Health Care Utilization and Adults Who Are Deaf: Relationship with Age at Onset of Deafness

Abstract

Objectives

To evaluate the health care utilization of a nationally representative sample of U.S. deaf adults while accounting for the age at onset of deafness, an indicator of linguistic and sociocultural group affiliation.

Data Sources/Study Setting

Data from the 1990 to 1991 National Health Interview Surveys, the most recent years the Hearing Supplement was administered. The data were collected during in-home interviews of a sample of the U.S. civilian noninstitutionalized population.

Study Design

Cross-sectional analyses comparing health-related measures of adults deafened before (prelingually) and after (postlingually) the age of 3 and those of a representative sample of the general population, adjusting for sociodemographics and health status. Key measures were physician visits and preventive health care services utilization.

Principal Findings

Compared with the general population, prelingually deafened adults had fewer physician visits and were less likely to have visited a physician in the preceding 2 years, whereas postlingually deafened adults had more physician visits and were more likely to have visited a physician in the preceding 2 years. Postlingually deafened women were less likely to have had a mammogram within the previous 2 years.

Conclusions

In terms of health care utilization, the deaf population is heterogeneous. Prelingually deafened adults' use of health care is similar to that of other language minority groups. Postlingually deafened adults' use of health care services appears similar to people with chronic illness. Future studies must distinguish different groups of people with hearing loss in order to identify barriers and monitor improvements in health care services access.

Hearing impairment, with a prevalence of 9.35 percent, is the sixth most common chronic condition in the civilian, noninstitutionalized U.S. population (Collins 1997). Of the 20.3-million persons with hearing trouble, 4.8 million cannot hear and understand normal speech (Ries 1994). American Sign Language, the language of the U.S. deaf community, is the third most commonly used language in the United States (Lotke 1995). The prevalence of hearing trouble is increasing faster than the U.S. population is aging (Ries 1994). Despite the size of this population, there is limited information regarding the health services utilization of its constituents. Understanding the health care needs of people with hearing loss has gained legal implications since the passage of the Americans with Disabilities Act.

Prior studies of health care and deaf people have produced conflicting results. Some studies indicate that people with hearing loss have poorer health status and increased health services utilization (Kurz et al. 1991; Ries 1982; Zazove, Niemann, Gorenflo, et al. 1993). Other studies suggest that people with hearing loss have decreased health services utilization (Ebert and Heckerling 1995b; Pollard 1994). Inconsistencies in these findings may be related to how the study populations are defined. Studies have ignored sociodemographics (Ries 1982), the difference between unilateral and bilateral hearing loss (Kurz et al. 1991), or the age at onset of deafness (Kurz et al. 1991; Pollard 1994; Ries 1982; Zazove, Niemann, Gorenflo, et al. 1993). Other limitations include small sample size (Ebert and Heckerling 1995b; Zazove, Niemann, Gorenflo, et al. 1993) and limited generalizability (Ebert and Heckerling 1995b; Pollard 1994).

The relative lack of data and the inconsistencies of previous findings make it difficult to identify and address the health care needs of deaf and hard-of-hearing people. Although typically studied as a single homogeneous population, people with hearing loss form distinctive subpopulations based on sociocultural characteristics (Benderly 1980; Phillips 1996). The age at onset of a severe bilateral hearing loss can have a significant effect on communication and socialization. People deafened after early adulthood and completion of their basic education are more likely to communicate well in English. People deafened prelingually, before the development of language (usually considered to be before the age of 3), are more likely to communicate using a sign language. These different communication modes have significant social implications for their users. They may also have an impact on health services utilization and health status.

We hypothesized that age at onset of deafness differentially affects language barriers and social isolation, which, in turn, influence health services utilization. The Andersen Behavioral Model of Health Services Use explains the interaction of these factors (Andersen 1995). Using this model, language barriers are an impeding factor to health care services use, and social isolation is a predisposing factor for health care services use. We theorized that adults deafened later in life, who are more likely to communicate primarily in English, would find their hearing loss socially isolating. People who are socially isolated are more likely to visit the health care system (Kouzis and Eaton 1998). In contrast with postlingually deafened adults, prelingually deafened adults are more likely to communicate primarily in a sign language, socialize mostly with other deaf people, and not find their hearing loss socially isolating. Because of language barriers, however, we theorized that prelingually deafened adults would have less access to the health care system. Having a common language with health care practitioners is an enabling factor for health care services use. Members of other language minority groups are less likely to visit physicians (Horberger, Gibson, Wood, et al. 1996; Woloshin, Bickell, Schwartz, et al. 1995). In the analyses presented here, we tested two hypotheses based on these theories: (1) When compared with the general population, prelingually deafened adults would have lower health care services utilization. (2) When compared with the general population, postlingually deafened adults would have higher health care services utilization.

