Workers in certain industries (e.g., agriculture) do not participate in a hearing conservation program and rarely receive this important preventive care. The National Institute for Deafness and Communication Disorders (NIDCD) has published a questionnaire to assess a person's need for medical examination of their hearing. The use of a self-administered questionnaire may pose a simple, low cost opportunity to assess the hearing health of workers not included in a hearing conservation program. However, the validity of the NIDCD questionnaire has not been tested. A study was conducted to compare the results of audiometric screening and the NIDCD questionnaire in identifying persons at risk for hearing impairment who may benefit from a medical referral. Data were collected from 103 farm operators at a regional farm show. Receiver-operator characteristic curves, measuring the ability of the questionnaire to correctly classify those with and without the disease suggested that the level of performance of the questionnaire as a whole, ranged from 0.74 to 0.69 (fair to poor). However, selected questionnaire items (e.g., Do you have to strain to understand conversations?) had moderate agreement (0.38 to 0.41) with screening audiogram results. Overall, the 10-item instrument did not perform as well as instruments with fewer items reported in separate studies. These findings suggest that, while less desirable than audiometry, alternative self-administered instruments might perform the better in this group. This approach may be useful as an alternative screening method to detect risk of hearing loss and identifying the need for medical evaluation of hearing sensitivity, particularly among farm operators.
Keywords: Noise-induced hearing loss, prevention, screening
|How to cite this article:|
McCullagh MC. Validation of a questionnaire to identify hearing loss among farm operators. Noise Health 2012;14:32-8
| Introduction|| |
Hearing loss is highly prevalent in American society, affecting approximately 29 million adult Americans. This number is expected to rise as the US population ages, and with the high popularity of personal listening devices. 
An estimated 22 million workers are exposed to hazardous noise at work,  placing them at risk for noise-induced hearing loss (NIHL). This condition is among the most common work-related diseases, and the second most self-reported occupational disease or injury.  The Hearing Conservation Standard provides for systems to protect workers from NIHL, such as noise level monitoring and a hearing conservation program for at-risk employees, which includes audiometric testing, training, and provision of hearing protection devices. 
However, not all workers are included in the standard; among these are farm operators. The USDA estimates that there are about 3,281,000 farm operators,  and over 1 million farm laborers.  Exposure to hazardous noise has been found to be essentially universal among farm operators, , while regular use of hearing protection is very uncommon. ,, Estimates of prevalence rates for NIHL among farm operators vary greatly, and have been reported to be 17%,%,  22%,  38%,  65%,  and 72%. 
Methods of determining prevalence of hearing loss include audiometric testing and self-report (such as questionnaires or interviews). While audiometric testing is considered the gold standard, the equipment, personnel, and costs of administering these tests may be prohibitive in some settings. However, questionnaires are subject to bias, and there are few studies evaluating the validity of these measures,  and no such studies used a sample of farm operators. Establishment of the validity of a self-reported measure of hearing ability would aid in the estimate of the prevalence of hearing loss among subpopulations, such as farm operators, and would provide useful information in the prioritization of health services and development of health policy.
Results of selected investigations of self-reported measures of hearing ability are summarized in [Table 1]. Of these studies, ,,,, sample characteristics, questions used, and analysis methods are greatly varied, making comparisons between studies difficult. Most studies used general samples of adults, with the exception of studies by Gomez  and Choi,  who used samples of farmers. Most questionnaire items directly queried about the person's perceived hearing ability (e.g., "Do you have any difficulty with your hearing?"),  and the number of questionnaire items ranged from 14  to 1. , Comparisons between measures were accomplished by a variety of statistical analysis methods, including sensitivity/specificity, ,,,, receiver-operator characteristics (ROCs),  and Kappa. 
