Personal listening device usage among Canadians and audiometric outcomes among 6–29 year olds

Abstract

Objective

To describe personal listening device (PLD) usage and sociodemographic variables among a nationally representative sample of Canadians and examine audiometric outcomes among a subsample.

Design

Audiometry and in-person questionnaires were used to evaluate hearing and PLD usage across age, sex, household income/education level. PLD exposure was quantified using a common occupational noise limit.

Study sample

A randomised sample of 10,460 respondents, aged 6–79, with audiometric analysis of a subsample (n = 4807), aged 6–29, tested between 2012 and 2015.

Results

Loud PLD usage was reported by19.5% of Canadians. The highest prevalence was among teenagers (44.2%) and young adults (36.3%). Among children, 13.1% of users listened at loud volumes. High PLD usage (equivalent to or above 85 dBA, LEX 40) among 12–19 year olds was double that of 20–29 year olds: 10.2% versus 5.1%^E. Five years or more of loud PLD usage was associated with significantly higher mean hearing thresholds compared to less years. No association between loud or high PLD usage and mean thresholds were found.

Conclusion

The majority used PLDs safely, however a small proportion reported high risk usage which will impact hearing should this pattern persist over many years.

Keywords:

• Personal listening device usage
• audiometry
• hearing loss
• sociodemographic
• Canadian population

Introduction

The World Health Organisation (WHO) has estimated that over a billion young people worldwide are at risk of hearing loss due to unsafe listening practices from recreational exposure to loud sound (WHO 2019). The use of personal listening devices (PLDs) is ubiquitous and represents a common source of loud leisure noise exposure. There is concern that the misuse of these devices may have debilitating and potentially irreversible long term consequences for hearing health.

Historically, the largest threat to hearing health was occupational noise which led to the creation of workplace noise limits and hearing conservation programs. In many Canadian workplaces and worldwide, an occupational limit of 85 dBA averaged over a typical 8-h day, exists to protect adults working in potentially hazardous workplace noise settings (Arenas and Suter 2014; Canadian Centre for Occupational Health and Safety 2007). This limit minimises the risk of noise-induced hearing loss (NIHL) from occupational exposure and also assumes that there are no significant noise exposures outside the workplace (Arenas and Suter 2014). In contrast, the onus is on the individual to limit their exposure to hazardous leisure noise through mitigation strategies such as hearing protection usage, moving further away from the source or avoiding the exposure altogether.

However, recently a global mitigation strategy aimed at PLD manufacturers was undertaken by the International Telecommunication Union (ITU) in collaboration with the WHO geared towards decreasing the risk of hearing loss from unsafe PLD usage. The WHO-ITU standard for safe listening devices details a requirement that duration and volume exposure tracking software be used, with a reference exposure of 80 dBA and 75 dBA over 40 h, for adults and children, respectively (ITU-T H.870 2018). Daily and weekly PLD listening habits are to be displayed graphically with user warnings and prompts for action when 100% of weekly allowance are exceeded (ITU-T H.870 2018). In addition, the WHO has recommended a limit of 80 dBA for 8 h for children’s leisure noise exposure, indicating that this limit would be protective for 99% of children from more than 2.1 dB of hearing loss throughout their childhood (Neitzel and Roberts 2019). The development of both the WHO-ITU global standard for safe listening and leisure noise exposure limits for children highlights the public health concern surrounding the risk of hearing loss from recreational noise exposure.

PLD usage is unique in that, unlike most leisure noise activities, the user has control over volume and duration. Sound is delivered directly into the ear with high volume listening over time representing a potential risk to hearing health. Research has shown that average PLD user listening levels range from 75 to 105 dBA (Serra et al. 2014; Sulaiman, Seluakumaran, and Husain 2013, Sulaiman, Husain, and Seluakumaran 2014) with maximum output levels from 101 to over 125 dBA (Breinbauer et al. 2012; Keith, Michaud, and Chiu 2008). The recommended daily noise exposure dose could be exceeded in 5 minutes for those listening at the highest volume levels (NIOSH 1998). For many young people, the risk of NIHL is not well known or understood (Portnuff 2016). The early warning signs of NIHL begin insidiously and develop slowly over time such that individuals are often unaware of early deficits. Hearing damage is cumulative with prolonged exposure to loud sounds leading to impairment of hair cells of the inner ear. In children and adolescents, the consequences of even minimal or unilateral hearing loss can adversely affect auditory and language development, academic performance and social development (Dodd-Murphy and Mamlin 2002; Fischer and Lieu 2014; Fitzpatrick et al. 2019; Kennedy et al. 2006; Wake et al. 2006). Among adults, hearing impairment has been associated with depression, anxiety, low income and decreased employment opportunities (Emmett and Francis 2015; Girard et al. 2009; Li et al. 2014), with particularly far reaching impacts on younger adults, as this stage of life often involves career exploration and engagement in higher education.

Several studies have profiled and analysed PLD usage among adolescents and or young adults using different methodologies to determine the potential risk to hearing. One common approach has been the use of an occupational noise limit equivalent of 85 dBA, over 8 h or 40 h, based on calculated duration and self-reported or measured volume, as a benchmark for identifying risky PLD usage. Twardella et al. (2017) reported high risk PLD usage among high school students (n = 2143) where approximately one-quarter exceeded a 40-h week equivalent of 85 dBA. Levey, Levey, and Fligor (2011) reported much higher estimates among college students (n = 189) where 51.2% exceeded an 85 dBA 40 h weekly equivalent. A systematic review by Jiang et al. (2016) found that up to 58.2% of adolescents and young adults exceeded the recommended daily noise dose of 85 dBA over 8 h, indicating that these young people are at risk of developing NIHL.

To assess the impact of risky PLD listening behaviour, several studies have used audiometric evaluation to assess the hearing of PLD users with some reporting elevated mean hearing thresholds among high risk PLD users compared to low risk or non-users (Chen et al. 2011; Feder et al. 2013; Le Prell et al. 2011; Peng, Tao, and Huang 2007; Sulaiman, Husain, and Seluakumaran 2015; Taneja et al. 2015). However, other studies did not report any significant threshold differences (Ahmed et al. 2007; Kumar et al. 2009; Mostafapour, Lahargoue, and Gates1998; Sulaiman, Husain, and Seluakumaran 2014; Twardella et al. 2017). Overall, a wide range of high risk PLD usage has been reported, from 10% to 25%, depending on the risk methodology used, age group, sample size and study design (Ahmed et al. 2007; Le Prell et al. 2013; Levey, Levey, and Fligor 2011; Peng, Tao, and Huang 2007; Shah et al. 2009; Sulaiman, Husain, and Seluakumaran 2014, 2015; Taneja et al. 2015; Torre 2008; Williams 2005). The majority of the aforementioned studies examined PLD usage among University students, using small to medium size convenience samples, with recruitment taking place at campuses through flyers, emails or advertisements. Some exceptions include a randomised prospective study of 500 subjects (undergraduates, graduates, postgraduates), aged 16–30, enrolled through a medical college-based clinic, a small study (n = 55) of white collar workers and students, aged 15–48, recruited on a noisy street and lastly, a study (n = 94) of 18–65 year olds, including faculty and staff as well as students, recruited outside a University recreation centre (Shah et al. 2009; Taneja et al. 2015; Williams 2005). The widespread usage and popularity of PLDs among children, adolescents and young adults, underscores the importance of understanding PLD user patterns and risk across a broad age spectrum, including but not limited to University students, with random selection of subjects to allow generalisation of findings.

There have been several studies examining PLD usage among males compared to females with conflicting findings. Some studies have found that males listen to PLDs at higher volumes and longer durations compared to females (Hussain et al. 2018; Taneja et al. 2015; Torre 2008; Williams 2005), while others have reported no significant sex differences in terms of high risk PLD listening (Gilliver et al. 2017; Levey, Levey, and Fligor 2011; Vogel et al. 2009). Sociodemographic factors such as income or education level and the association with risky PLD listening has not been studied as extensively across age groups. Some adolescent studies have reported associations between high risk PLD users or participants at risk of HL with living in single parent households, low parental education and attending pre-vocational or vocational training schools (Twardella et al. 2017; Vogel et al. 2011) while a study of 10–17 year olds found that subjects from lower income families (<100k per annum) had significantly higher typical PLD sound levels compared to higher income families (≥ 100k per annum) (Feder et al. 2013). Few PLD studies using randomly selected subjects, ranging from children to young adults, have examined the association between high risk PLD usage and household income or education level.

