Cross-Check Principle
Introduction :
A test battery approach is especially useful with special populations. When limited information is obtained from traditional tests, converging evidence acquired from a carefully selected test battery helps in developing a “picture” of the patient’s auditory capabilities so that appropriate recommendations can be made. Obtaining valid information through this approach was first proposed by Jerger and Hayes (1976) and was labeled the “cross-check principle.” This principle takes into consideration that many patients may be able to provide only limited behavioral audiometric information, and the validity may be questioned. When this occurs, clinicians may use another test procedure such as an objective measure (i.e., immittance, otoacoustic emissions, ABR) to verify the behavioral results or to provide additional information about auditory status.
Jerger and Hayes (1976) proposed the cross-check principle over twenty years ago because of their legitimate concern as to the reliability of the behavioral tests available at that time. Given the poor performance of these tests, the probability of error was high. They recommended that behavioral results be confirmed by a physiological test, immittance or auditory brainstem response audiometry (called BSER by Jerger and Hayes), before making a final decision as to hearing status. Thus, their motivation was not that more tests are always better, but that in this particular situation, there must be some confirmation of behavioral results (Jerger, personal communication).
The cross-check principle actually consists of a concept and a specific test protocol.The concept was explained by Jerger and Hayes. They stated, “...our audiologic evaluation of the child does not stop with behavioral test results. We always insist that the behavioral result be confirmed by a crosscheck.”Thus, the basic concept is that one test result is used to confirm another test result. As a concept, the cross-check principle has a certain face validity. Requiring agreement among several tests increases confidence in the test results. In addition, disagreement among test results raises a flag that there may be a problem. These characteristics are particularly useful when evaluating a population where testing may be difficult and unreliable. That is why the cross-check principle has been particularly popular in pediatric evaluation. The cross-check principle is also a specific test protocol as recommended by Jerger and Hayes. In their own words, “We used impedance audiometry to confirm behavioral test results... In most cases, impedance audiometry serves as an effective cross-check of behavioral audiometry. If they disagree, however, BSER audiometry can serve as a further cross-check… We use BSER to crosscheck behavioral test results whenever impedance audiometry is noncontributory... We also use BSER to cross-check the results of impedance audiometry when behavioral testing yields no useful information.” Essentially, they proposed a three-test strategy. All individuals receive the first two tests, some type of behavioral test and immittance. The third test, BSER, is administered only when the first two tests disagree. Their cross-check protocol is shown in Figure 1.
Much has changed since Jerger and Hayes proposed the cross-check principle; however, the cross-check principle is still recommended for pediatric and, occasionally, adult testing. Recently, this author was reviewing a book chapter that discussed the use of test batteries for pediatric hearing evaluation and the importance of the crosscheck principle. Current guidelines for early hearing detection and intervention state, “Adequate confirmation of an infant’s hearing status cannot be obtained from a single test measure. Rather, a test battery is required to cross-check results of both behavioral and physiologic measures” (Finitzo et al., 2000). What has actually survived is the concept of the cross-check principle, the basic understanding that one test result is used to confirm another test result. What has not survived is the specific test protocol recommended by Jerger and Hayes. This is not surprising since that specific protocol was determined by the cost and availability of tests in the 1970s. Since that time, tests have been modified and new tests, such as otoacoustic emissions, have been introduced. There is no reason to believe that their specific protocol would be optimum today. Gone unnoticed is the fact that the protocol recommended by Jerger and Hayes differs from more traditional test protocols. Thus, the objective of this paper is to evaluate the particular test protocol described by Jerger and Hayes to determine if this testing strategy offers any advantages over more traditional test protocols. The objective is not to quantitatively evaluate the performance of the cross-check principle as implemented by Jerger and Hayes in the 1970s.The necessary data are probably not available and the need at that time to confirm behavioral results was evident. Before evaluating the crosscheck protocol, it is useful to consider more traditional test protocols.