|Table 1: Selected previous studies comparing self-reported hearing ability and pure tone audiometry|
Click here to view
The NIDCD has posted a hearing ability questionnaire on its website.  The 10 yes-no items address subjective areas that may present challenges to persons with hearing loss, such as hearing conversation in a crowded room, or using the telephone. The items included in the questionnaire are different from items used by other researchers. ,,,, The NIDCD questionnaire includes instructions for persons who answer three or more items in the affirmative to seek hearing testing from an otolaryngologist or licensed audiologist. However, a review of the literature failed to reveal any reports of studies establishing the validity of the questionnaire. The purpose of this study was to compare the results of the NIDCD questionnaire with audiometric screening as a method of identifying persons at risk for hearing impairment who may benefit from a medical referral.
| Methods|| |
The study protocol was approved by the author's university institutional review board. Study participants included attendees of a regional farm show who were at least 18 years of age and active in farming operations at least 20 hours per week. Participants were asked to complete a written questionnaire, including the 10 questionnaire items and demographic information.
After questionnaire completion, participants were queried about their occupational and recreational noise exposure, use of hearing protection, and presence of hearing concerns (e.g., ear pain, drainage). The screening protocol provided the immediate referral of persons with the listed concerns. Audiometric screening was administered by trained technicians. Pure tones were presented in each ear in the following order: 1000, 2000, 4000, 6000, 8000, and 500 Hz. Each frequency was presented at 5 dB and progressed by 5 dB increments until participants indicated they heard the tone. Testing took place in a quiet room using a recently calibrated manual audiometer (Maico model MA20 series 102, Maico, Eden Prairie, MN). Results were recorded on paper forms, and later transcribed to computer for analysis. Following administration of the screening audiometry, results of the test were interpreted for the client, and an individually tailored educational session addressing the person's risk for NIHL and implications for self-care was provided. Referrals were made to licensed hearing professionals (i.e., primary care provider, audiologist, or otolaryngologist) for all persons who measured >25 dB hearing loss at any tested frequency.
For purposes of this study, three definitions of hearing loss were used: low, mid, and high frequency. Low frequency hearing loss was defined as binaural hearing thresholds (non-weighted mean of both ears) averaging greater than 25 dB at 500, 1000, and 2000 Hz. Mid-frequency hearing loss was defined as binaural hearing thresholds averaging greater than 25 dB at 1000, 2000 and 4000 Hz. High-frequency hearing loss was defined as binaural hearing thresholds averaging greater than 25 dB at 4000, 6000, and 8000 Hz.
Hearing questionnaire responses were coded and scores were calculated using the NIDCD criteria (three or more positive responses qualify a failure) and SPSS (version 17) software. Measured hearing status was determined for each of the three separate definitions of hearing loss. Chi-square analysis and Kappa statistic were conducted for each of the 10 hearing questionnaire items for each of the three frequency ranges tested. Measured hearing status was determined for each of the three separate definitions of hearing loss. ROC curves were used to analyze the relationships between sensitivity and specificity of the questionnaire.
| Results|| |
The mean age of participants was 53.48 years (SD = 14.71). Most respondents were men (88%), White (97%), working on family-owned farms, and producing crops. More than half (53%) worked on farms of at least 1000 acres. Sample demographics are summarized in [Table 2].
Hearing screening questionnaire
No persons presenting at the screening event were identified with concerns of ear pain or drainage. The distribution of responses by item is summarized in [Table 3]. The item receiving the lowest number of affirmative responses was, "Do you have a problem hearing over the telephone," with 24% of the sample responding yes. The item receiving the highest number of affirmative responses was, "Do you have trouble hearing in a noisy background," with 75% of the sample responding yes.
|Table 3: Distribution of questionnaire item responses and agreement (using Kappa) with audiogram results for three definitions of hearing loss|
Click here to view
The mean number of positive responses to items was 4.64 (SD = 2.93). Using the NICDC pass/fail criteria, over 70% of the sample failed, responding affirmatively to three or more items.
