Impact of Masks on Speech Recognition in Adult Patients with and without Hearing Loss

Masks are a critical component of personal protective equipment. During the coronavirus disease 2019 (CO-VID-19) pandemic, they have become ubiquitous for all clinical interactions. Universal masking is mandated in Hong Kong, Singapore, and across the majority of the USA [1, 2]. In our clinic, providers wear either a surgical or N95 mask, and patients are provided with medical grade masks for clinical encounters. Otolaryngologists frequently wear an N95 mask given increased risk of particle aerosolization from manipulating the nasal cavity, oral cavity, and oropharynx [3]. Although masks are necessary, concern has been raised that wearing a mask may decrease effective patient-physician communication. Poor communication with healthcare providers worsens health outcomes [4-8], which is important to consider given that 20% of Americans have hearing loss [9]. The current mask policies will remain in place for patients and healthcare workers for the foreseeable future, and therefore it is important to identify communication gaps and develop strategies to ensure effective communication.

The effect of masks on patient-physician interactions is not currently well understood; however, in a survey of 1,000 healthcare providers in Israel and Portugal, Parush et al. [10] found that over 55% of providers reported difficulty with hearing and speech understanding associated with an N95 mask. A recent editorial suggested that patients were mostly concerned with the inability to lip read (56%) and decreased sound level (44%) associated with a facemask in the emergency department [11]. Pre-CO-VID-19 research has investigated the impact of masks on communication and hearing ability with mixed results. One study determined that surgical masks decreased word recognition in normal-hearing subjects [12]. However, other studies have shown no difference [13, 14]. Radonovich et al. [15] found decreased speech understanding among healthcare workers in an intensive care unit when wearing an elastomeric respirator but not with other masks. These nascent studies as well as our collective clinical experience in the COVID-19 era raise concern that quality care may be decreased due to poor communication [4-8]. In the era of shared decision-making, it is particularly important to consider the patient’s ability to hear because it impacts their overall comprehension, engagement, and participation in their care.

The goal of our study was to determine the effect of N95 and surgical masks on speech recognition in clinic patients in a clinical setting. We enrolled patients presenting to an academic otorhinolaryngology clinic with self-reported normal hearing and self-reported hearing loss. We hypothesize that there will be a decrease in word recognition when words are presented with a surgical or N95 mask and that this effect will be greatest in patients with self-reported hearing loss.

This study was approved by the University of North Carolina Institutional Review Board (IRB No. 20-1939). Informed written consent was obtained prior to testing. Forty-five adult participants were recruited from an ambulatory clinical setting. Demographic information including age, gender, presence or absence of hearing loss, regular use of hearing aids, and otologic disease process was collected. All-comers were included regardless of sex, gender, or hearing loss status. Exclusion criteria included non-English speaking and age <19 years.

Thirty-six spondaic words were selected from a standardized audiological list (Central Institute for the Deaf Auditory List W-1) to determine the speech recognition score (Table 1) [16]. The words were randomly recorded in one of the 3 conditions: unmasked (12 words), surgical mask (12 words), or N95 mask (12 words) by either a male or female speaker using a portable digital recording device (TapMedia Ltd. Voice Recorder and Audio Editor app) at a normal speech intensity level (45–60 dB). The speakers were speaking at the same volume regardless of each mask. The recording was not manipulated to account for attenuation of sound due to the mask type. The recording was performed in a quiet room with the speaker positioned 6 inches from the microphone. Each speaker had a general American dialect, was clearly intelligible, and without any abnormality in vocal characteristics as determined by a board-certified otolaryngologist.

Data were collected at the University of North Carolina Department of Otolaryngology – Head and Neck Surgery outpatient clinic. The words were presented to participants in a closed clinical exam room at a standardized distance (6 feet) at a normal speech intensity level (45–60 dB). The speaker was calibrated each day of testing by placing the decibel meter at the head of the patient’s chair, and the speaker volume was adjusted until the decibel meter read between 45 and 60 dB. The speaker was located approximately 6 feet from the patient’s chair. All spondees were presented at the same loudness regardless of mask condition. Attenuation of sound loudness by fabric thickness was not studied. The words were presented in a randomized order. Participants were instructed to repeat the word; they were given 4 s to repeat the word before another word was presented.