There have been no population-based studies profiling PLD user listening habits, sociodemographic variables or audiometric hearing thresholds among PLD users, allowing estimates of the potential risk to hearing health among a nationally representative sample. The objectives of the present study, based on 2012–2015 Canadian Health Measures Survey (CHMS) data, are to examine PLD usage patterns and sociodemographic variables across age, sex, household education and income among a nationally representative sample of Canadians, aged 6–79. Specific study objectives include the examination of self-reported PLD listening volume and duration in relation to pure-tone average hearing loss thresholds, the presence of an audiometric notch, and mean hearing thresholds among a subsample, aged 6–29. A final study objective is to calculate the proportion of participants in high, medium and low PLD user categories, using a common occupational limit as a guide, in order to determine those at risk of developing NIHL, as predicted by ANSI or ISO standards.

Methods

Study design and study population

Recruitment of CHMS participants took place for Cycle 3 (2012–2013) and Cycle 4 (2014–2015) consecutively, across five regions of Canada: Atlantic, Quebec, Ontario, Prairies, and British Columbia. In this ongoing cross-sectional survey, one or two individuals were randomly selected from each household using a multistage weighted person-level sampling strategy by age and sex. The CHMS excludes full-time members of the Canadian Forces, residents of the three territories, First Nations Reserves and other Aboriginal settlements, certain remote regions, and institutional residents. Despite these exclusions, CHMS data are considered to be representative of the Canadian population. The CHMS uses a rigorous methodology for validating responses, as outlined in the CHMS User Guide (Statistics Canada 2015). Available at: http://www23.statcan.gc.ca/imdb-mdi/document/5071_D4_T9_V2-eng.htm

Using the 2011 Census as the sampling frame, collection sites were selected that included a population of at least 10,000 respondents, with a maximum travel distance of 50–75 km to a mobile examination clinic (MEC) site, resulting in 360 collection sites across Canada. This sampling protocol covers approximately 96% of the Canadian population. The CHMS, carried out on approximately 10,000 individuals over a four-year period, allows estimates for sex and age groups, as presented in this study (Statistics Canada 2015).

This study was approved by the Health Canada and Public Health Agency of Canada Review Ethics Board (Protocol #2005-0025).

Data collection

Individuals were invited to complete a household questionnaire, administered in the individual’s home. The individual was then invited to participate in the physical measures component of the CHMS which included a hearing evaluation requiring travel to a MEC site. Participants who agreed to take part in the physical measures component were allotted a $100 honorarium. Proxy interviews were accepted in cases of physical and/or intellectual impairment and for children under age 12.

Household questionnaires including noise exposure questions and a hearing evaluation module were administered to participants, aged 6–79 years (n = 10,460). Participants were excluded at different stages of the hearing evaluation module as indicated in Figure 1.

Figure 1. Canadian Health Measures Survey: Exclusions for hearing evaluation module (unweighted). Note: DPOAE = distortion product otoacoustic emissions.

Household questionnaire

Self-reported data were collected in-person, in English or French, using a computer-assisted household questionnaire, comprised of 12 different modules, requiring approximately 90–110 min to complete. The noise exposure module consisted of 52 questions and required approximately 30–40 min to administer. In the household questionnaire, participants were asked about a wide range of topics from sun exposure to dietary habits in addition to gathering demographic, socioeconomic, health and lifestyle information. The complete CHMS household questionnaire (Cycle 3 and 4) can be accessed online.

In the noise exposure module, exposure to potentially hazardous noise sources were estimated by asking participants about specific sources of loud noise exposure during the previous year, including duration and frequency. The first portion of the noise exposure module was dedicated to loud workplace noise exposures, and these findings have been published previously (Feder et al. 2017). The second portion of this module focussed on specific sources of loud leisure noise, including PLD usage with headphones or earbuds. Participants were asked to indicate “Yes” or “No” to having listened to audio devices i.e. PLDs, using headphones or earbuds in the past 12 months. An affirmative response was followed by: “On average, in the past 12 months, how many hours per week did you usually listen to music, movies or other types of audio using headphones or earbuds?” Examples of audio devices provided to the participant included mp3 players, iPods, cell phones, stereo systems, televisions or computers. Participants were then asked about volume level using two questions: A: “Is the volume level usually at or above three quarters of the maximum volume?” and B: “Is the volume usually so loud that someone standing an arm’s length away would have to speak in a raised voice for you to understand what they were saying?” An affirmative answer to either of these questions indicates loud PLD usage . This was followed by the question: “How many hours per week do you usually listen to audio devices, using headphones or earbuds, at a volume that is (A and/or B)?”

Other loud leisure noise exposures or activities that participants were asked about in terms of frequency and duration in the previous 12 months included: power tools, heavy industrial, farming or construction equipment, hand-held gasoline engines, motorcycles or snowmobiles driven at highway speeds, sporting or entertainment events, amplified music (concerts, nightclubs, band member), loud music listening without headphones or earbuds (home or car stereos), firearms (number of rounds fired). The prevalence of loud leisure noise exposures among a representative sample of Canadians have been reported in Feder et al. (2019).

Hearing evaluation protocol

Participants were evaluated in a MEC. The possibility that a temporary threshold shift (i.e. a temporary increase in hearing threshold following loud noise exposure) may have influenced measurements was considered by asking participants to recall their exposure to loud noise/music in the 24 h before testing. Among 6–29 year olds, loud noise/music exposure was reported by 11.1% (95% CI: 8.7–13.5) of participants in the 24 h prior to testing, with significant differences observed at 3 kHz (p < 0.01), 6 kHz (p < 0.05) and HF PTA (p < 0.05) where those who reported 24 h prior noise exposure showed a higher likelihood of having a hearing loss compared to the non-exposed group (data not shown). This group of participants (n = 452) were therefore removed from the hearing analysis (Figure 1). Individuals who refused or had apparent cognitive deficits that could potentially interfere with testing were excluded from audiometric evaluation.

All testing was carried out by trained health measures specialists with supervision provided by a certified audiologist who conducted periodic on-site visits to ensure quality control. Hearing tests, with the exception of otoscopy and tympanometry, were conducted in a portable audiometric sound booth (Eckel, AB-4230 ).

A Casella CEL-633 sound level metre monitored ambient sound pressure levels inside the sound booth. Testing was paused if sound pressure levels exceeded 55 dB (A-weighted) for more than 2 seconds. This testing protocol safeguard was developed in the case of loud environmental or people noise (from inside or outside the MEC) which may potentially interfere with hearing test results.

The hearing evaluation consisted of otoscopy, tympanometry and audiometric evaluation (0.5–8 kHz). Otoscopy was performed, prior to testing, using the Welch Allyn otoscope (Model 25020) to identify gross abnormalities, including the presence of blood, pus, excessive or impacted ear wax, a growth, tumour or foreign object in the ear canal, a collapsed ear canal, or other occlusion. Criteria for otoscopic exclusion included obstructed ear canal, acute pain or infection, open wounds or bandages covering the ear(s), refusal to remove hearing aid, or chronic abnormalities including congenital atresia or microtia of the ear canal (one or both ears). No further testing was performed if the individual was excluded by otoscopy.

Tympanometry was conducted using the A GSI 39 Auto Tympanometer. A normal tympanogram was defined as compliance between 0.2 and 1.8 cm³ with middle ear pressure between −150 and +50 daPa in an equivalent ear canal volume of between 0.75 cm³ and 2.0 cm³ (Jerger 1970).