Pure tone audiometry
Mean binaural threshold averages for three frequency ranges were calculated, and are displayed in [Table 4]. Mean binaural thresholds were progressively higher as frequency levels increased. The distribution of participants passing the audiometric exam at each set of frequencies (low, mid, and high) declined as the frequency levels increased [Table 4].
The agreement between each questionnaire item and the audiogram results for each of the three frequency levels was compared using the Kappa statistic [Table 3]. Kappa scores were below 0.30 for most items and frequency ranges, except for selected frequencies for questionnaire item 4 (strain to understand), item 1 (problem using telephone), item 7 (people seem to mumble) and item 9 (trouble understanding women and children).
An ROC curve is a plot of the true-positive rate against the false-positive rate for the different possible cut points of a diagnostic test.  ROC curves were used to examine the sensitivity and specificity of the NIDCD hearing questionnaire relative to the "gold standard" of the audiometric examinations; results are displayed in [Figure 1].[Figure 2] and [Figure 3]. The area under the curve (AUC) is a measure of test accuracy.  AUCs for the respective measures of hearing loss (low frequency, mid-frequency, and high frequency) are 0.71, 0.74, and 0.69, respectively. These results suggest that the performance of the hearing questionnaire as a whole as compared to audiometry can be considered poor to fair, with the questionnaire performing slightly better with the mid-frequency definition of hearing loss, compared to the low and high frequency definitions.
|Figure 1: ROC curve, for low (500, 1000, 2000 Hz) frequency hearing loss (AUC = 0.71)|
Click here to view
|Figure 2: ROC curve for mid (1000, 2000, 4000 Hz) frequency hearing loss (AUC=0.74)|
Click here to view
|Figure 3: ROC curve for high (4000, 6000, and 8000 Hz) frequency hearing loss (AUC = 0.69)|
Click here to view
Sensitivity and specificity of the questionnaire were calculated for each definition of hearing loss [Table 5]. For the prescribed fail threshold of three or more items in the affirmative, sensitivity ranged from 0.80 to 0.89, while specificity ranged from 0.37 to 0.43.
|Table 5: Questionnaire item sensitivity and specificity, compared to audiogram results using three definitions of hearing loss (Low, mid, and high frequency)|
Click here to view
| Discussion|| |
The effectiveness of the questionnaire in correctly identifying persons in this sample as having hearing loss ranged from 80 to 89%. Conversely, the percentage of non-hearing impaired people who would be correctly identified as not having hearing loss ranged from 37 to 43. ROC curves incorporate sensitivity and specificity measures, indicating the ability of the questionnaire to correctly classify those with and without the disease. ROC curve analysis of this sample suggested that the level of performance of the questionnaire was fair to poor.  These findings suggest the need to use alternative screening methods to detect risk of hearing loss.
There was generally low concordance between most questionnaire scores and audiogram results. Three items (1, 4, and 7) showed moderate agreement with mid-frequency audiogram results, suggesting these items may be more useful as screening questions. These findings suggest the potential to select items from the NIDCD questionnaire to develop a revised version that may be useful as an alternative screening method to detect risk of hearing loss and identify the need for medical evaluation of hearing sensitivity, particularly among at-risk and underserved populations.
The results of the study reported here can be compared to those of other researchers. The two-item Gomez  instrument ("Which statement best describes your hearing in your left/right ear without a hearing aid?") performed better than the NIDCD questionnaire, producing agreement of 55% (as measured by Kappa statistic) in mid-range frequencies, as compared to individual NIDCD items (Kappa range = 0.15 to 0.41). Sindhusake's  one-item ("Do you feel you have a hearing loss?") questionnaire also had higher sensitivity (but lower specificity) than the NIDCD questionnaire. Other studies found smaller AUC,  and lower sensitivity , than the NIDCD questionnaire.