TRADITIONAL TEST PROTOCOLS
EXAMPLE
I n this example, we are testing 1000 children with a prevalence of hearing loss of 10%; therefore, 100 of the 1000 children have hearing loss. To simplify this example, we will make two assumptions. First, the tests are uncorrelated; that is, the tests have zero correlation. Test correlation is the tendency of two test to identify the same individuals as positive or negative. Zero correlation means that no such tendency exists. Second, all tests have the same performance, HT/FA = 80%/15%. The results of testing are shown in Figure 1b. Of the 100 children with hearing loss, 64 test positive on both tests, 32 have one positive and one negative result, and 4 test negative on both tests. With a strict criterion (both tests must be positive), the hit rate would be 64% (64/100). With a loose criterion (only one positive result is required), HT = 96% (96/100). Results for the children without hearing loss are also shown in Figure 1b. With a strict criterion, False Alarm Rate is 2% (20/900). With a loose criterion, FA = 28% (250/900). These results are summarized in Table 1. As expected, a strict criterion reduces test protocol FA below that of the individual tests, which is good, but also reduces HT below that of the individual tests, which is bad. A loose criterion increases test protocol HT and FA above that of the individual tests. Thus, criterion can be used to manipulate HT and FA, but there is generally a trade-off: increasing HT increases FA and decreasing FA decreases HT. This trade-off is one reason why a test protocol may not be better than a single test. Which is better, the performance of the individual test (80%/15%), the performance of the protocol with a strict criterion (64%/2%), the performance of the protocol with a loose criterion (96%/28%)? There is no unequivocal answer to this question. The “best” testing strategy would depend on a number of factors including the particular testing objective, e.g., screening vs. diagnostic.
CROSS-CHECKING TEST RESULTS
The major reason that an audiologist uses a diagnostic battery is to be able to check the results of individual tests with each other. The idea that “the results of a single test are
cross-checked by an independent test measure” is referred to as the cross-check principle (Jerger and Hayes, 1976, ). Since the cross-check principle was first proposed, many manuscripts have revisited the concept as new diagnostic tests have been developed and different test batteries have been proposed to diagnose specific disorders. The goal of comparing the results of two or more tests is to increase the rate of correct identification of disorders (hit rate) and to decrease the rate of diagnosing a disorder when no disorder exists (false alarm rate) (Turner, 2003).
Cross-checks for Puretone Air Conduction
If you only obtained puretone air-conduction thresholds then you would not be able to accurately diagnose the type of hearing loss. Air-conduction audiometry is normally cross checked with bone-conduction audiometry or tympanometry to rule out a conductive component of the hearing loss. If a difference greater than 10 dB exists between the airconduction and bone-conduction thresholds at the same frequency, a conductive component is indicated. Similarly, air-conduction thresholds for an ear may be within normal limits; however, if a tympanogram for that ear falls outside of the norms for middle-ear pressure and compliance (e.g., Jerger Type B or Type C), a conductive component may be present. ARTs can reveal more information about the type of loss based on the pattern of responses obtained, thus serving as an additional cross-check for puretone air conduction.
Cross-checks for Puretone Audiometry
When puretone audiometry (air- and bone-conduction testing) suggests a significant air–bone gap, tympanometry and ARTs can be used to reinforce the diagnosis of the conductive element and to contribute to a specific diagnosis. OAEs also can be used as a cross-check of puretone audiometry. OAEs are used to assess the health of the outer hair cells of the cochlea, but their measurement may be affected by disorders in the conductive pathway. An audiologist might use OAEs as a cross-check to aid in potentially ruling out a nonorganic hearing loss, to verify outer hair cell function and the degree of cochlear hearing loss, and to further assist with the diagnosis of conductive components, auditory neuropathy spectrum disorder (ANSD), and other retrocochlear disorders. In addition, ARTs have been used to cross check puretone audiometry (Jerger et al., 1974), although other objective tests, such as tone-burst–stimulated auditory brainstem response (ABR), are considered to be better procedures for estimating hearing thresholds. Acoustic reflexes can be used to help identify the presence of hearing loss in young children as well as in adults with language and/or cognitive issues that may reduce the validity and reliability of behavioral measures (Hall, 2010). Acoustic reflexes can also be used to determine site of lesion within the auditory pathway, specifically in differentiating between cochlear and retrocochlear pathologies.
Cross-check for Puretone Average
A puretone average (PTA) is usually calculated as the average of the air-conduction thresholds at 500, 1,000, and 2,000 Hz for each ear. Normally, the PTA should agree with the speech recognition threshold (SRT), meaning that the PTA and SRT should be within 10 dB of one another in the same ear. One instance in which the audiometric thresholds may cause the PTA to be greater than the SRT by 10 dB is when the audiogram configuration is sharply sloping or sharply rising. In such instances, it is preferable to use a two-frequency PTA by averaging the two lowest (e.g., best) thresholds at 500, 1,000, and 2,000 Hz. The two-frequency PTA should then be in agreement with the SRT. Another instance in which the PTA and SRT may disagree is if a person is malingering or intentionally exaggerating a hearing loss. Outside of these special circumstances, we would expect SRTs and PTAs to be highly correlated (except when language or foreign language is a major factor). This allows us to use the SRT to validate the PTA (American Speech-Language-Hearing Association, 1988).