Audiometric evaluation was carried out using a computer-controlled CCA-100 mini audiometer with insert earphones (EAR 5 A). In cases where a participant was excluded from tympanometry for the reasons noted above, audiometry was performed using supra-aural headphones (TDH-39) instead of insert earphones. The audiometer was calibrated daily using the Bio-Acoustic simulator BAS-200. Hearing thresholds were assessed at 0.5, 1, 2, 3, 4, 6 and 8 kHz. Testing followed procedures recommended for standard audiometry using automatic mode, except when the respondent could not physically press the response button, had very slow response times, or when difficulties were noted with automatic mode. Manual mode was carried out using the modified Hughson–Westlake ascending method (Carhart and Jerger 1959).To avoid interference with audiometric evaluation, participants were asked to refrain from chewing.

Definitions

Audiometric means for individual frequencies (0.5–8 kHz) and for each pure-tone average (PTA) were calculated using the worse ear audiometric value for each participant. HL was defined as an impairment of more than 15 dB HL for 6–19 year olds and 26 dB HL or greater for 20–29 year olds at the following PTAs: four frequency (0.5, 1, 2, and 4 kHz) (FFPTA), high frequency (3, 4, 6 and 8 kHz) (HFPTA) and low frequency (0.5, 1 and 2 kHz) (LFPTA). Individuals with either bilateral or unilateral loss were classified as having a hearing loss.

An audiometric “notch” was defined as follows: Thresholds at 0.5 and 1 kHz are less than or equal to 15 dB, and the maximal threshold at 3, 4, or 6 kHz is greater than or equal to 15 dB higher or worse than the highest threshold for 0.5 and 1 kHz, and the threshold at 8 kHz is greater than or equal to 10 dB lower or better than the maximal threshold for 3, 4, or 6 kHz, unilaterally or bilaterally (Niskar et al. 2001).

PLD user groups, based on self-reported PLD usage in the previous 12 months and the criteria for loud were defined as follows: 1) Loud PLD user – a participant who used a PLD, at loud volume; 2) Non-loud PLD user – a participant who used a PLD, but not at loud volumes; 3) Non PLD user – a participant who had not used a PLD.

Loud PLD user groups were further divided into high, medium and low noise weighted PLD user groups. These were based on participants’ self-reported loud duration of weekly PLD exposure. In accordance with published literature, a typical sound pressure level concomitant with loud PLD usage of 90 dBA was used, with 13 h being the calculated weekly duration equivalent to 85 dBA (LEX, 40 h). Participants were assigned to the high PLD user category if their calculated noise exposure was equivalent to (or above) the occupational limit of 85 dBA (LEX, 40 h; ISO, 2013). The low category represents a noise exposure at or below 75 dBA (effective quiet) and the medium category was selected to be exclusive of the high and low. A calculated noise exposure equivalent at or above the occupational limit of 85 dBA (LEX, 40 h) was categorised as high (40 h per week or more) or medium (greater than 4 to less than 40 h per week), and a noise exposure of 75 dBA (LEX, 40 h) or less (4 h or less per week) was categorised as low noise exposure from PLD usage. For more information on the procedure used for categorising PLD users, see Feder et al. (2019).

In summary, definitions for PLD usage were categorised as follows: 1. five noise weighted PLD user groups (high, medium, low, non loud PLD user, non PLD user); 2. number of years of loud PLD usage (5 years or more, up to 5 years); 3. number of hours per week (unweighted) of PLD usage (8 h or more, less than 8 h, non loud PLD usage and no PLD usage); and 4. PLD usage (any volume and non PLD user).

Statistical analysis

Descriptive statistics of sociodemographic and PLD characteristics were carried out for all participants, aged 6–79. Hearing thresholds and PLD user group analyses were focussed on a subset of participants, aged 6–29. These analyses included weighted mean audiometric thresholds for individual frequencies and all PTAs according to sociodemographic variables of age, sex, household income, highest household education as well as for each CHMS cycle. Weighted means, frequencies and cross-tabulations were used to describe each of the three PLD user groups (loud PLD users, non-loud PLD users and non PLD users) by the same demographic variables. All estimates were weighted at the person level to represent the Canadian population. The weighting procedure carried out was based on the principle that the individual selected in a probability sample such as the CHMS, “represents” himself or herself in addition to several other individuals not in the sample. The weighting procedure that was applied corresponds to the number of individuals represented by the participant in the population as a whole. Therefore, the weighting, which takes into account the age and sex distribution of the population, non-response and the sampling strategy of the survey, allows means and frequency calculations that are representative of the population. See the CHMS User Guide (Statistics Canada 2015) for further details about the weighting procedure and its application.

In order to analyse the relationship between audiometric means (continuous outcome variable) and PLD usage (predictor variable), linear regression models were used. PLD usage included: 1. five noise weighted PLD user groups; 2. number of years of loud PLD usage; 3. number of hours per week (unweighted) of PLD usage. Audiometric means included individual frequencies as well as HFPTA, FFPTA and LFPTA. Logistic regression models were used to examine the relationship between prevalence of HL or audiometric notch (binary outcome variables) and PLD usage (predictor variable). PLD usage only included “used a PLD, any volume” and “non PLD usage”. These two user groups were collapsed from the five user groups (high, medium, low, non-loud and non PLD user) due to the low prevalence of HL and the Niskar notch among the five PLD user groups used in other models. All models were adjusted for CHMS cycle and sociodemographic variables of age, sex and household income. Lastly, loud PLD usage and exposure to other loud leisure noise activities were analysed using logistic regression modelling. Household education level was removed from this analyses to avoid multicollinearity, as this variable was highly correlated with household income.

Analyses were conducted using SAS Enterprise Guide 7.1 and SUDAAN 11.0.0 software. To account for the complex survey design, p values, 95% confidence intervals, and coefficients of variation (CV), were estimated using the bootstrap technique with 22 degrees of freedom (Rao, Wu, and Yue 1992; Rust and Rao 1996). Statistical significance was specified as a p value of less than 0.05. All comparisons were carried out using Satterthwaite F test, and Bonferroni corrections were conducted where multiple pair wise comparisons were required to compare more than two groups. In the tables and text, estimates with a coefficient of variation (CV) between 16.6% and 33.3% are flagged with an E (interpret with caution); those with a CV that exceeds 33.3% are not releasable and are designated F.

Results

Study population

A total of 10,460 respondents aged 6–79, took part in Cycle 3 and 4 of the CHMS, 2012–2015, representing approximately 30.8 million Canadians (49.9% males). The combined response rate for the household questionnaire and the hearing evaluation module was 89.6% and 78.3%, respectively. Among all Canadians, loud PLD usage was reported by 19.5% of participants. Loud usage was most pronounced among 12–19 year olds (44.2%) and 20–29 year olds (36.3%), with these age groups showing significant differences compared to other loud PLD user age groups (p < 0.01) (Table 1).

Table 1. PLD user group by age and sex among Canadians, aged 6–79.

		Loud PLD Users			Non Loud PLD Users			Non PLD Users
Variable description	Variable	n	N ('000)	% (95% CI)	n	N ('000)	% (95% CI)	n	N ('000)	% (95% CI)
All participants	Overall^§	2016	5956	19.5 (17.5–21.4)	3224	8300	27.1 (24.8–29.5)	5110	16361	53.4 (50.5–56.4)
Sex	Male^§	1059	3236	21.2 (18.8–23.6)	1636	4270	28 (25–30.9)	2468	7768	50.9 (47.1–54.6)
	Female^§	957	2720	17.7 (15.5–19.9)	1588	4030	26.3 (23.7–28.8)	2642	8593	56 (53–59)
Age Groups	6–11^§	260	283	13.1 (10.6–15.5)	818	866	40 (36.2–43.8)	949	1017	46.9 (41.9–52)
	12–19^§	851	1394	44.2 (40.6–47.9)	955	1479	46.9 (43–50.9)	221	279	8.8 (6.7–11.5)
	20–29	248	1726	36.3 (30–42.7)	238	1797	37.8 (30.4–45.2)	212	1226	25.8 (18.5–33.1)
	30–39^§	300	1124^E	24.1 (15.6–32.6)^E	417	1405	30.2 (25.1–35.2)	702	2127	45.7 (37.4–54)
	40–49^§	200	697	14.3 (10.9–17.8)	356	1119	23 (17.3–28.8)	770	3046	62.7 (55–70.3)
	50–59^§	78	479^E	9.6 (6.8–13.2)	169	869	17.3 (12.7–22)	553	3663	73.1 (68.1–78.1)
	60–69^§	65	205^E	5.1 (3.5–7.3)^E	208	627	15.6 (12.4–18.8)	1069	3189	79.3 (75.5–83.2)
	70–79^§		^F	^F	63	138	6.9 (5–9.4)	634	1814	90.7 (87–93.4)

Note: 110 participants did not respond to some or all of the PLD user questions. Source: 2012/2013; 2014/2015 Canadian Health Measures Survey (CHMS).