METHODS

The National Health Interview Survey (NHIS) collects data on the civilian noninstitutionalized population of the United States. Using in-home interviews, these data are obtained in a continuing nationwide sample of households. The sampling follows a multistage probability design and includes oversampling of minority populations. Members of the armed forces, U.S. nationals living abroad, homeless people, and institutionalized persons, including nursing home residents, are excluded from the data collection. Information collected includes sociodemographic and health characteristics of people living in the surveyed households. In addition to core questions asked of all participants, sets of questions are administered to randomly sampled subsets of participants. The overall response rate for the NHIS has been 95 to 98 percent (Ries 1994).

In 1990 and 1991, the NHIS core questionnaire included questions related to hearing ability in the Hearing Supplement. People were identified as having hearing loss using three screening questions. For people with hearing loss, the supplement included two scales to rate hearing ability, as well as questions regarding the age at onset of hearing loss. In our analyses, the responses to these questions were used to define the study populations.

One hearing scale used was the self-rated scale (SRS). People rate the hearing ability of each ear without the use of hearing aids on a four-point scale, ranging from “good” to “deaf.” For people older than 3 years of age, a second scale, the Gallaudet Hearing Scale (GHS), was also administered. This five-point scale rates how well a person can usually hear and understand speech without the use of hearing aids, ranging from the “ability to hear and understand whispered speech” to the “inability to hear or understand any speech.”

For those with hearing difficulties, respondents were asked to identify the age at which the hearing problems began or the age at which they became deaf, as well as the age range for the onset of hearing trouble or deafness (before or after the 19th birthday and before or after the 3rd birthday).

In this study, deafness is defined using criteria established in previous demographic analyses (Gentile 1975; Ries 1994). Individuals are defined as deaf if they (1) indicated on the SRS that they at least “had a lot of trouble hearing” in both ears (a SRS score of 3 or greater in each ear) or (2) indicated on the SRS that, at best, they had “a little trouble hearing” in their better ear and indicated on the GHS that they could not hear and understand any speech (a SRS score of 2 or more in each ear and a GHS score of 5). The members of the resulting “deaf study population” have bilateral hearing loss that interferes with understanding speech.

For analyses in this study, the deaf study population is subdivided by age and age at onset of hearing loss. Because children access the health care system through their parents or other adults, only data on adults (19 years and older) were analyzed. Within the deaf community, age at onset of hearing loss often predicts communication mode and social group (Benderly 1980). The relative frequencies at different ages of hearing loss onset are illustrated in a histogram of the responses of deaf adults to the NHIS question regarding the exact age at onset of hearing loss (n = 2,449; Figure 1). Two peaks are evident, one before the age of 3 and a second after the age of 60. We divided the deaf study population data into two subpopulations based on whether the hearing loss is prelingual or postlingual (whether it occurred before or after age 3 years). For this dichotomy, we used the responses of deaf adults to the NHIS questions regarding the age range for hearing loss onset (n = 2,728; Table 1). People (n = 12) in the third age-at-onset category (hearing loss before their 19th birthday but unsure whether it occurred before or after age 3 years) were included in the postlingual group for the analyses.

Figure 1.

Percent Distribution of Age at Onset of Deafness for Deaf Adults

Table 1.

Range of Age at Onset of Deafness: Distribution for Deaf Adults

The two deaf study populations were compared with a control group comprised of the nondeaf adults with hearing problems and adults without hearing problems. Because of the size of the data set, a 20 percent random sample of the adults without hearing problems was used. Weights were adjusted appropriately.

