Accuracy of screening instruments in identifying central auditory processing disorders: an integrative literature review

Corresponding address: Sheila Andreoli Balen Rua Desembargador Hemetério Fernandes, 1162, 102 Natal, RN, CEP: 59015-110, Brasil E-mail: sheila@sheilabalen.com.br ABSTRACT Purpose: to assess the literature about the accuracy of screening instruments for identifying the Central Auditory Processing Disorders (CAPD). Methods: search strategies were performed in the following databases: CINAHL, LILACS, PubMed /MEDLINE, Scopus, Speechbite and Web of Science. A search was also carried out in the grey literature. Four independent reviewers selected the included articles using a two-phase process based on the eligibility criteria. Two reviewers independently collected the required information from the included articles. The diagnostic methods were minimal batteries of behavioral tests to assess auditory processing skills. Results: from 1,366 articles found on all databases, after analysis of title and abstract, 36 were selected for the next phase, when 5 articles were finally included. It was found that the studies included were related to five instruments applied in children. The specificity was higher than 70%, but just the Mottier test and Screening Test for Auditory Processing (STAP), and Screening Checklist for Auditory Processing (SCAP) showed sensibility higher than 70%. Conclusion: Mottier was the most accurate CAPD screening test. There was no homogeneity in the presentation of the pass/fail criterion, or in the gold reference test used to establish the presence of CAPD.


INTRODUCTION
The Central Auditory Processing Disorder (CAPD) is defined by the difficulty in interpreting verbal and nonverbal auditory stimuli, resulting from the presence of auditory dysfunction of the central auditory nervous system (CANS) in its afferent pathways, with bottom-up modulation, and/or in the efferent pathways, which receive top-down modulation and involve language, speech, cognition, attention, memory and fluid reasoning¹ , ². The International Statistical Classification of Diseases and Related Health Problems (ICD-10) characterizes the Auditory Processing Disorder as an Impairment of Auditory Discrimination, under the classification "Other abnormal auditory perceptions" (H93.25) 3 .
Currently, there is an understanding that there is a correlation between CAPD and other developmental disorders in children, such as Autism Spectrum Disorder, Dyslexia and specific language impairment 1 . Thus, there are gaps in the literature regarding evidence of purely auditory deficit in children diagnosed with CAPD 2 , since this alteration can manifest itself as a co-occurrence of other disorders. These changes in the Central Auditory System can cause several damages to the development and learning process of children 4 , impacting the performance of personal activities and quality of life 1 .
In the literature, it is possible to categorize auditory processing screening instruments into various types, such as tests with recorded stimuli 5 , questionnaires and checklists 6 , scales 7 , software 8 and online program 9 and tests with uncalibrated sounds 10 .
In the study by Volpatto et al. 11 , it is possible to see, through a systematic review, the questionnaires, and checklists most used in Brazil for auditory processing screening available in Portuguese. Among the translated and/or adapted tools, the Auditory Processing Domains Questionnaire (APDQ) is the only questionnaire that has been validated 11 . Another review that focused on analyzed studies that applied CAPD screening in schoolchildren did not use the eligibility criterion studies of comparing the CAPD screening with diagnosis procedures in CAPD including behavioral assessment battery tests 12 . This eligibility is necessary for measuring the accuracy of the CAPD screenings protocols. So, this gap in literature remains.
The broad spectrum of alterations inherent to the diagnosis and clinical implications of the CAPD makes it very important that the screening instruments have diagnostic validity for them to be apt to quickly and effectively detect this disorder. The more valid the screening of CAPD is, greater are the chances of the subject to be properly identified as with a high or low probability to have CAPD and be effectively referred for diagnostic evaluation when necessary.
Therefore, it is of great importance to screen, as early as possible, the central auditory processing disorders. It is important that the process of improving, or even developing methods of auditory screening be approached with caution 13 , as it is the moment when the professionals collect information on the subjects' health, education and well-being, which are important parameters for them to perform their daily tasks 14 .
Based on the identification, the child can be adequately diagnosed and, afterwards, referred for intervention 13 .