^EInterpret with caution, coefficient of variation between 16.6% and 33.3%.

^Ftoo unreliable to be published coefficient of variation was greater than 33.3%; n = study sample; N = population size; CI: confidence interval; PLD: personal listening device.

^§PLD user groups significant at the p < 0.01 level within the level of the demographic variable.

Among children, aged 6–11, just over half were PLD users (53.1%), with 13.1% engaged in loud PLD listening for an average of 4.3 h per week, according to parental report (Table 2). Participants, aged 12–19, had a significantly higher average number of hours per week of loud PLD usage (9.1 h per week) compared to those aged 40 and over, for whom average hours of listening per week ranged from 3.1 to 4.6 h per week (p < 0.05) (Table 2).

Table 2. PLD weekly usage (any volume and loud volume) among Canadians, aged 6–79.

		PLD Usage (Any Volume) (hours/week)				Loud Volume PLD Usage (hours/week)
		n	N('000)	Median (95% CI)	Mean (95% CI)	n	N('000)	Median (95% CI)	Mean (95% CI)
Overall		5240	14256	3.3 (2.8–3.8)	7.9 (7–8.7)	2016	5956	2.9 (2.5–3.2)	6 (5.4–6.5)
Sex	Male	2695	7506	3.8 (3.1–4.5)	8.8 (7.6–10)	1059	3236	3 (2.5–3.5)	6.5 (5.7–7.2)
	Female	2545	6750	2.9 (2.5–3.4)	6.8 (6–7.7)	957	2720	2.6 (1.7–3.5)	5.4 (4.2–6.6)
Age groups	6–11	1078	1149	1.6 (1.2–2)	4.2 (3.2–5.2)	260	283	1.6 (0.7–2.5)^E	4.3 (2.9–5.7)
	12–19	1806	2873	6.5 (6.1–6.9)	12.3 (10.9–13.6)	851	1394	4.8 (3.8–5.8)	9.1 (7.5–10.6)
	20–29	486	3522	4.6 (3.6–5.5)	8.1 (6.2–9.9)	248	1726	3.3 (2.1–4.5)^E	6 (4.9–7.1)
	30–39	717	2529	2.7 (1.8–3.6)	7.1 (5.4–8.9)	300	1124	F	5.2 (3.3–7.1)^E
	40–49	556	1816	2.7 (2–3.5)	5.4 (4.1–6.6)	200	697	1.9 (0.9–2.8)^E	3.1 (2.5–3.8)
	50–59	247	1349	2.7 (1.7–3.7)^E	6.3 (4.1–8.6)^E	78	479	F	4.6 (2.9–6.3)^E
	60–69	273	831	2.7 (1.7–3.7)^E	7.1 (5–9.1)	65	205	2 (0.7–3.2)^E	4.2 (2.8–5.6)
	70–79	77	187	3.6 (1.4–5.7)^E	8.1 (4.8–11.4)^E	14	49	^F	^F

Source: 2012/2013; 2014/2015 Canadian Health Measures Survey (CHMS).

^EInterpret with caution, coefficient of variation between 16.6% and 33.3%

^Ftoo unreliable to be published coefficient of variation was greater than 33.3%; n = study sample; N = population size; CI: confidence interval; PLD: personal listening device.

Due to the elevated PLD usage patterns observed among the under 30 aged cohorts, further analysis was restricted to 6–29 year olds. The sample size was therefore reduced to 4807 respondents, aged 6–29, representing 10.1 million Canadians, with fairly equal male to female representation and equal distribution across household income levels (Table 3). A discrepancy among highest household education level was observed with the majority having some postsecondary education or higher (n = 3680) compared to those reporting secondary education or less (n = 831). Once participants with 24 h prior noise exposure were excluded and audiometric test exclusions were applied, the final study sample was comprised of 4125 respondents.

Table 3. Sociodemographic variables among 6–29 year olds.

Sociodemographic variable	Categories	*n	*N ('000)	% (95% CI)
Total		4807	10148
Age groups	6–11	2055	2185	21.5 (20.7–22.4)
	12–19	2048	3183	31.4 (30.1–32.6)
	20–29	704	4781	47.1 (45–49.2)
Sex	Male	2410	5104	50.3 (47.7–52.9)
	Female	2397	5045	49.7 (47.1–52.3)
Household income	<$50,000	1408	3068	30.2 (26.2–34.3)
	$50,000 to <$100,000	1665	3761	37.1 (33.5–40.6)
	$100,000 or more	1734	3319	32.7 (28.6–36.8)
Education (highest level in household)	Secondary education or less	831	1879	19.9 (16.9–22.9)
	Some post secondary or higher	3680	7563	80.1 (77.1–83.1)
CHMS Cycle	2012–2013	2405	5086	50.1 (48.2–52)
	2014–2015	2402	5063	49.9 (48–51.8)

Note: All participants completed the household questionnaire. Source: Canadian Health Measures Survey (CHMS) 2012–2015.

n = study sample; N = population sample; CI: confidence interval.

*Includes participants who were excluded from audiometric evaluation.

Sociodemographic variables, high and loud PLD usage and other loud leisure noise exposure: ages 6–29

Overall, among 6–29 year olds, approximately one-third were loud PLD users. No significant sex, household income or education differences were observed within each PLD user group. Just over one-third of males (35.6%) and slightly fewer females (31.9%) used their PLD at loud volumes (Table 4). The prevalence of loud PLD usage by individual age, presented in Figure 2, shows that among PLD users as young as 6, approximately one-third (33%^E) listened to their PLD at loud volumes. This pattern increases, peaking at age 22, with 82% of participants reporting loud PLD listening before gradually decreasing. The prevalence of high PLD usage (equivalent to or above 85 dBA, over 40 h or more per week) among 6–11 years olds was 1.3%^E (95% CI: 0.7–2.2; N = 28,000^E); among 12–19 years, high PLD usage was 10.2% (95% CI: 8.5–11.9; N = 321,000) and among 20–29 year olds, 5.1%^E (95% CI: 2.6–9.6; N = 240,000^E) were high PLD users.

Figure 2. Prevalence of loud PLD usage among Canadian PLD users, aged 6–29. ♦ Prevalence of loud PLD usage: CV <16.6%; ♦ Prevalence of loud PLD usage: CV ≥ 16.6% and <33.3%, interpret with caution; PLD = personal listening device; CV=coefficient of variation. Note: Prevalence rates with a CV ≥ 33.3% not shown.

Table 4. PLD user groups by age, sex, household income, education and participation in other loud leisure noise activities among Canadians aged 6–29 years.