The word recognition score was calculated by determining the percentage of correctly repeated words in each condition and overall. The percentage correct by mask and speaker sex was also calculated. The scores were compared between patients with and without self-reported hearing loss.

For data analysis, continuous variables were summarized by means and standard deviations. Categorical variables were summarized by frequencies and percentages. The difference in means between 2 groups was compared using a 2-sample t test. Categorical variables were evaluated using a χ² test for independence. Statistical analyses were performed using RStudio (version 0.99.903; RStudio, Inc., Boston, MA, USA).

A total of 45 participants were included in the study. Patient demographics are reported in Table 2 and Table 3. The mean age of adults was 56 years (18–93 years) with a slight female preponderance overall (53.3%; Table 2). Twenty-three subjects (51.1%) reported some level of hearing loss. The hearing loss cohort had a higher mean age (63 years) compared to the cohort without hearing loss (48 years), and this relationship was statistically significant (p < 0.01). Of those with self-reported hearing loss (n = 23),  18 (78.3%) reported bilateral hearing loss with 9 (39.1%) indicating regular hearing aid use. Among the 9 patients that reported using hearing aids on a regular basis, only 5 (55.5%) of them were wearing their hearing aids for their clinical visit and for this study (Table 3).

The average word recognition score across all participants was 87% (95% CI: 80–94%) without a mask, 78% (95% CI: 70–86%) with a surgical mask, and 61% (95% CI: 53–70%) with an N95 mask (Fig. 1). N95 masks resulted in a significantly lower word recognition score compared to no masks (p < 0.001, 95% CI: 19–31%) or surgical masks (p = 0.0085; 95% CI: 4–28%; Fig. 1). Surgical masks resulted in a nonsignificant decrease in word recognition compared to no mask (p = 0.09; 95 CI: −1.6 to 20%).

Patients with self-reported normal hearing had an average word recognition score of 91% (95% CI: 85–97%) with a surgical mask compared to 66% in patients with self-reported hearing loss (p = 0.002; 95% CI: 10–39%; Fig. 2). With an N95 mask, normal-hearing patients had an average word recognition score of 79% (95% CI: 68–88%) compared to 46% (95% CI: 35–57%) in the hearing loss cohort. There was a statistically significant difference between the normal and hearing-impaired cohorts when using surgical masks (p = 0.002; 95% CI: 10–39%) and N95 masks (p < 0.001; 95% CI: 19–47%) (Fig. 2).

In the normal hearing cohort (Fig. 3), the word recognition score with a surgical mask was 93% and 87% for male and female recordings, respectively (p = 0.31; 95% CI: −5 to 18%). Likewise, the word recognition scores between male and female speaker were 84% and 74% (p = 0.024; 95% CI: 1–19%), respectively, with an N95 mask. In those with self-reported hearing loss (Fig. 3), lower scores were observed with the female speaker across surgical and N95 mask conditions. In participants with self-reported hearing loss, word recognition scores in the surgical mask condition for male and female speakers were 73% and 55%, respectively (p = 0.0653; 95% CI: −1 to 38%). This same cohort exhibited word recognition scores in the N95 mask condition for male and female speakers of 54% and 36%, respectively (p < 0.001; 95% CI: 10–26%).

The COVID-19 pandemic has posed numerous challenges to healthcare. One of these challenges is the use of masks in all clinical encounters. While masks are critical to providing safe care to patients and protecting healthcare providers, a growing body of evidence suggests that masks can impact communication with patients. Previous research involving masks and hearing has been inconclusive and has focused primarily on the healthcare provider [13-15]. Few studies have investigated how N95 masks in particular impact a patient’s ability to hear within a clinical setting.

Our results demonstrate a significant decrease in the word recognition score when using an N95 mask for all participants regardless of hearing loss status. A similar trend was observed with a surgical mask and no mask; however, this comparison was not significant. These results are in agreement with previous studies investigating the impact of masks on hearing [12, 15, 17, 18]. Specifically, word recognition scores declined in a step-wise fashion in the no mask, surgical mask, and N95 mask conditions (Fig. 1). Participants with self-reported hearing loss performed worse across all mask conditions compared to participants with normal hearing and dropped to <50% word recognition of spondees with an N95 mask (Fig. 2). These surprisingly low word recognition scores were obtained using a spondaic word list which is one of the easier word recognition tasks as opposed to the CNC test [19]. Furthermore, word recognition typically was lower with female speakers, in alignment with previous studies [20, 21].