In this way, this review aimed at studying what is referred to in the literature regarding the accuracy of screening instruments in identifying the CAPD.

METHODS
This integrative literature review was guided by the question: "What is the accuracy of the screening instruments in identifying the CAPD?". After the question had been developed in the first phase, the second phase of the integrative review was begun, involving the survey of the literature in six databases, namely, Cumulative Index to Nursing and Allied Health Literature (CINAHL), Literature in the Health Sciences in Latin America and the Caribbean (LILACS), PubMed (including MEDLINE), Scopus, SpeechBITE and Web of Science, in addition to two grey literature databases, Google Scholar and OpenGrey.
Mesh terms, keywords, and other free terms related to "questionnaires"; "hearing tests"; "mobile application"; software; "mass screening hearing"; "auditory perception"; and "auditory perceptual disorders" were used with Boolean operators (e.g. OR and AND) to combine searches. Additional information on the search strategies are provided on the Appendix 1. The lists of references of included studies were also hand-searched to identify additional relevant studies 15 . Experts were also consulted by email to improve search findings.
Studies of the accuracy screening instruments applied for identification of the CAPD were included. For this end, it was necessary that the study had compared CAPD screening instruments either already validated or with a battery of tests for the diagnosis of CAPD. The inclusion of studies had no restrictions regarding language, age, gender, and time of publication.
Studies excluded from the analysis were: 1 -Studies whose subjects had hearing loss; 2 -Studies presenting any other screening protocol, whose focus was not the CAPD; 3 -Studies not using the reference standard, which could be either a battery of behavioral tests for the auditory processing assessment, or instruments validated for the diagnosis; 4 -Studies not presenting accurate measurements (sensitivity and specificity), or not presenting sufficient data to calculate them; 5 -Case-control studies, cohort studies, clinical trials, conference summaries, personal opinions, congress annals, reviews, editorials, and letters to the editor; and 6 -Unavailable complete articles. Those criteria were applied on screening stage and full text analyses.  16 was used to enable, between the reviewers, the blind reading of the titles and abstracts.
In the third phase of the review, four reviewers selected the articles blindly and independently. On the screening stage, these four reviewers examined the titles and abstracts of all the studies through the Rayyan.qcri (Rayyan, Qatar Computing Research Institute) 16 applying eligibility criteria. Then the same four reviewers evaluated the full text with eligibility criteria. The four reviewers cleared any disagreement in both stages by discussing and getting to a mutual agreement. When they could not get to a consensus, the fifth reviewer was asked to make a final decision.
Two reviewers blindly and independently collected from the selected articles author, year and country of publication of the study, sample size, mean age in years, characteristics of the subjects and other alterations, screening instruments, auditory processing skills, sensitivity and specificity values, pass/fail criterion, and the diagnostic method used. Any disagreement was settled by discussion and mutual agreement between the two reviewers. The sixth author was involved, as necessary, to make possible for the final decision to be formulated.

LITERATURE REVIEW
The search on the databases returned 1,366 studies, as shown on the diagram (Figure 1). After the duplicated files had been removed, the first selection of 1,174 articles was conducted (stage 1), through the reading of title and/or abstract. Following the eligibility criteria, 1,144 studies were excluded. Of the 30 remaining articles, three could not be obtained, even though the authors of the articles had been contacted by e-mail in different days and hours to make possible for the articles to be obtained in full. Neither were the articles available in full text on the repositories to which the researchers' university of origin has access, nor was there availability for them to be acquired. The search on the grey literature identified 1,864 studies; the first 100 results from each database were taken into consideration in the initial title and abstract analysis. None of these met the eligibility criteria. It is also important to highlight that, after the reference lists had been sought through and the articles had been solicited to the experts, no additional study was included.
With the update of the searches, 114 were analyzed based on the title and abstract. Three articles met the eligibility criteria and went on to analyze the full text. Therefore, 33 articles passed on to the stage 2, when the full texts were read, and 28 of them were excluded following the same eligibility criteria (shown in Appendix 2). Lastly, five studies remained and were included in the qualitative analysis.
