The auditory mechanism is a complex sensory system and as such requires a wide selection of specific procedures for assessing its functional integrity. In general, these procedures may be grouped according to the differential information they supply about peripheral auditory disorders (i.e., external and middle ear, cochlea, and cranial nerve VIII), central auditory dysfunction (i.e., neural pathways of the brainstem and auditory cortex), and pseudohypacusis (i.e., hearing loss of nonorganic origin). The results of a battery of procedures contribute information about the auditory processes that are normal as well as those that are abnormal.

In differential audiologic assessment, the audiologist seeks procedures that provide optimum information about which levels of the auditory system are disordered. Patients can—and often do—have coexisting disorders at several levels, with the most dominant problem masking clues to the presence of others. Because no single test can represent the integrity of the entire auditory system, the best overall measure is obtained by combining test results, whereby each test within the test battery evaluates some aspect of the auditory mechanism. Jaeschke, Guyatt, and Sackett (1994) propose that useful diagnostic tests distinguish among disorders or states that might otherwise be confused, add information beyond that otherwise available, and lead to a change in management that is beneficial to the patient.

Some tests are designed specifically to assist in identifying the site of the lesion, while others are designed to determine the presence and nature of an auditory deficit. The diagnostic outcomes sought from the pediatric population vary little from adult counterparts. That is, audiologic tests are selected to differentiate peripheral versus central hearing loss, conductive versus sensorineural hearing loss, cochlear versus neural site of lesion, and varying middle ear conditions. The audiologist neither expects nor gets complete agreement on all the different tests performed. Age, physical and cognitive/intellectual conditions, individual variability, and peculiarities of different audiologic conditions can result in paradoxical outcomes and may affect the consistency of audiologic findings. Thus, an extensive battery of audiologic tests is intended to provide a profile of data that may be compared with findings obtained in individuals with previously documented auditory conditions. Evidence for a specific interpretation exists when the profile of results is consistent with expected findings.

The test battery approach in pediatric audiology is focused on confirming suspected hearing loss in infants referred from universal newborn hearing screening programs and the ongoing assessment of infants at risk for delayed-onset or progressive hearing loss (Joint Committee on Infant Hearing, 2000). Refinements in audiologic tests (e.g., conditioned behavioral tests including visual reinforcement audiometry and conditioned play audiometry; acoustic immittance; auditory-evoked potentials), as well as the addition of new audiologic tests (e.g., otoacoustic emissions), provide the audiologist with a sophisticated test battery from which to initiate clinical decisions (Folsom and Diefendorf, 1999).

When individual tests are combined into a test battery, results can be viewed from a holistic framework. In this approach, findings across tests are integrated to establish a working diagnosis that often goes beyond the sum of the individual parts. The use of a test battery offers several advantages, including (1) avoidance of overgeneralizing the results from a single test, (2) increasing the data set from which to draw conclusions, and (3) enhancing the confidence in a clinical decision as the number of test results consistent with a specific interpretation increases. Conversely, combining tests into a battery may not be advantageous, cost-effective, or time-efficient when the tests are highly correlated (that is, different tests testing for the same disorder). The more positive the test correlation, the less performance varies when tests are combined. When tests have high to maximum positive correlation, test battery performance cannot be better than the best single test in the battery; thus, there is no value in combining tests just to satisfy the faulty assumption that more tests are always better. Each test must be selected on the basis of the patient's complaints, and associated with the highest hit rate and lowest false alarm rate for the suspected disorder.

When selecting tests as part of a test battery, it is essential to balance quality patient care with fiscal responsibility. Therefore, the general rule to apply in pediatric assessment when selecting appropriate audiologic tests is not to administer a test unless its results provide new information for patient management. In fact, the real advantage of tests in a test battery comes when negative correlation is determined between tests, indicating that each test tends to identify different disorders.

Not only does the test battery delineate hearing loss, it also provides opportunities for making appropriate cross-checks. The cross-check principle in audiology, originally outlined by Jerger and Hayes (1976), undergirds the concept of a test battery approach so that a single test is not interpreted in isolation, but various tests act as a cross-check on the final outcome. The principle is that the results of a single test are never accepted as conclusive proof of the nature or site of auditory disorder without support from at least one additional independent test. That is, the error inherent in any test and in patient response behavior is recognized, and the probability of an incorrect diagnosis is minimized when the results of several tests lead to the same conclusion. Moreover, the test battery approach and cross-check principle provide a statistical advantage when compared with the utilization of but a single test. The multiplicity of judgments in the test battery renders the entire differential assessment more reliable and valid. Statistically, multiple judgments from nonduplicative, negatively correlated tests lend safety to the interpretation of raw data when compared to the outcome and potential for error from a single judgment. To implement the cross-check strategy successfully, clinicians need to recognize the importance of selecting tests based on the child's physical status, developmental level, and test correlation.

A test battery is paramount when the clinician is evaluating children with multiple disabilities. These children exhibit diverse medical problems that can diminish the accuracy of behavioral and physiological hearing tests. Complicating factors may include but are not limited to severe neurological, motor, and sensory problems. These factors can adversely influence test results, in turn compromising the validity of a single test approach. In addition, the limitations of the tests themselves can impose barriers when evaluating children with complex problems. Test constraints (e.g., limitations of behavioral observation audiometry [BOA] in eliciting an observable response; impact of developmental age on visual reinforcement audiometry [VRA]; middle ear pathology compromising acoustic reflex measures; the impact of central nervous system damage on the auditory brainstem response [ABR]) frequently dictate what procedures are feasible for a child with special needs. No test or clinician is infallible, and mistakes made with infants and young children can have crucial implications for medical and educational management.

Gans and Gans (1993) tested children with special needs by a test battery made up of BOA, VRA, ABR, and the acoustic reflex. The primary goal of the study was to rule out bilateral hearing loss greater than a mild degree. Stringent criteria were established for ruling out hearing loss with each of the tests to minimize the chances of missing a child (false negative error) with a moderate hearing loss or greater. The tests were performed in a serial manner until one test result ruled in essentially normal hearing. Once achieved, further testing was discontinued.

The results demonstrated that BOA passed approximately 35% of the children, VRA passed approximately 10%, acoustic reflex measurement passed approximately 22%, and ABR passed approximately 57%. Yet when conducted in a serial strategy (individual tests administered until one “normal” result was obtained), 80% of the children under study were determined to have hearing better than the cutoff criterion. Although ABR alone was better at predicting hearing sensitivity than the other tests, the total percentage score accomplished by a serial test battery was more than 20% greater than for ABR alone. Factor analysis failed to find a strong relationship among the tests and suggests that different factors caused changes in threshold estimation across different children. That is, successful outcomes were based on interactions between the individual's disabilities and the individual tests. These factors included but were not limited to low chronological or developmental age (neurological, motor, skeletal, and respiratory abnormalities), medications, and conductive hearing loss. Certain factors will adversely influence the results of one test more than another. Therefore, reliance on a single test for a child with disabling conditions would give an erroneous clinical impression that a large proportion of these children sustain hearing loss.

An audiologst prone to using a single-test approach might be tempted to rely on ABR as the only test method. This approach certainly relies on the assumption that the test of choice is a valid hearing test for all individuals and is not susceptible to variables that could lead to errors in outcome. The important findings from the work of Gans and Gans (1993) provide evidence that challenges this assumption for children with special needs, and provides a strong rationale for using a battery of appropriately selected tests.

The selection of individual tests for use in a test battery must be supported by clinical and experimental evidence. If individual tests and their use in test batteries are not evidence based, are not cost-effective in outcomes, and do not positively impact patients, we diminish the quality of services provided.