How many individuals in the United States have speech, voice, swallowing, and/or language problems? The answer depends on how the terms are defined. It is difficult to specify precise incidence and prevalence data when terms used to describe communication and swallowing disorders are inconsistent. For example, there is controversy about what to call (central) auditory processing disorder and how to define it. Various terms used include auditory processing disorders, central auditory processing disorders (without the parentheses), receptive language disorder, and information processing disorder.
ASHA's 2005 Technical Report on (Central) Auditory Processing Disorders states that diagnosis of (C)APD "generally requires performance deficits on the order of at least two standard deviations below the mean on two or more tests in the battery." What constitutes the test battery, however, can and does vary, as "the selection of a (central) auditory test battery should be individualized and based on the referring complaints and additional information obtained." Our relative lack of precise terminology can have consequences for the discipline, and we would be well served to examine whether we are doing everything we can to understand the number and characteristics of people who stand to benefit from speech-language pathology or audiology services.
When Terms Are Inconsistent
Ironically, one of the consequences of our frequent failure to agree on what we mean when we refer to a client as having (C)APD, autism, or a moderate hearing loss, is that the disagreement hampers our ability to communicate about these important professional issues. If professionals within our discipline do not necessarily mean the same thing when using some of these terms, imagine the confusion that can ensue when we discuss clients with professionals from other disciplines, or communicate with a client or caregiver attempting to learn more about the disorder.
A second consequence is our inability to evaluate public health and other preventive efforts to combat these disorders. The establishment of baseline estimates of the incidence or prevalence of a disorder is a necessary step to assessing whether such efforts have made a difference. If we don't know how many people previously had the disorder, it is impossible to know that impact.
At the level of an individual client, research that doesn't fully describe the characteristics of the study participants is limited in its ability to help us determine which treatments are likely to work for which clients under which circumstances.
A third consequence is the effect on our ability to document the scope of communication and swallowing disorders as we compete for finite health care and/or educational resources. The federal Centers for Medicaid and Medicare Services stated, "Based on prevalence, health care expenditure, and ability to make a positive impact through evidence-based approaches, the Center for Medicaid and State Operations, in partnerships with states, will develop additional quality improvement program options."
Finally, the discipline of communication sciences and disorders (CSD) depends upon basic, translational, and clinical practice research if the knowledge base is to evolve. Inconsistent use of terminology has long stood as a major barrier to the ability of individual research studies to coalesce into a more meaningful body of literature. ASHA's National Center for Evidence-Based Practice recently completed an evidence-based systematic review that highlights the problem of inconsistent terminology in research. The review focused on studies of hearing loss in persons who receive treatment with the drug gentamicin for an infection. Two of the 17 reviewed studies investigated how the incidence of hearing loss varied when different definitions were applied. One study found that the incidence of hearing loss in its subjects could vary from 8% to 58%, depending on which definition of hearing loss was used (Whatley, 2006). The remaining 15 studies each used a single definition, but rarely the same one. Eleven different definitions of hearing loss were used in the studies, with no single definition common to more than three studies (see Table 1 [PDF]). This variability makes it difficult, if not impossible, to compare findings across studies.
A comparison of the Chong (2003) and Indudharan (2005) studies illustrates the challenges presented by different definitions. Chong used a much more restrictive definition of hearing loss—a loss of at least 30 dB at one frequency—relative to the much broader definition used by Indudharan, who classified hearing loss as a loss of 5 dB or more. The Chong study involved children with urinary tract infections who received gentamicin intravenously; it observed hearing loss in none of its 161 subjects. The Indudharan study involved children and adults with chronic otitis media who received gentamicin in ear drops. Of the study's 95 participants, 28 (29.5%) were observed to have a hearing loss subsequent to gentamicin administration.
Was the markedly higher rate of hearing loss in the Indudharan study due to some characteristics of the participants (e.g., children versus adults; otitis media versus urinary tract infection), the route of administration (e.g., intravenous versus drops), or some other factor (e.g., dosage schedule)? Even under ideal circumstances, simply comparing the studies wouldn't yield a definitive answer. Yet, with such wide variation in how hearing loss is generalized in these two studies, it is difficult to know whether the discrepancy may be due to these or other factors.
With the different definitions of hearing loss used across the two studies, however, we can't say that there was a greater prevalence of hearing loss among participants in the Indudharan study than in the Chong study. We know that none of the Chong participants had a loss of 30 dB or more. It is conceivable, however, that some—or even many—of the participants could have had a hearing loss of between 5 dB and 25 dB, levels that would have categorized them differently in the Indudharan study. Perhaps if the Indudharan definition had been applied to the Chong participants, a similar rate of roughly one-third of subjects with hearing loss would have been noted. However, it is also plausible that the rate would have been substantially higher or lower.
Viewing the studies from the opposite perspective, it is equally impossible to translate the findings from the Chong study to the Indudharan participants. Although we know that 29 of the Indudharan participants had a loss of at least 5 dB, we don't know whether all, some, or none of them had loss of at least 30 dB, as the exact amount of loss for each participant was not provided.
Next Big Step
One obvious answer to the confusion is the development and use of standardized diagnostic criteria, or what epidemiologists would call case definitions. This idea is not new, as ASHA has been working on terminology issues since at least the 1930s (Stinchfield, 1931) and has been a member of the International Group on Terminology Frameworks–Communication Sciences and Disorders (IGOTF-CSD) since the group was created in 2005. An ASHA staff member serves on a National Association of Insurance Commissioners working group developing standard definitions for medical and health insurance terms, as mandated by the Patient Protection and Affordable Care Act.
Although a manual of agreed-upon definitions has obvious appeal, this route need not be the only way to address the problem. Recent technological advances, particularly the growth of online publishing, afford researchers the opportunity to share much more data with readers than has traditionally been the case with print media. In addition to reporting x number of y participants had a hearing loss of at least z decibels, a researcher could submit complete information about each participant in supplementary material that then becomes part of the archival version of the article. Regardless of which definition the researcher had chosen, readers would have the raw data available to characterize the findings according to whatever definition they choose (see Table 2 [PDF] for an example).
Making these data available would remove much of the pressure to develop consensus definitions that researchers would feel forced to use, and it could serve as the impetus for the eventual refinement of definitions that could be agreed upon by all at some later point. Although this approach would be simpler than developing consensus definitions for every aspect of CSD, it would not be simple. A great deal of effort would be required to establish—and keep current—the universe of potential diagnostic indicators that speech-language pathologists and audiologists consider when making a diagnosis. Indicators for measuring hearing loss would include not only the decibels, but potentially the particular instrumentation and its calibration standard, whether repeat or single measurements are required, and other sources of information relevant to making the diagnosis. For other disorders, such as aphasia, individual scores on particular tests and even subscores could be reported. Connected speech and language samples also could be included as supplementary material, either in video or audio formats (with permission of study participants and Institutional Review Board approval).
Not all researchers would necessarily use the same indicators, let alone interpret them in the same way, so seamless comparisons across studies would be far from assured. Making this additional information available, however, would help to further our understanding of the disorders faced by our clients. Readers and reviewers will expect—and require—as much information about individual participants as possible. This approach to delineating detailed patient data applies equally to treatment research in which the use of supplementary material is likely to similarly revolutionize treatment advances.