The five studies included had been published in the United States 17 , in Germany 18 , in the United Kingdom 19,20 and in Hong Kong 21 . The sample size of the studies varied from 81 to 201 subjects, aged from 6 to 16 years.
The screening instruments used were a Screening Test for Auditory Processing Disorders (SCAN) 17 ; auditory processing behavioral test battery 18 , Children's auditory performance scale (CHAPS) 19 , the SCAN-3 subtest Auditory Figure-Ground (AFG) in the signal/ noise relations of +8 dB and 0 dB 20 , the Screening Test for Auditory Processing (STAP) and the Screening Checklist for Auditory Processing (SCAP) 21 . The pass/ fail criterion was presented in four articles, and they were the signal-noise relation of 0 dB in the AFG 20 ; failing in two or more tests of the battery conducted, with score of 1 to 1.5 SD below average 19 ; the parameters recommended by Keith 5 , i.e., 1 SD below average composite score of the subtests 18 ; and a cut-off score criteria of 6 on the SCAP. In the articles from 2000 18 and 2014 21 , there was the need of support from another study to present this information 5,22 . It should be noted and 100%. The CAPD diagnostic methods observed were minimum batteries of behavioral tests aimed at assessing all the auditory processing skills, or the SCAN test battery aimed at diagnosing. The detailed characteristics of the studies included are shown in Table 1.
that article of 2014 21 did not mention the parameters of the STAP.
All the studies achieved the primary outcome, proving the accuracy of the instruments analyzed. The values of sensitivity ranged from 42.1% to 97.2%, whereas those of specificity reached between 65.5%  The five studies included in this review used different procedures for screening auditory processing and were conducted with children. There was no homogeneity in the presentation of the pass/fail criterion, or in the gold reference test used to establish the presence of the CAPD.
All the articles included presented the sensitivity and specificity measurements of the screening instruments [17][18][19][20][21] . The instruments used were two versions of the SCAN 17,20 , Mottier test 18 , STAP and SCAP used separately and in combination 21 , and the questionnaire CHAPS 19 , in spite of the questionnaires, especially those developed more recently based on psychometric characteristics, being promising screening instruments to detect subjects at risk of CAPD 6 . A systematic review with questionnaires for Brazilian Portuguese found only one validated instrument, the APDQ, indicating that there was a rate of 100% in the sensitivity and specificity measurements 11 . This datum indicates that it is a tool apt to detect all the subjects who actually have CAPD (sensitivity), as well as exclude the subjects without alteration (specificity) 23 .
Despite this, the values of sensitivity and specificity of the SCAP may reflect that the use of questionnaires by other professionals, such as teachers, tends to correctly identify CAPD in suspected children. However, they are likely to be over-referred. Screening becomes more accurate when there is an association between two types of tools, such as the use of tests that involve recorded sound stimuli 21 .
Sensitivity and specificity vary across studies. Nevertheless, some studies have used similar instruments, they could belong to different categories, such as the SCAN 16,19 which has variations. Probably the different characteristics from tests can interfere on the task executed by the child, and consequently on sensitivity and specificity measures. CAPD complex nature 1 could reflect different results for different symptoms, too.
Five instruments applied in children for CAPD screening were identified: CHAPS, AFG, SCAN, Mottier, SCAP and STAP. The specificity of all the instruments CHAPS, AFG, SCAN, Mottier test, STAP and the associated use between SCAP and STAP were higher than 70%, but just the Mottier and STAP test, and SCAP showed sensibility higher than 70%.
The Mottier test was the instrument with better specificity and sensibility. This test assesses memory, which is only worked on in the STAP tests. This gives greater sensitivity to the instrument since this is one of the most affected skills in this population 21 . In addition, memory is linked to sequences of meaningless syllables which makes the test more difficult.