While face masks are vital to limiting the spread of aerosolizing particles, providers must be aware that communication is decreased with patients. In order to minimize the impact of masks on hearing, patients would benefit from good communication techniques from providers. The Hearing Loss Association of America recommends ensuring attention, facing patients directly, speaking slowly rather than shouting, avoiding noisy backgrounds, and choosing to rephrase rather than repeat information [22, 23]. Especially when using a mask, it is important for the patient to repeat back key points from the clinical visit to ensure understanding. Even if and when the mask policy is eliminated, these strategies have the potential to improve communication between patients and providers, in turn improving health outcomes and patient satisfaction [4, 5].

Other authors have suggested the need for technological advances to improve the hearing ability of patients. Various tools suggested include handheld amplifiers, videoconferencing, speech to text applications on smartphone devices, and virtual reality [24-29]. These techniques would likely improve communication regardless of hearing loss status. Furthermore, there is an urgency to improve the design of the currently available respiratory masks. The material used to create an N95 mask is capable of repelling viruses and microbes but may also attenuate sound. Future research is needed to determine an appropriate material that is medically sound and with negligible impact on communication.

This study is not without limitations. First, objective diagnostic data on hearing ability were not obtained via a standardized audiogram, and patient-reported hearing loss is not perfect. However, most patients for clinical encounters across all specialties do not have a formal audiogram, and the use of patient-reported concerns of hearing loss makes our study more generalizable. Future investigations should aim to obtain a hearing level threshold to compare results with mild, moderate, and severe to profound hearing loss using formal audiometry. Next, this study only investigated patients within an otolaryngology clinic. Patients that present to an otolaryngology clinic are often affected by hearing loss. However, all recruited patients were visiting a rhinologist for sinonasal complaints thus minimizing the impact on hearing loss. Second, this study was performed in a single clinical exam room with ambient noise from the clinic. The intent of the study was to increase our understanding of how masks influence clinical interactions with patients in the clinical setting, not in an insulated sound booth. Additionally, there would have been more heterogeneity of participants had we included patients presenting to various clinics, rather than an adult rhinology office. Future studies and primary care practices would be helpful to confirm the generalizability of our study. Future investigations should also include words and phrases that more closely resemble those used in clinical settings. Last, we did not test word recognition of providers. It is important to consider the implications when the treating physician is not able to hear or understand what the patient is saying when using a mask. As providers may also suffer from hearing loss, decreased hearing of their own patients may also lead to poor health outcomes. It is our hope that improved communication between the patient and provider will lead to better medical care for all patients, regardless of their hearing ability.

COVID-19 has altered how healthcare is delivered to patients. As masks will continue to be necessary to maintain the health of patients and providers alike, it is important to understand how this impacts healthcare visits. Our results demonstrate a dramatic decrease in the word recognition score when the provider enunciates words through an N95 mask and especially when the speaker is a female. Although we do not endorse eliminating mask utilization, our results do cast light on potential challenges associated with masks and proper communication during healthcare visits. Future studies are needed to identify optimal methods of communication while providers and patients are wearing masks. The use of masks in daily practice is here for the foreseeable future, and therefore we must adapt our practice to better serve our patients.

This study was approved by the University of North Carolina Institutional Review Board (IRB No. 20-1939). Informed written consent was obtained prior to testing.

The authors have no conflicts of interest to declare.

The authors disclose receipt of the following financial support for the research, authorship, and/or publication of this article: this work was supported by the National Institutes of Health (KL2TR002490) to Adam J. Kimple. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.

Elizabeth Ritter contributed to design, conduct, analysis, presentation, and manuscript preparation; Craig Miller and Justin Morse contributed to analysis and presentation; Princess Onoruah and Abdullah Zeaton contributed to conduct and presentation; Adam Zanation, Brian Thorp, and Charles Ebert contributed to design, analysis, and presentation; Brent Senior and Adam Kimple contributed to design, conduct, analysis, and presentation.

All data generated or analyzed during this study are included in this article. Further enquiries can be directed to the corresponding author.