Analyses

Because the NHIS uses a complex multistage stratified probability design to sample households, the statistical package SUDAAN (Research Triangle Institute, 1992) was used for the analyses reported here. The SUDAAN program uses a Taylor series approximation method to compute variances that allow adjustment for the multistage probability sampling strategy. The weights provided on the NHIS public use tapes were used to adjust for survey oversampling and nonresponse to yield population estimates in the reported analyses. Chi-square tests and analysis of variance were used to examine the univariate relationships between each variable and hearing status (categorized as normal hearing, prelingual deafness, or postlingual deafness).

Sociodemographic variables included age, sex, race (dichotomized as Caucasian or not), place of residence (dichotomized as whether or not residence is in a metropolitan statistical area defined by the U.S. Office of Management and Budget), marital status (dichotomized as whether or not the individual is currently married but not separated), telephone status (dichotomized as having a telephone or not), education level (categorized as less than 12 years, high school graduate, or beyond high school), and family annual income (four categories: less than $10,000, $10,000–$19,999, $20,000–$34,999, and $35,000 or more). Health care utilization variables included whether a doctor had been seen within the preceding 2 years; the number of doctor visits within the preceding year (natural logarithm is used to normalize the distribution); whether the individual had a regular source of health care; whether the individual had private health insurance in the past month; whether blood pressure had been measured by a health care professional within the preceding 2 years; whether the cholesterol level had ever been tested; whether women had a Pap smear performed within the preceding year; for women 50 years of age and older, whether a mammogram had been performed within the preceding 2 years; and whether a dentist had been seen within the preceding year. Several analyses also adjusted for overall self-reported health status, using the Healthy People 2000 Years of Healthy Life (YHL) measure (Centers for Disease Control and Prevention [CDC], 1995).

The YHL measure is constructed from questions on the annual NHIS (Adams and Benson 1992) and is used by the Public Health Service to track the health of the nation during the current decade. The measure assesses health on a continuum from 0.0 (death) to 1.0 (optimal health). The YHL measure is composed of two health domains—self-rated health and role limitations—and considers age and social role (CDC 1995). The measure exhibits reasonable validity. It has been found to produce consistent and plausible estimates of the effects on the quality of life of smoking and smoking cessation, in different age groups (Fiscella and Franks 1996). It produces health-related quality-of-life scores with reasonable face validity for a nationally representative sample of persons with a wide range of morbidities; the condition scores also exhibit high correlations with those obtained from other studies (Gold et al. 1998).

The impact of the age at onset of deafness on health care utilization, adjusting for sociodemographic factors and the health status variable, was examined in a series of regression analyses. Prelingual and postlingual deafness were included in the regression analyses as two dummy variables, with the general population as the reference group. Analyses adjusted for sociodemographic factors and health status variables that were statistically significant or affected the parameter estimates for the effects of prelingual or postlingual deafness by 10 percent or more. Logistic regression analyses were used for dichotomous dependent variables, and ordinary linear regression was used for the continuously dependent variables.

Because of the difference between the mean age in the postlingually deafened group and the hearing group, further analyses examined the effect of age–interaction variables (age squared and the interaction between age and other key variables). When the interaction between age and the deaf category was significant, additional analyses, stratified by age group, were conducted.

RESULTS

The distribution of sociodemographic and health-related variables for the prelingually deafened group, postlingually deafened group, and the hearing group are shown in Tables 2 and 3. Postlingually deafened adults were more likely to be men and Caucasian than the general population. Postlingually deafened adults were also older and reported lower overall health status and more physician visits in the preceding year than the general population. Prelingually deafened adults also reported lower overall health status but did not have more physician visits. Table 4 shows the results of the multivariate analyses with postlingually deafened adults and hearing adults. Postlingually deafened adults were more likely to have seen a doctor in the preceding 2 years and had more physician visits than hearing adults; the association of increased visits and postlingual deafness was much more pronounced in younger adults. The increased number of physician visits is not statistically significant when analyses are adjusted for health status. Postlingually deafened women 50 years and older were less likely have had a mammogram regardless of health status. There were no statistically significant associations between postlingual deafness and the other health care utilization variables.

Table 2.

Sociodemographic Characteristics of the Deaf and Hearing Samples

Table 3.

Health Care Utilization and Health of the Deaf and Hearing Samples

Table 4.