Among the articles included in this review, it is possible to verify that there is no strong relationship between questionnaires and CAPD screening tests. Questionnaires can provide information on comorbidities and behavioral issues that allow them to know the individual in the educational, social and communicative contexts. Meanwhile, the performance-based screening tools present varied tasks that focus on different listening skills and some of them may have more consistent results to determine the CAPD or the need for a detailed behavioral assessment 24 .
The description of the methodological procedures and the pass/fail criterion used in the studies was not homogeneous either. Kiese-Himmel; Nickisch 18 did not present the pass/fail criterion used, making it difficult for the instrument to be concisely analyzed, and the methodological procedures to be reproduced by an independent author. The remaining four articles coherently described the criterion used 17,19-21 . However, they diverged from one another, as the tests present their particularities, related to the use of stimuli with different acoustic characteristics and tasks. Despite the SCAN being used in two studies 17,20 , the manner of assessment and the parameters of analysis were not the same. Domitz; Schow 17 propose the screening to be conducted with all the subtests, based on the recommendations of the author of the instrument 5 . Ahmmed 20 , in his turn, deals only with the use of the Speech-in-Noise subtest, studying the performance of the subjects in the signal/noise relations of 0 dB and +8 dB to define which of these would better reflect the actual listening difficulties.
The age range of the participants of the included studies encompass school age, the youngest being six years old 19,20 , and the oldest, 16 years old 19 . Such proximity reiterates the relevance of studies with this public, since the screening can provide important information for the diagnosis. In the adult public, there is greater probability that the signs and symptoms of CAPD have not been identified or could be confounded with other health conditions. However, the results of this review evidenced the poor studies with this public, that need an adequate instrument to screening too.
Regarding the reference standard to determine the presence or absence of CAPD, there was no homogeneity. This is certainly a reflection of the complex and heterogeneous nature of the CAPD 1,25 and, as a consequence, of its diagnosis, whose definition presents diverging concepts.
The recommendations concerning the diagnosis indicate that the different categories composing the auditory processing -temporal processing, dichotic listening, low-redundancy speech perception, and binaural interaction -should be evaluated 26 . The German study that investigated the accuracy of different combinations of the tests concludes that, for the diagnosis, the ideal is the combination of the ten tests of the battery studied 18 . Following the same line of investigation, one study from the United Kingdom uses as gold standard the SCAN-3 20 , whereas the other study from the United Kingdom uses the SCAN-C battery as diagnostic criterion 19 . Although these two studies use SCAN, different versions are used, with distinct diagnostic criteria, even though the tests are similar. Another study used the MAPA battery 17 , and another, a combination of tests 21 , as recommended for the diagnostic assessment 26 .
A lot of points are discussed about the definition and diagnosis of CAPD 2,27,28 . It cannot yet be precisely stated that the CAPD is characterized as a solely auditory deficit 2 , since the literature evidences that the intelligence, memory, attention and language characteristics in children with CAPD overlap the same characteristics in children diagnosed with specific language impairment, dyslexia, attention deficit hyperactivity disorder (ADHD) and learning difficulty. This supports the idea that the various diagnosis of disorders interrelates, forming a sort of spectrum, instead of distinct alterations, with well-defined characteristics 29 .
There is also the premise that the CAPD goes beyond a disorder, its symptoms being a marker for the Neurodevelopmental Syndrome, which associates various auditory, speech, attention, memory, and behavioral difficulty markers. Depending on the severity and predominance of one or more markers, the child would manifest a unique development profile, with modulated unfoldment through the genetic composition, the environment to which the child is exposed, the age, and the academic demands 30 .
For Moore 27 , it is implausible that children with auditory difficulties (auditory inabilities, in spite of audiometry within standard normality) present a disorder in the central auditory function and it can be diagnosed only through a combination of tests. This occurs because the symptoms clinically reported in the subjects do not follow logically along with the tests used, or with an intervention based on evidence 27 . On the other hand, a second line of thought states that the term "auditory difficulties" is not specific enough, just as it does not contribute to the differential diagnosis. Due to its complexity, CAPD needs a thorough assessment through the central auditory processing tests and the multidisciplinary insight for the differential diagnosis to be precise and to manage to identify comorbidities, thus maximizing an effective intervention, directed to the specific processes identified through the diagnostic battery [26][27][28] .