		Loud PLD Users			Non Loud PLD Users			Non PLD Users
Variable	Category	n	N('000)	% (95% CI)	n	N('000)	% (95% CI)	n	N('000)	% (95% CI)
All participants	Overall^§	1359	3402	33.8 (30.7–36.9)	2011	4142	41.1 (37.4–44.9)	1382	2521	25 (21–29.1)
Age Groups	6–11^§	260	283	13.1 (10.6–15.5)	818	866	40 (36.2–43.8)	949	1017	46.9 (41.9–52)
	12–19^§	851	1394	44.2 (40.6–47.9)	955	1479	46.9 (43–50.9)	221	279	8.8 (6.7–11.5)
	20–29	248	1726	36.3 (30–42.7)	238	1797	37.8 (30.4–45.2)	212	1226	25.8 (18.5–33.1)
Sex	Male^§	693	1807	35.6 (31.4–39.9)	999	2139	42.2 (38.4–46)	697	1126	22.2 (17.7–26.7)
	Female*	666	1595	31.9 (27.6–36.3)	1012	2003	40.1 (34.7–45.5)	685	1395	27.9 (22.5–33.4)
Household Income	<$50,000	451	1051	34.5 (28.8–40.2)	515	1177	38.6 (32.3–44.8)	427	822	26.9 (20.7–33.2)
	$50,000 to <$100,000*	483	1293	34.6 (29.7–39.5)	692	1525	40.8 (33.9–47.7)	473	919	24.6 (18.4–30.8)
	$100,000 or more^§	425	1058	32.3 (26–38.5)	804	1440	43.9 (37.4–50.4)	482	780	23.8 (18.3–29.3)
Education (highest level in household)	Secondary education or less	274	697	37.4 (29.9–44.9)	299	659	35.4 (27.3–43.4)	244	508	27.2 (19.4–35.1)
	Some post secondary or higher^§	984	2474	33 (29.6–36.4)	1599	3121	41.6 (37.3–45.9)	1060	1907	25.4 (20.9–29.9)
^aParticipation in other loud leisure noise activities excluding PLD use	^bhigh (40 h or more per week)^§	340	1296	51.1 (44.5–57.7)	217	874	34.5 (28.4–40.5)	114	366E	14.4 (9.3–19.6)^E
	^bmedium (between 4 and 40 h/week)^§	452	1130	41.9 (36.1–47.6)	446	1007	37.3 (31–43.6)	202	563	20.8 (14.5–27.2)
	^blow (up to 4 h/week)^§	379	652	23.5 (18.7–28.4)	798	1243	44.9 (41–48.7)	551	874	31.6 (26–37.2)
	non-user^§	185	322	15.6 (10.6–20.7)	546	1016	49.4 (41.1–57.7)	515	719	35 (27.8–42.1)
CHMS Cycle	2012–2013^§	688	1773	35 (30.5–39.5)	989	2008	39.6 (35.2–44.1)	709	1284	25.4 (20.5–30.2)
	2014–2015^§	671	1629	32.6 (28.7–36.4)	1022	2134	42.7 (36.8–48.6)	673	1237	24.7 (18.1–31.4)

Note: A total of 55 participants did not respond to some or all of the PLD user questions. Source: 2012/2013; 2014/2015 Canadian Health Measures Survey (CHMS).

n = study sample; N = population size; CI: confidence interval; PLD: personal listening device

^aOther loud leisure noise activities or exposures include power tools, heavy industrial, farming or construction equipment, hand-held gasoline engines, motorcycles or snowmobiles driven at highway speeds, sporting or entertainment events, amplified music (concerts, nightclubs, band member), loud music listening without headphones or earbuds (home or car stereos), firearms.

^bNoise exposure classifications were based on adjusted weekly self-reported duration of exposure or usage that was equivalent to or above 85 dBA SPL or greater than or equal to 40 h per week (high), greater than 4 h to less than 40 h per week (medium), and 4 h or less per week (low) (Feder et al. 2019).

^EInterpret with caution, coefficient of variation between 16.6% and 33.3%.

*PLD user groups significant at the p < 0.05 level within the level of the demographic variable.

^§PLD user groups significant at the p < 0.01 level within the level of the demographic variable.

Among loud PLD users, the majority also participated in other loud leisure noise activities at either a medium (between 4 and 40 h/week) or high (40 h or more per week) noise exposure classification. In contrast, among the non loud PLD users and non PLD users the majority either did not participate in other loud leisure noise activities or only participated at a low noise exposure classification (4 h or less per week) (Table 4).

As seen in Table 5, among loud PLD users, participants were more than twice as likely to also report loud amplified music exposure (concert/nightclub attendance, being in a band) (OR: 2.2; 95% CI: 1.6–3) and three times as likely to listen to loud car or home stereo systems without headphones or earbuds (OR: 3.0; CI: 2.1–4.3).

Table 5. Association between loud PLD usage and other loud leisure noise activities.

Loud leisure noise activity	Participation	n	N ('000)	% (95% CI)	OR (95% CI)
Power tools	Yes	502	1248	48.5 (42.5–54.4)	1.2 (0.9–1.7)
	^aNo	857	2154	43.4 (38.2–48.5)
Heavy industrial /farming/ construction equipment	Yes	222	606^E	49.9 (35.6–64.2)	1.3 (0.6–2.5)
	^aNo	1136	2795	44.2 (39.3–49)
Gasoline engines	Yes	311	746	48.3 (39.1–57.5)	1.2 (0.8–1.8)
	^aNo	1047	2654	44.3 (39.7–48.8)
Motorcycle/snowmobiles at highway speed	Yes	276	611	56.8 (44.8–68.8)	1.7 (1–3.1)
	^aNo	1083	2791	43.2 (38.5–47.8)
Sporting/entertainment event	Yes	490	1203	47.3 (40.4–54.1)	1.1 (0.8–1.7)
	^aNo	869	2200	44 (38.6–49.4)
Amplified music: concerts, clubs, band member	Yes	704	2230	53.8 (48.9–58.6)	2.2 (1.6–3)^§
	^aNo	655	1172	34.5 (29.7–39.3)
Car or home stereo music (no headphones/earbuds)	Yes	829	2210	58.4 (53.1–63.7)	3 (2.1–4.3)^§
	^aNo	530	1193	31.8 (26.6–36.9)
Firearms	Yes	108	255^E	43.1 (27.6–58.7)^E	0.8 (0.4–1.7)
	^aNo	991	2864	49.4 (43.9 –54.8)

Source: 2012/2013; 2014/2015 Canadian Health Measures Survey (CHMS).

n = study sample; N = population size; CI: confidence interval; OR: odds ratio; PLD: personal listening device.

^aReference group.

^EInterpret with caution, coefficient of variation between 16.6% and 33.3%.

^§p < 0.01.

Audiometric outcomes and sociodemographic variables among 6–29 year olds

Analysis of mean audiometric thresholds (individual frequencies and PTAs) and sociodemographic variables showed significantly higher mean thresholds for frequencies 0.5, 2, 3, 4 and 6 kHz and all PTAs (p < 0.05) among 20–29 year olds compared to other age groups. Among individuals from lower income households (<50 K), significantly higher means were found for LFPTA (p < 0.05) and at 1 kHz (p < 0.01) compared to those from high income households. In addition, significantly higher means were observed at 0.5 kHz and LFPTA (p < 0.01), at 1 kHz, 2 kHz and FFPTA (p < 0.05) among those reporting secondary education or less compared to higher education households (Supplementary Tables 1a–1c).

Relationship between PLD usage (using five user groups) and audiometric mean thresholds

Linear regression models were used to examine the relationship between each of five PLD user groups (high, medium, low, non-loud PLD user and non PLD user) and audiometric means for individual frequencies and PTAs while accounting for CHMS cycle and the sociodemographic variables of age, sex and household income. No significant differences in audiometric mean thresholds were found among the five PLD user groups, however, a tendency towards higher mean thresholds was observed among 20–29 year olds. Notably, those who were classified as high PLD users (40 h or more per week of loud PLD usage) had higher audiometric means compared to the other user groups (Figure 3). No significant relationship or tendency was observed between mean thresholds and high PLD user groups among 12–19 year olds or 6–11 year olds (Figures 4 and 5).

Figure 3. Audiometric *means by PLD user group – 20–29 year olds. LF = low frequency pure tone average (0.5, 1 and 2 kHz); FF = four frequency pure tone average (0.5, 1, 2 and 4 kHz); HF = high frequency pure tone average (3, 4 ,6 and 8 kHz). *adjusted for age, sex, family income and CHMS cycle.

Figure 4. Audiometric *means by PLD user group – 12–19 year olds. LF = low frequency pure tone average (0.5, 1 and 2 kHz); FF = four frequency pure tone average (0.5, 1, 2 and 4 kHz); HF = high frequency pure tone average (3, 4, 6 and 8 kHz). *adjusted for age, sex, family income and CHMS cycle.

Figure 5. Audiometric *means by PLD user group – 6–11 year olds. LF = low frequency pure tone average (0.5, 1 and 2 kHz); FF = four frequency pure tone average (0.5, 1, 2 and 4 kHz); HF = high frequency pure tone average (3, 4, 6 and 8 kHz). *adjusted for age, sex, family income and CHMS cycle.