Adjusted Relationship Between Health Care Utilization and Postlingual Deafness

Results of analyses that examine the influence of health status on the health care utilization of prelingually deafened adults are presented in Table 5. Prelingually deafened adults had fewer doctor visits in the preceding year and were less likely to have had a doctor visit in the preceding 2 years than hearing adults in analyses that adjusted for health status. There were no statistically significant associations between prelingual deafness and the other health care utilization variables.

Table 5.

Adjusted Relationship Between Health Care Utilization and Prelingual Deafness

DISCUSSION

The results reported here may account for some of the discrepancies observed in previous studies (Ebert and Heckerling 1995b; Kurz et al. 1991; Pollard 1994; Ries 1982; Zazove, Niemann, Gorenflo, et al. 1993). Our results support our hypotheses and demonstrate that prelingually deafened adults and postlingually deafened adults represent distinct subpopulations in terms of health care utilization. These differences in health care utilization may reflect differences in sociocultural group affiliation.

Prelingually deafened adults had fewer doctor visits in the preceding year and were also less likely to have seen a doctor in the last 2 years. The lower health care services utilization persisted even after adjusting for the lower education level and lower income known to be associated with hearing loss (Ries 1994). Although their poorer health status resulted in more physician visits, prelingually deafened adults did not go to the doctor as often as expected based on their poorer health status. That language barriers play a role in decreased utilization of health care is supported by findings from other studies. In a study of the health care practices of deaf adults (73 percent prelingually deafened), deaf patients in a primary care program with sign language-accessible services had more physician visits and better compliance with health maintenance recommendations than deaf people who sought health care elsewhere (MacKinney et al. 1995). In a survey of deaf adults (66 percent prelingually deafened) regarding their health care experiences, 80.5 percent suggested the use of a sign language interpreter as a means to improve communication between hospital staff and themselves (Schein and Delk 1980). Like prelingually deafened adults, other language minority groups have fewer physician visits and poorer health status (Hornberger, Gibson, Wood, et al. 1996; Woloshin, Bickell, Schwartz, et al. 1995).

Postlingually deafened adults had higher health services utilization when compared with the general population. Others have reported social isolation and depression associated with postlingual deafness (Cooper 1976; Jones and White 1990). Because the NHIS in 1990 and 1991 did not include measures of depression or social isolation in a significant number of households surveyed, we were unable to test the effects of these factors on health care utilization. It is possible physician visit frequency is higher because the medical needs of postlingually deafened adults were inadequately addressed. The lower mammography compliance found with postlingually deafened adults suggests that some of their specific medical needs are not being met. The presence of some chronic health problems in older adults is associated with lower levels of cancer screening and more physician visits (Fontana et al. 1997). This may also be true with deafness; if physicians and deaf patients need to spend more time trying to communicate, there will be less time to discuss recommended preventive services.

The strength of the relationship between postlingual deafness and number of doctor visits decreased with age, which is probably related to the associated increase in non–hearing-related health care needs with aging. The same pattern is seen when analyses adjust for health status, although the differences between postlingually deafened adults and hearing adults are not statistically significant. Adjusting for health status also decreases the difference between postlingually deafened adults and hearing adults in terms of likelihood of not having seen a doctor in the preceding 2 years, although the difference remains significant. These findings suggest, in part, the increased utilization of postlingually deafened adults is related to their lower health status.

Limitations

The categories used in this study are likely to result in some misclassification bias. For example, not all people deafened prelingually (before age 3 years) use American Sign Language; thus, their sociocultural ties to the deaf community may be weaker than those of ASL users (Ebert and Heckerling 1995a). The analysis also does not examine the possibility that additional categories based on the age at onset of hearing loss might reveal insight into significant differences among postlingually deafened adults. However, the categories of people with hearing loss used in these analyses yield a more socially valid way to study the deaf population than used in many previous studies of health care services utilization. Categories of people with hearing loss based on hearing loss severity and age at onset of hearing loss have been used in health care-related analyses elsewhere (Cooper 1976; Ebert and Heckerling 1995b; Jones and White 1990; Lass et al. 1978; McEwen and Anton-Culver 1988), including studies with NHIS data (Gentile 1975; Ries 1994). Diversity in the sociocultural identity of people with hearing loss has also been discussed previously (Padden and Humphries 1988). Age at onset of hearing loss is a good surrogate measure of the preferred communication mode of deaf people. Preferred mode of communication might be a more valid way to categorize deaf people for the purpose of these analyses, but those data are not available. The findings of this study suggest that more direct questions about modes of communication of deaf persons are warranted in future surveys. It is unlikely that survey nonresponse or the use of self-report produce significant bias. The overall nonresponse rate of the NHIS is 4.4 percent; 2.7 percent is a result of respondent refusal, and the remainder is a result of failure to locate the respondent (Ries 1994). Self-reported health status (Stewart, Hays, and Ware 1988) and self-reported hearing status on the NHIS Hearing Supplement (Schein, Gentile, and Haase 1970) have been shown to be valid measures.