Considerations for Assessing Speech Understanding
One additional step that audiologists may take to address a patient’s complaint of not being able to understand speech in noisy environments is to administer a speech-in-noise test in addition to the word recognition testing in quiet. Although this is technically not a cross-check, the addition of a speech-in-noise test, especially with sentence stimuli, will provide a more realistic test environment to evaluate a common patient complaint. The puretone audiogram does not necessarily correlate with the amount of difficulty a listener will have in noise (Killion and Niquette, 2000). In addition, when word recognition testing is performed in quiet at a single speech presentation level, no guarantee exists that the test is measuring the patient’s maximum speech understanding (Wiley et al., 1995).
Cross-check Considerations for Pediatric Testing
For children, it is imperative that the audiologist utilize the cross-check principle. The behavioral responses obtained via behavioral observation audiometry (BOA) or visual reinforcement audiometry (VRA) are considered to be accurate reflections of a child’s true thresholds when these tests are conducted carefully (Madell and Flexer, 2008). However, because children often do not respond as consistently or as quickly as adults, it is possible that a child’s behavioral responses may still be elevated compared to actual thresholds. As a result, the audiologist may judge the child’s responses as unreliable (Baldwin et al., 2010). Regardless of the judged reliability of such measures, audiologists should use objective tests such as OAEs and tympanometry as cross-checks for pediatric behavioral responses (Baldwin et al., 2010; Littman et al., 1998; Madell and Flexer, 2008). In addition, OAEs and acoustic reflexes have been shown to be good cross-checks for ABR in young children (Berlin et al., 2010; Stach et al., 1993). The Joint Committee on Infant Hearing Position Statement (JCIH; American Academy of Pediatrics, 2007) also recommends that electrophysiological measures be employed as a cross-check for behavioral response audiometry for children younger than 6 months chronological age. The statement further stresses the importance of obtaining behavioral thresholds as soon as possible using the most age-appropriate method “to cross check and augment physiologic findings” (American Academy of Pediatrics, 2007, p. 910).
Electrophysiological Tests as Cross-checks
Although beyond the scope of this chapter, it should be noted that certain electrophysiological tests can be used to cross check behavioral measures, as well as to cross check each other and to help confirm diagnoses of certain disorders (Bachmann and Hall, 1998; Berlin et al., 2010; Gravel, 2002; Hall and Bondurant, 2009; Stapells, 2011). For example, Berlin et al. (2010) discussed the use of cross-checking test results to diagnose ANSD: “ . . . the presence of a [cochlear microphonic] or reversing waves at the beginning of the trace does NOT make a diagnosis of ANSD . . . without the cross-check of middle-ear muscle reflexes (MEMR), OAEs, and an ABR latency-intensity function” (p. 32). For further information about these tests, the reader is referred to the chapters that discuss electrophysiological tests in the text. Table 2 summarizes many of the cross-check tests that are used in audiology.
LIMITATIONS OF THE AUDIOLOGIC TEST BATTERY
The combination of well-validated test measures, precise patient instruction, careful scoring, and application of the cross-check principle should result in accurate diagnostic and rehabilitative decisions for most patients. It is important to remember, however, that real-world patients usually
do not present as textbook cases. The case studies contained in this chapter and the diagnostic criteria published in the audiologic test literature should be treated as guidelines rather than absolute rules. High-quality diagnosis depends on both the construction of a high-quality test battery and skill in interpreting ambiguous or seemingly contradictory test results. A good rule for daily practice is this: When test results seem in disagreement, first check the tester (rule out the clinician’s own mistakes); then, check the equipment (rule out malfunction or equipment performing out of calibration); and finally, check the patient (rule out patient error or pseudohypacusis).
References :
- Roeser, R. J., Valente, M., & Hosford-Dunn, H. (2007). Audiology: Diagnosis. Thieme.
- Katz, J., Medwetsky, L., Burkard, R. F., & Hood, L. J. (Eds.). (2007). Handbook of Clinical Audiology (6th revised North American edition). Philadelphia: Lippincott Williams and Wilkins.
- Gelfand, S. A. (2009). Essentials of Audiology. Thieme.
THANK YOU FOR READING
Comments
Post a Comment