Relationship between PLD usage and non-usage with HL PTAs and the Niskar audiometric notch

The prevalence of HL (any PTA) and the presence of an audiometric notch was low in each of the five PLD user groups (previously defined) and across age groups. For example, the prevalence of the audiometric notch ranged between 13.2%^E (95% CI: 8.0–18.4) (6–11 year olds) and 19.4% (95% CI: 14.4–24.4) (12–19 year olds), with a similar prevalence of 18.9% (95% CI: 12.8–25.0) for young adults (20–29 year olds). HL PTAs also showed a wide but low prevalence across age groups. Therefore, the five PLD user groups were collapsed into two groups for this analysis: used a PLD, at any volume and no PLD usage. Across all age groups, those who used a PLD, any volume was n = 2889 (74.8%; 95% CI: 70.8–78.7) and no PLD usage was n = 1199 (25.2%; 95% CI: 21.3–29.2). Among 6–11 year olds, PLD usage was n = 968 (53.6%; 95% CI: 48.0–59.2) and no PLD usage was n = 823 (46.4%; 95% CI: 40.8–52.0; among 12–19 year olds, PLD usage was n = 1514 (90.5%; 95% CI: 87.5–92.9) and no PLD usage was n = 190 (9.5%; 95% CI: 7.1–12.5); among 20–29 year olds, PLD usage was n = 407 (74.4%; 95% CI: 66.9–82.0) and no PLD usage was n = 186 (25.6%; 95% CI: 18.0–33.1). Logistic regression models used to examine the relationship between PLD usage and HL at all PTAs (HFPTA, FFPTA, LFPTA) and the audiometric notch found no significant associations (data not shown).

Loud PLD usage and audiometric means: 5 years or more versus less than 5 years

Among 20–29 year olds who have used their PLD at loud volumes for 5 years or longer, significant increases were observed in mean audiometric thresholds for HFPTA and individual frequencies of 3 kHz, 4 kHz and 8 kHz compared to those reporting less than 5 years of loud PLD usage (Table 6). In contrast, among 12–19 year olds who reported loud PLD usage for 5 years or more, lower audiometric means at 2 kHz were observed compared to less than 5 years.

Table 6. Adjusted^a audiometric means and duration of loud PLD usage (in number of years) among Canadians, aged 12–29 years.

Frequency	Duration of loud PLD usage	12–19 year olds (n = 672, N = 1096) dB (95% CI)	20–29 year olds (n = 203, N = 1465) dB (95% CI)
500 Hz	5 years or more	9.91 (7.85–11.97)	13.25 (10.12–16.39)
	less than 5 years	11.65 (10.65–12.64)	12.77 (10.45–15.08)
1 kHz	5 years or more	6.79 (4.42–9.16)^E	9 (6.42–11.59)
	less than 5 years	6.49 (5.46–7.52)	5.92 (3.58–8.25)^E
2 kHz	5 years or more	5.48 (3.22–7.73)^E,^§	10.71 (6.94–14.47)^E
	less than 5 years	8.88 (7.7–10.07)	9.31 (6.86–11.76)
3 kHz	5 years or more	6.69 (5.16–8.23)	13.37 (8.42–18.31)^E,*
	less than 5 years	7.99 (6.62–9.36)	7.58 (4.48–10.68)^E
4 kHz	5 years or more	4.8 (3.13–6.46)^E	10.61 (5.65–15.58)^E,^§
	less than 5 years	4.73 (3.59–5.87)	4.51 (1.45–7.57)^E
6 kHz	5 years or more	14.94 (11.04–18.85)	19.69 (14.07–25.31)
	less than 5 years	12.8 (10.21–15.39)	13.61 (9.64–17.58)
8 kHz	5 years or more	9.57 (7.01–12.13)	16.97 (10.84–23.1)^E,^§
	less than 5 years	8.56 (6.68–10.43)	6.79 (3.36–10.22)^E
FF PTA	5 years or more	6.53 (4.89–8.18)	10.49 (7.23–13.74)
	less than 5 years	7.63 (6.72–8.55)	8.02 (5.97–10.07)
HF PTA	5 years or more	8.5 (6.48–10.51)	14.36 (9.2–19.52)^E,*
	less than 5 years	8.1 (6.64–9.56)	7.91 (4.92–10.91)^E
LFPTA	5 years or more	7 (5.16–8.85)	10.8 (7.83–13.78)
	less than 5 years	8.71 (7.76–9.67)	8.97 (6.83–11.1)

Note: loud PLD usage was self-reported and was not calculated using weightings. Source: 2012/2013; 2014/2015 Canadian Health Measures Survey (CHMS).

n = study sample; N = population size given in ‘000; FFPTA: four frequency pure tone average (0.5, 1, 2 and 4 kHz); HFPTA: high frequency pure tone average (3, 4, 6 and 8 kHz); LFPTA: low frequency pure tone average (0.5, 1 and 2 kHz); dB: decibels; CI: confidence interval.

^aModels were adjusted for age, sex, family income and CHMS cycle.

^EInterpret with caution, coefficient of variation between 16.6% and 33.3%.

*Significant difference between audiometric means among those who listened to their PLD at loud volumes for 5 years or more and less than 5 years p < 0.05.

^§Significant difference between audiometric means among those who listened to their PLD at loud volumes for 5 years or more and less than 5 years p < 0.01.

Duration of loud PLD listening and mean audiometric thresholds

No significant differences were observed in mean audiometric thresholds among those who listened to their PLD at loud volumes for 8 h per week or more compared to those who listened for up to 8 h per week and non PLD users (which includes both non PLD users and non-loud PLD users). However, among 20–29 year olds, there was a tendency observed such that those who listened to their PLD at loud volumes for less than 8 h per week had lower audiometric mean thresholds compared to those who listened at loud volumes for 8 h or more per week (Figure 6). Although this observation was not statistically significant, it can be observed at all frequencies and PTAs.

Figure 6. Audiometric *means by loud PLD listening per week – 20–29 year olds. LF = low frequency pure tone average (0.5, 1 and 2 kHz); FF = four frequency pure tone average (0.5, 1, 2 and 4 kHz); HF = high frequency pure tone average (3, 4, 6 and 8 kHz). *adjusted for age, sex, family income and CHMS cycle.

Discussion

This nationally representative study characterising PLD user listening behaviours among a broad age group underscores the ubiquity of these devices. In this first population-based study of 10,460 Canadians, aged 6–79, nearly one-half reported using a PLD in the previous 12 months, with three-quarters of this cohort aged 29 or younger and just over 20% aged 50–79. Adolescents, aged 12–19, had the highest prevalence of PLD usage in the present study (91%), a finding confirmed by research from different countries (Gilliver et al. 2017; Kim et al. 2009; Vogel et al. 2010). Understanding PLD usage characteristics among demographic groups in a population is important, however of particular interest are behaviours such as high volume and duration of listening which have the potential to impact hearing health.

Based on the representative sample in this study, 19.5% or 5.9 million Canadians, aged 6–79, were loud volume PLD users while among 6–29 year olds, just over one-third or 3.4 million Canadians, listened at loud volumes. Not surprisingly, the highest prevalence of loud usage was among teenagers followed by young adults, with nearly one-half of 12–19 year olds and over one-third of 20–29 year olds listening at loud volumes; the average weekly listening time for these age groups was 9.1 and 6.0 h, respectively. While direct comparison is difficult due to study design and noise quantification differences, estimates of loud PLD usage in a Dutch study of 1687 adolescents were only somewhat lower. Among the 90% of PLD users in Vogel et al. (2011) study, loud PLD usage (89 dBA or greater) for 1 h or more per week was reported for close to one-third of participants. Collectively, these findings point to higher risk of NIHL among this younger age group.

Some PLD studies have found that adolescents from among socially disadvantaged groups or attending vocational school had higher PLD volume levels and or listening durations (Feder et al. 2013; Twardella et al. 2017; Vogel et al. 2011). Although not statistically significant, the present study found a higher prevalence of loud PLD users with low household education levels (i.e. secondary education or less) compared to non PLD users, among 6–29 year olds. Income levels were fairly equally distributed for each PLD user group. However, there were slightly more loud PLD users compared to non PLD users from low income households (i.e. less than $50 K per year). These non-significant findings show a pattern in line with previous studies. However, further corroborating research is needed to fully understand the impact of low income or parental education on PLD user listening habits.