The sample showed a high rate of risk for hearing loss, with 28 to 57% of participants demonstrating screening threshold averages above 25 dB, depending on frequency range. This high rate of risk for hearing loss is not unexpected, given the high rate of work-related noise exposure of this worker group, the high rate of hearing loss reported in previous studies, ,,,, and the low rate of use of hearing protection reported in this population. ,,
Although farm operators work in a hazardous noise environment, hearing health services (such as hearing conservation programs and audiometric testing) are not generally offered. Moreover, most farm operators experience hearing loss before their risk for NIHL is identified. While development of policies to deliver hearing health services to this high risk group is indicated, implementation of this is not anticipated in the near future. Meanwhile, the use of a self-administered questionnaire may present a simple, low cost opportunity to assess the hearing health of workers not included in a hearing conservation program, and to initiate a referral for medical evaluation of hearing. The results of the study reported here suggest that the performance of screening questionnaires vary significantly, and while less desirable than audiometry, the Gomez and Sindhusake instruments might perform the best in this group.
This study used a small, convenience sample of farm operators, limiting the generalizability of results. Screening audiometry was conducted in field conditions that were less controlled than that found in clinical settings where diagnostic testing is conducted. Consequently, the screening audiometry results may have included some biases resulting from measurement or other errors such as occasional background noise. Further, the qualifications of examiners were not comparable to that of licensed and experienced professional audiologists or other hearing specialists, possibly introducing additional errors of measurement. The study should be repeated using standardized clinical equipment and procedures with licensed audiologists, and should include an evaluation of its cost effectiveness. Furthermore, the study cannot determine the nature or cause of risk of hearing loss, e.g., presbycusis or other pathology.
Despite sampling limitations, the demographics of the sample closely matched those of the population of farm operators in the region.  This study is one of the few studies examining the performance of self-reported measures of hearing ability among farm operators, who are known to experience high rates of hearing loss and to be underserved by programs designed to conserve workers' hearing.
| References|| |
|1.||Agrawal Y, Platz EA, Niparko JK. Prevalence of hearing loss and differences by demographic characteristics among US adults: Data from the national health and nutrition examination survey, 1999-2004. Arch Intern Med 2008;168:1522-30. |
|2.||Tak S, Calvert GM. Hearing difficulty attributable to employment by industry and occupation: An analysis of the national health interview survey-United States, 1997 to 2003. J Occup Environ Med 2008;50:46-56. |
|3.||Conway H, Simmons J, Talbert TThe purposes of occupational medical surveillance in US industry and related health findings. J Occup Med 1993;35:670-86. |
|4.||Suter AH. The hearing conservation amendment: 25 years later. Noise Health 2009;11:2-7. |
|5.||USDA. Farm labor. Available from: http://usda.mannlib.cornell.edu/usda/nass/FarmLabo//2000s/2008/FarmLabo-08-15-2008.pdf. [Last accessed on 2009 Nov 5]. |
|6.||USDA. Briefing rooms: Rural labor and education: Farm labor. Available from: http://www.ers.usda.gov/Briefing/LaborAndEducation/FarmLabor.htm#Numbers. [Last accessed on 2010 Dec 9]. |
|7.||Tak S, Davis RR, Calvert GM. Exposure to hazardous workplace noise and use of hearing protection devices among US workers-NHANES, 1999-2004. Am J Ind Med 2009;52:358-71. |