The NHIS is a cross-sectional study; causality cannot be assessed. It is unlikely, however, that the health characteristics analyzed caused the different categories of hearing loss. It is possible that unmeasured variables explain the findings. For example, some comorbid conditions may cause deafness and affect health care utilization. The adjustments used for health status as well as sociodemographics should limit biases due to confounding.

Some NHIS questions regarding health characteristics were asked only of a subset of households in the survey. The resulting small samples for deaf population categories, particularly in the prelingually deafened respondents, limit the power to detect differences among the deaf and general populations. Conclusions about analyses finding no differences among the general and deaf populations may be misleading.

Conclusions from statistically significant results of analyses may be questioned because of the small unweighted sample size for the prelingually deafened study population. Because these data are from a nationally representative sample, these analyses are valuable despite the sample size limitations. Future surveys could oversample for people with deafness.

Insufficient data were available to allow adjustment for health insurance status. Approximately one third of survey respondents were asked about private health insurance, and no data are available regarding other health insurance types. No information is available from other sources on health insurance coverage of prelingually or postlingually deafened adults, although Purohit, Zazove, and Gorenflo (1993) mentioned that deaf people had difficulty obtaining life insurance. Future studies could collect information about health insurance coverage among deaf people.

The health status index used may not be entirely independent of deafness. The YHL measure includes role limitations in major activities and instrumental activities of daily living (CDC 1995), both of which may be affected by deafness. Thus, analyses adjusting for health status may have resulted in overadjustment.

The NHIS excludes institutionalized persons. Results from this study do not directly measure the health and health care utilization of older nursing home residents, 22 percent of whom are hearing impaired (Dey 1997).

CONCLUSIONS

This study demonstrates, in terms of health care utilization, that deaf people are a heterogeneous population. Prelingually deafened adults use fewer health care services, which is similar to other language minority groups. Postlingually deafened adults use health care services similar to people with chronic illness. These 1990 and 1991 data can be used as a baseline to measure changes in health care access for deaf people resulting from implementation of the Americans with Disabilities Act of 1990. Some results, such as lower mammogram compliance among postlingually deafened women, highlight areas needing improvement. As the geriatric population increases and the large cohort of people deafened during the rubella pandemic of 1964 (Benderly 1980) age, a better understanding of the specific health care needs of deaf people will become increasingly important.

To help explain some of the differences reported here, researchers can examine each of the distinct subpopulations of people with hearing loss using qualitative methodologies. To monitor health care needs and utilization, future surveys and other quantitative data sets should collect information that allow for analyses to consider the different subpopulations of deaf people. Oversampling of deaf people on national surveys may help in analyses that examine health insurance coverage, site of access to health care services, and racial and ethnic minorities within the deaf populations. Qualitative and quantitative studies can help determine how many subpopulations of deaf people should be considered in health analyses and whether age at onset of deafness, preferred mode of communication, or a combination of these and other factors is the best way distinguish these subpopulations. It will also be valuable to learn the perspectives of health care practitioners on the challenges of providing equitable access to health care services for their deaf and hearing patients. As we gain a better understanding of the relationship of deafness, health, and health care, policy makers, administrators, and clinicians working with members of the different deaf communities will be able to develop and monitor programs designed to improve the health of deaf people. Until the appropriate studies are done, decision makers should be aware that some research may be misleading because it does not recognize the distinct subpopulations of people with hearing loss and because some data collection methodologies may exclude certain segments of the deaf population (Barnett and Franks 1999).