Among 6–29 year olds, no significant sex differences were found among PLD user groups or in the prevalence of loud PLD listening volume or prolonged listening duration, which is consistent with several adult studies (Fligor, Levey, and Levey 2014; Gilliver et al. 2017; Levey, Levey, and Fligor 2011). However, other researchers have reported a higher prevalence of males showing high risk listening behaviours compared to females among both adolescent and university student cohorts (Le Prell et al. 2011; McNeill et al. 2010; Taneja et al. 2015; Vogel et al. 2011; Widen et al. 2017). Despite the findings of smaller convenience sample studies, the present population based findings may be an indication that high risk PLD usage is one leisure noise activity which is engaged in fairly equally by both males and females. Therefore, educational messaging would be beneficial for all individuals in this age cohort, regardless of sex.

Older age groups had lower prevalence rates of loud PLD usage compared to younger age groups, with lower weekly average durations of listening, a finding corroborated by Gilliver et al. (2017). Nonetheless, the representative proportions of loud PLD users in the over 29 age groups were not insubstantial. For example, among 30–39 year olds, just over one-half (54.3%) used a PLD and nearly one-quarter listened at loud volumes for an average of 5.2^E h per week, representing 1.12 million Canadians while 37.3% of 40–49 year olds used a PLD and 14% were loud PLD users (listening average of 3.1 h per week), representing 697,000 Canadians. Among children, aged 6–11, approximately 13.1% or 283,000 were engaged in loud PLD listening for an average of 4.3 h per week, according to parental report. There are no other population-based studies to date which report the prevalence and duration of loud PLD usage among a broad age spectrum allowing a snapshot of PLD usage patterns. The finding that loud PLD usage is also pervasive among older cohorts, over age 30, and among children, highlights the benefit of educational outreach for a wide range of age groups.

Much of the PLD research has focussed on teenagers and young adults due to the high risk listening behaviours reported among this demographic. The present study findings of loud PLD listening by nearly one-half of 12–19 year olds is fairly consistent with previous adolescent studies, which reported high volume listening among one-half and nearly one-third of cohorts (Vogel et al. 2009; Pellegrino et al. 2013). The duration of loud listening is an important variable enabling a more accurate estimate or identification of high risk PLD usage. Taneja et al. (2015) found that 57.2% of subjects (n = 500), aged 16–30, were characterised as high risk PLD users, in that they listened to their PLD at loud volume for a mean of 25.3 h weekly. In comparison, the present study found that 39.5% of 12–29 year olds listened at loud volume for a mean of 7.4 h weekly. This discrepancy may be due to factors such as recruitment through a single facility and the use of a 10-point rating scale for self-reported PLD volume in the Taneja et al. (2015) study compared to the present study, which used vocal effort and volume setting to capture loud volume listening among participants. The present study findings were fairly consistent with some studies of college students or young adults which found approximately one-third of PLD users listened to music above a level considered safe (80 dBA for 9 h) or listened at full volume (Hoover and Krishnamurti 2010; Kumar et al. 2009).

Some studies have quantified PLD exposure as a 40-h week equivalent using self-reported volume and duration, in order to compare to specific occupational noise limits and further refine identification of high PLD users. The present study, which used this approach, estimated that among 12–19 year olds and 20–29 year olds, 10.2% and 5.1%^E were high PLD users, respectively. A similar approach, used in a study of 13–19 year olds, found that 22% exceeded 85 dBA over 40 h per week (Twardella et al. 2017). The increased prevalence of high risk PLD usage reported by Twardella et al. (2017) may be due to the restricted recruitment (students uniquely recruited from one German city) and the skewed age of participants in that study (which were primarily 15–16 year olds) compared to the nationally representative data collection in the present study.

Gilliver et al. (2017) found similar estimates of high risk PLD usage, despite the self-selection bias limitation, as subjects completed questionnaires and were recruited online, and despite using a methodology which incorporated transducer type. Approximately 15% of participants, aged 18–35, were in the high and very high risk categories with 10% of PLD users or 400 participants showing exposure estimates in excess of the daily noise dose (DND) recommended by workplace exposure guidelines (Gilliver et al. 2017). This similarity in findings despite the differences may point to the strength of using an equivalent occupational limit as a reference for quantifying high PLD usage. A similar conclusion to that of Gilliver et al. (2017) can likewise be drawn from the present study, which is that individuals in the high or loud PLD user groups may be at increased risk of developing NIHL considering that PLD use is unlikely to be their only source of loud noise exposure.

In fact, the present study findings confirm this premise, showing that loud PLD users were more than twice as likely to report exposure to loud amplified music (such as concert or nightclub attendance or being in a band) and three times as likely to listen to loud car or home stereos. Furthermore, the majority of loud PLD users in the present study took part in other loud leisure noise activities for a weekly duration of either 40 h or more or between 4 and 40 h. A recent study found that approximately 28.5% of respondents who reported loud leisure noise were in the high cumulative leisure noise exposure category, representing 6.6 million Canadians (Feder et al. 2019). Therefore, the risk of developing NIHL from loud leisure noise needs to consider cumulative leisure noise exposure from multiple sources.

In examining the impact that high risk PLD usage has on hearing across conventional audiometric frequencies, research findings have been variable. Several studies of high risk PLD users have reported either some degree of measured hearing loss or “clinically” normal hearing, but higher mean hearing thresholds at conventional frequencies compared to controls (Chen et al. 2011; Feder et al. 2013; Hussain et al. 2018; Le Prell et al. 2011; Peng, Tao, and Huang 2007; Sulaiman, Husain, and Seluakumaran 2015; Taneja et al. 2015; Widen et al. 2017), while other studies have not found notable differences (Kumar et al. 2009; Mostafapour, Lahargoue, and Gates 1998; Sulaiman, Husain, and Seluakumaran 2014; Twardella et al. 2017). The present study showed no significant mean threshold differences at conventional frequencies among five PLD user groups (high, medium, low, non-loud and non-user) for the subset of 6–29 year olds. In addition, there was no association between the Niskar notch and PLD usage at any volume while other studies found no association between high risk PLD usage and the Niskar or 4 kHz notch (Sulaiman, Seluakumaran, and Husain 2013; Twardella et al. 2017). Interestingly, the notch prevalence among 12–19 year olds in the present study was considerably higher than reported by Twardella et al. (2017) (19.4% versus 2.3%) and similar to the notch prevalence among 20–29 year olds (18.9%). These findings may be an indication of high cumulative leisure noise exposure among both adolescents and young adults. The presence of a notch in the audiogram is often an indication of NIHL in individuals with a noise exposure history. However, according to Schlauch and Carney (2011), the Niskar notch definition may lead to high false positive rates and overestimation of NIHL. Further research examining the Niskar notch definition and prevalence among noise exposed individuals is needed.

Among 20–29 year olds, a tendency was observed whereby high PLD users showed consistently higher audiometric means, ranging from 3 to 6 dB, for the majority of individual frequencies (0.5, 2, 3, 4 and 6 kHz) compared to low PLD users. This finding may be indicative of a subclinical or subtle degradation taking place which has not yet developed into a measurable hearing loss in this young age group. It is well known that NIHL begins insidiously and requires a number of years of loud noise exposure to develop.

Some researchers have evaluated risk by examining hours per week of loud PLD usage and mean hearing thresholds. One study by Marron et al. (2014) involving 115 participants, aged 18–84, reported that those who listened to their PLD for more than 8 h per week had higher mean thresholds, accounting for age, compared to 8 h or less per week of listening, without considering PLD volume. Although not statistically significant, the present study found that among 20–29 year olds, loud PLD listening for less than 8 h per week was associated with lower mean thresholds at all frequencies and PTAs when compared to 8 h or more per week. This observation may be explained by the cumulative exposure to loud leisure noise from both PLDs and other sources of leisure noise (e.g. amplified music from nightclub or concert attendance). This trend was not observed among 12–19 year olds, which may be due to this being a younger age group in which hearing threshold differences are not yet apparent. Further research examining this premise may be of interest.

