00)	87	94	−17.5
LP 2	0.87 (0.74–1.00)	80	94	−17.3
LP 3	0.98 (0.95–1.00)	93	94	−16.9
LP 4	0.94 (0.86–1.00)	87	100	−16.3

Figure 4. Scatterplots of SRT values against PTA_3,4,6 (upper figure), and test against retest results (lower figure), for test version LP 3, for NH and HFHL subjects.

Table 5. Bivariate correlation coefficients (Pearson’s r) of the SRT values against the PTA of the frequencies 0.5, 1, 2, 4 kHz (PTA_0.5,1,2,4) and 3, 4, 6 kHz (PTA_3,4,6) for all test versions (OEC test).

	All subjects (N = 33)		HFHL subjects (N = 15)
Test version	PTA0.5,1,2,4	PTA3,4,6	PTA0.5,1,2,4	PTA3,4,6
LTASS	0.65**	0.74**	0.62*	0.75**
LP 1	0.68**	0.76**	0.59*	0.56*
LP 2	0.66**	0.83**	0.5	0.73**
LP 3	0.73**	0.85**	0.5	0.68**
LP 4	0.68**	0.83**	0.41	0.61*

*Significant at p < 0.05. **Significant at p < 0.01.

Table 6. Test–retest characteristics of NH and HFHL subjects.

Test version	Group	Testmean SRT(dB SNR) (SD)	Retestmean SRT(dB SNR) (SD)	MeanΔtest–retest (dB)	Meanwithin-subject SD (dB)	ICC**All (N = 33)	ICC**HFHL (N = 15)
LTASS	NH	−11.7 (1.3)	−11.9 (0.9)	0.3	1	0.63*	0.60*
	HFHL	−9.7 (1.7)	−10.0 (1.3)
LP 1	NH	−19.6 (1.9)	−19.0 (1.6)	−0.5	1.4	0.84*	0.84*
	HFHL	−14.2 (3.0)	−13.8 (3.0)
LP 2	NH	−20.6 (2.6)	−21.2 (2.3)	0.4	1.6	0.87*	0.83*
	HFHL	−14.2 (4.0)	−14.3 (4.0)
LP 3	NH	−19.1 (1.7)	−19.7 (1.7)	1.0*	1.5	0.84*	0.68*
	HFHL	−12.7 (2.5)	−14.2 (3.2)
LP 4	NH	−20.3 (2.3)	−21.9 (2.3)	1.2*	1.7	0.87*	0.74*
	HFHL	−13.3 (3.5)	−14.0 (3.3)

*Significant at p < .01 All p values are corrected using Bonferroni correction for multiple comparisons.

**ICC: using a two-way random model, type: absolute agreement, single measures.

Figure 5. Psychometric functions for NH subjects (N = 18) (left), and for HFHL subjects (N = 15) (right), per test version.