In spite of this, all the studies included at least one test that assesses the abilities of hearing integration and/or target-to-masker listening 17,[19][20][21] . Filtered speech was used 17,19 , as well as speech in noise 21 , and figureground 20 , encompassing the degraded speech 1,25,26 .
Given the particularities of each instrument, it is challenging to think which pass-failure criterion would be the most relevant. For STAP 21 , whose criteria are based on existing diagnostic tests, there is greater reliability in the result when the total score has correct answers above 50% of the stimuli presented in each subtest. The SCAN 17,20 , which had lower sensitivity values, presented parameters involving um standard deviation below the average of the composite score 17 and the percentile score ≤ 10 in the AFG subtest, both determined by reflecting more adequately to the real hearing difficulties. This comparison is also reinforced with the SCAP 21 and CHAPS 19 questionnaires. In the first, the cut-off point the score was six, with a score for each symptom related to CAPD 21 , while the second is based on a score below 1.5 standard deviation 19 .
Attention should also be called to the fact that, of the 28 excluded articles, seven were so because of the methodological design that diverged from the accuracy study, which was the focus of this review. Ten articles were excluded for not presenting all the accuracy measurements; and eight did not report the comparison with a battery of auditory processing diagnostic tests, considered as the reference standard assessment. This reflects the shortage of studies in audiology with methodological rigor to assess the accuracy of the CAPD screening instruments, and the need of primary studies presenting this scope and overcoming the difficulties with the diagnostic and the establishment of the reference standard. Only after this it will be possible to broaden the discussions and present more effective evidence, strengthening the importance of CAPD screening in the clinical context.
The heterogeneity found in the gold reference standard test and in the index test are presented here as limitations to this study, since it makes the comparison and the conclusion less effective. Nevertheless, it reflects the complexity of the CAPD, and points to the need for further primary studies with rigorous methodology in order for them to be replicated, thus making comparisons easier. A methodological limitation of this review was the lack of access to three studies on the second phase, as they were unavailable in full text version, though all the possibilities had been tried, even requesting the corresponding author to provide the article.
Another limitation to the study was exactly the non-homogeneity of the pass/fail criteria to consider the CAPD in the screening instruments and in the diagnostic criteria. These two points wind up making evident the fragility in this field of knowledge; it is thus necessary to invest in primary studies presenting sturdier scientific methodology so as to bring about clearer and more valid scientific contributions to the clinical context of the practice of audiology in the field of the CAPD.

CONCLUSION
The CAPD screening test with the best accuracy was the Mottier test. There was no homogeneity in the presentation of the pass/fail criterion, or in the gold reference test used to establish the presence of the CAPD. The specificity of the instruments identified (CHAPS, AFG, SCAN, Mottier test, STAP and the associated use between SCAP and STAP) were higher than 70%, but just the Mottier and STAP test, and SCAP showed sensibility higher than 70%.

ACKNOWLEDGEMENT
Special thanks to the colleagues of the Laboratory of Technological Innovation in Health, Onofre Lopes University Hospital, Federal University of Rio Grande do Norte, for all the support and useful contributions throughout the process. We also thank Claudio Lingerfelt for the great work in translating the manuscript and Wesam Ashour for reviewing the translated manuscript. Scopus ("questionnaires" OR "questionnaire" OR "surveys" OR "questionnaire design" OR "hearing tests" OR "hearing test" OR "mobile app based interventions" OR "mobile application" OR "Mass Screening" OR "Mass Screenings" OR "Screening" OR "Screenings") AND ("Auditory Perceptual Disorders" OR "Auditory Perceptual Disorder" OR "Auditory Processing Disorder" OR "Auditory Processing Disorders" OR "Psychoacoustical Disorders" OR "Acoustic Perceptual Disorder" OR "Acoustic Perceptual Disorders") AND ("Hearing" OR "audition" OR "Auditory Perception" OR "Auditory Perception" OR "Auditory Perceptions")