Several studies examining the impact of PLD usage duration in years on hearing thresholds in adolescents or young adults found higher mean thresholds or HL PTA in subjects with 5 years or more of PLD usage compared to non-users or low risk PLD users (Feder et al. 2013; Kim et al. 2009; Le Prell et al. 2013; Peng, Tao, and Huang 2007). The present study found higher mean thresholds among 20–29 year olds who were loud PLD users for 5 years or more, compared to less than 5 years. These findings together with previous research suggests that years of PLD usage (loud or any volume) is an important variable to consider with regards to hearing health.

There have been fewer studies examining PLD usage among children compared to adolescents or young adults. The present study suggests that the prevalence of loud PLD usage among young school children is not insignificant. Among 6–11 year olds, according to parental report, just over one-half used a PLD, which is somewhat higher than the 40% PLD usage reported among 9–11 year olds (n = 3116) (Le Clercq et al. 2018). Our study findings are the first time that nationally representative data profiling loud PLD usage among children have been collected. Although some of the coefficient of variation estimates were flagged as “E” indicating high sampling variability and therefore findings should be interpreted with caution, these estimates nonetheless provide a snapshot of PLD usage among young children, filling a knowledge gap. Among 6–11 year old PLD users, 13.1% were classified as loud users, representing 283,000 Canadian while high PLD usage (i.e. calculated noise exposure equivalent to or above 85 dBA over 40 h or more per week) was 1.3%^E, representing approximately 28,000^E Canadian children. Due to children’s vulnerability and consequent hearing health repercussions, a more stringent leisure noise exposure limit of 80 dBA LEX was recommended by the WHO (Neitzel and Roberts 2019). If this noise exposure limit was applied to the present cohort, a larger proportion of children would be identified as high PLD users. The importance of protecting the hearing of children has been highlighted by research showing that even mild or minimal degrees of HL have been associated with poor academic achievement, lower language outcomes and negative social emotional development (Dodd-Murphy and Mamlin 2002; Lieu 2004; Lieu et al. 2010; Moeller et al. 2007; Wake and Poulakis 2004).

Studies of PLD usage and audiometric outcomes among children are limited, however two studies found significantly poorer thresholds in the right ear or possible mild high frequency HL among PLD users (Basjo et al. 2016; Le Clercq et al. 2018). In contrast, the present study found no significant mean threshold differences among high versus low PLD users, aged 6–11. Nonetheless, the present study pattern of loud PLD usage by age (illustrated in Figure 2), whereby one-third and one-quarter of 6 and 8 year olds respectively, listened at loud volumes, shows consistency with previous research (Basjo et al. 2016). Overall, the prevalence of loud PLD usage among young school-aged children found in the present study is striking and warrants careful surveillance. There is research suggesting an increased susceptibility to developing HL in adulthood when early exposure to loud noise occurs (Van Kemp and Davies 2013).

Recommendations for future studies

Numerous studies have shown higher mean thresholds in the extended high frequencies (EHF), 9–16 kHz, that were associated with unsafe PLD usage which has led some to conclude that early effects on hearing may be more clearly observed using EHF audiometric testing (Le Prell et al. 2013; Peng, Tao, and Huang 2007; Serra et al. 2005; Sulaiman, Husain, and Seluakumaran 2014; 2015). The inclusion of EHF audiometric testing in future longitudinal studies examining PLD usage patterns and EHF audiometric thresholds would be useful for monitoring usage trends. The use of other audiological evaluation tools such as otoacoustic emissions and speech in noise testing would be advantageous in identifying early NIHL. Tinnitus is another potential early warning sign of HL and future CHMS studies will be examining the prevalence of tinnitus and loud leisure noise exposure including PLD usage. Furthermore, studies which incorporate information regarding background or environmental noise levels during PLD usage would also be useful for examining PLD user patterns. Lastly, educational outreach with regards to safe PLD usage may be beneficial for reducing the risk of NIHL among PLD users.

Limitations

The strength of this study is the large sample size and robust study design, with data randomly collected yielding nationally representative data across a broad age range. However, a number of limitations must be highlighted. One limitation is the use of self-reported PLD volume and duration data reliant on participant recall. Apart from one study by Portnuff, Fligor, and Arehart (2013), others have found significant differences between self-reported and objectively measured PLD output levels, leading to conclusions that self-reported listening volumes and durations may be unreliable and underestimated (Hodgetts, Szarko, and Rieger 2009; Sulaiman, Seluakumaran, and Husain 2013; Vogel et al. 2010; Widen et al. 2017). A number of reasons for this discrepancy have been suggested including inattention to volume dial, poor recall, social desirability and high variability with self-reported volume reflecting only an average (Le Clercq et al. 2016; Portnuff, Fligor, and Arehart 2011). Additionally, the collection of PLD usage data by proxy or parental report for children, aged 6–11, may have resulted in inaccuracies.

This study did not include information regarding background or environmental noise levels during PLD usage. This information would have allowed more in-depth analysis with regards to conditions under which these devices were used. A further limitation is that no data was collected regarding transducer, headphone or earbud type which would have allowed for more accurate estimations of PLD volume output. Lastly, the calculated noise exposure estimates for high, medium and low PLD user categories, have some inherent uncertainties, one of which is the assigned noise weighting factor for loud PLD volume. While best efforts were made to assign a representative noise weighting factor, there is a margin of error associated with this value. Another limitation is the potential for misclassification into the appropriate PLD user category (high, medium and low) due to inherent inaccuracies in using self-reported data i.e. PLD volume and duration, which could in turn affect PLD user group estimates.

Finally, audiometric evaluation only included conventional frequencies (0.5–8 kHz) and did not include extended high frequencies, which some studies have shown to be more sensitive to the early onset of NIHL. While audiometric evaluation is the gold standard, there are limitations to using this instrumentation exclusively. Individuals showing difficulty understanding speech in noise may be missed which is a limitation. The use of targeted self-reported hearing health questions would have been a useful supplement to the hearing evaluation.

Conclusion

This is the first population-based study to profile PLD usage patterns across a broad age range of Canadians, aged 6–79. These study findings confirm previous research showing that an overwhelming majority of high risk PLD users are teenagers and young adults. Loud PLD usage prevalence among those aged 30–49 was not insubstantial, ranging from 14% to 24%. PLD usage among children may represent an emerging vulnerable group warranting surveillance. Among 6–11 year olds, just over half used a PLD, with 13.1% listening at loud volumes for an average of 4.3 h per week. An estimate of high PLD usage, using a calculated occupational noise equivalent of 85 dBA or greater, for 40 h or more per week, found that among 12–19 year old PLD users, twice as many were high PLD users compared to 20–29 year olds, 10.2% versus 5.1^E%. Among 6–11 year olds, 1.3%^E were in the high PLD user group.

These study findings did not show an impact on hearing from loud or high PLD usage with the exception of loud PLD usage for 5 years or more compared to less than 5 years, which showed significant elevations of mean hearing thresholds at several high frequencies (3, 4 and 8 kHz, HFPTA). No significant changes to mean hearing thresholds at conventional frequencies were found in loud or high PLD users compared to low PLD users. However, a non-significant tendency towards elevated mean hearing thresholds were associated with high PLD usage and loud usage for 8 h or more per week among 20–29 year olds.

The majority of Canadians reported using PLDs safely, however a small proportion of PLD users reported listening at volumes and durations which may result in hearing deficits should this usage pattern continue over many years. It is also important to highlight that loud PLD usage is unlikely to be the only loud leisure noise exposure for these young age groups. The long term ramifications of continued loud PLD usage may result in accelerated or early onset age-related HL (Kujawa and Liberman 2006).

Acknowledgments

The authors wish to thank all of the individuals, including children and families, who took part in this study. http://www23.statcan.gc.ca/imdb-bmdi/instrument/5071_Q2_V4-eng.htm#a29 (Last viewed 25 October 2020).

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This research was funded by Statistics Canada and Health Canada.