|8.||Beckett WS, Chamberlain D, Hallman E, May J, Hwang SA, Gomez M, et al. Hearing conservation for farmers: Source apportionment of occupational and environmental factors contributing to hearing loss. J Occup Environ Med 2000;42:806-13. |
|9.||Carruth A, Robert AE, Hurley A, Currie PS. The impact of hearing impairment, perceptions and attitudes about hearing loss, and noise exposure risk patterns on hearing handicap among farm family members. AAOHN J 2007;55:227. Available from: http://proquest.umi.com.proxy.lib.umich.edu/pqdweb?did=1279986041andFmt=7andclientId=17822andRQT=309andVName=PQD. [Last accessed on 2011 Apr 24]. |
|10.||Gates DM, Jones MS. Populations at risk across the lifespan: Case reports: A pilot study to prevent hearing loss in farmers. Public Health Nurs 2007;24:547-53. Available from: http://proquest.umi.com.proxy.lib.umich.edu/pqdweb?did=1279986041andFmt andFmt=7andclientId=17822andRQT=309andVName=PQD. [Last accessed on 2011 Apr 24]. |
|11.||Jenkins PL, Stack SG, Earle-Richardson GB, Scofield SM, May JJ. Screening events to reduce farmers' hazardous exposures. J Agric Saf Health 2007;13:57-64. |
|12.||Thelin JW, Joseph DJ, Davis WE, Baker DE, Hosokawa MC. High-frequency hearing loss in male farmers of missouri. Public Health Rep 1983;98:268-73. |
|13.||Gomez MI, Hwang SA, Sobotova L, Stark AD, May JJ. A comparison of self-reported hearing loss and audiometry in a cohort of New York farmers. J Speech Lang Hear Res 2001;44:1201-8. |
|14.||Stewart M, Scherer J, Lehman ME. Perceived effects of high frequency hearing loss in a farming population. J Am Acad Audiol 2003;14:100-8. |
|15.||Marvel ME, Pratt DS, Marvel LH, Regan M, May JJ. Occupational hearing loss in New York dairy farmers. Am J Ind Med 1991;20:517-31. |
|16.||Nondahl DM, Cruickshanks KJ, Wiley TL, Tweed TS, Klein R, Klein BE. Accuracy of self-reported hearing loss. Audiology 1998;37:295-301. |
|17.||Sindhusake D, Mitchell P, Smith W, Golding M, Newall P, Hartley D, et al. Validation of self-reported hearing loss: The Blue Mountains Hearing Study. Int J Epidemiol 2001;30:1371-8. |
|18.||Gomez MJ, Hwang SA, Sobotova L, Stark AD, May JJ. A comparison of self-reported hearing loss and audiometry in a cohort of New York farmers. J Speech Lang Hear Res 2001;44:1201-8. Available from: http://proquest.umi.com//pqdweb?did=98156130andFmt=7andclientId=17822andRQT=309andVName=PQD. [Last accessed on 2011 Apr 24]. |
|19.||Okamotoa M, Nakanishi N, Tatara K. The relationship between self-reported difficulty with hearing in the worse ear and the better ear. Clin Otolaryngol 2005;30:16-20. |
|20.||Choi SW, Peek-Asa C, Zwerling C, Sprince NL, Rautiainen RH, Whitten PS, et al. A comparison of self-reported hearing and pure tone threshold average in the Iowa farm family health and hazard survey. J Agromedicine 2005;10:31-8. |
|21.||NIDCD. Ten ways to recognize hearing loss. Available from: http://www.nidcd.nih.gov/health/hearing/10ways.htm. [Last accessed on 2011 Apr 24] |
|22.||Fawcett T. An introduction to ROC analysis. Pattern Recognit Lett 2006;27:861-74. |
|23.||Hanley JA, McNeil BJ. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology 1982;143:29-36. |
|24.||McCullagh MC, Lusk SL, Ronis DL. Factors influencing use of hearing protection among farmers: A test of the Pender health promotion model. Nurs Res 2002;51:33-9. |
|25.||McCullagh MC, Ronis DL, Lusk SL. Predictors of use of hearing protection among a representative sample of farmers. Res Nurs Health 2010;33:528-38. |
|26.||USDA. 2007 census of agriculture. Available from: http://www.agcensus.usda.gov/Publications/2007/Online_Highlights/Fact_Sheets/demographics.pdf. [Last accessed on 2009 Oct 5]. |
Marjorie C McCullagh
400 N. Ingalls Street, Suite 3182, Ann Arbor, MI 48109
Source of Support: None, Conflict of Interest: None
[Figure 1], [Figure 2], [Figure 3]
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5]