(PP1315) Lipreading Reliability and Validity

The current state of the technology, existing evidence for it's diagnostic value, and relevant ecological validity make it time to introduce a standard test of audiovisual speech perception into the typical audiometric assessment battery.  We’ve been able to measure the benefits, beyond listening alone, associated with speechreading for nearly 50 years, but it has been relatively impractical to test the visual speech perception element in the typical clinic until recently.  As an initial step in developing this effort, the current investigation will analyze and present reliability and validity results for lipreading data collected over the course of multiple test sessions.

Summary: 

Purpose:   Despite the ecological validity of conducting audiovisual speech perception assessments with clinical patients, the ability to perceive visual or audiovisual speech is almost never assessed in the clinical setting.  Clinicians counsel patient’s ability to hear and perceive speech while leaving a valuable piece of the patient’s communication profile out of the conversation.  If counseled on the benefits of watching a talker in difficult listening situations, the clinician can only give lip service (pun intended) to a patient’s ability to benefit unless they have assessed the patient’s ability to lipread and/or speechread and ultimately compared it to normative values.  Affordable flatscreen technology and multimonitor video cards allow for the easy insertion of an assessment screen inside the calibrated testing environment.  Further, at least two developments in the study of lipreading and speechreading support the inclusion of this valuable tool in the assessment of a patient’s communication ability.  First, lipreading ability has shown to be predictive of performance in difficult audiovisual listening conditions.   Second, we have established that there is a great deal of variability among listeners in the ability to lipread, and furthermore, among non-congenitally deafened adults, there is no relationship between the degree of hearing loss and the ability to lipread (i.e., if the clinician doesn’t measure it, they can’t assume they know anything about a patient’s lipreading ability.  In this study, we will take the required initial steps in establishing lipreading ability as clinical tool by determining if it can be a valid and reliable assessment tool. 

Method:   XX adults (XX older and XX younger) participated in a larger study of audiovisual speech perception that required three separate visits to the lab.  For the current discussion we only consider the lipreading (visual-only) testing conducted for that study.  Each day, lipreading was assessed using two different approaches.  The first included 24 sentences from the Build-a-Sentence test (BAS).  The BAS is a closed-set sentence matrix test that uses 36 possible words to make a three key-word sentence (e.g., The whale watched the dog and the girl, or The cow and the boys watched the mouse).  The second test was an open set test of XX different words taken from the LexList audiovisual test, with XX presented on each day. Daily lists were counterbalanced across participants. Data were analyzed for validity (was one test predictive of the other) and reliability (did each test produce the same results on each day). 

Results: Lipreading assessment using either the BAS or the LexList tests are reliable and appear to be valid.   Reliability coefficients as high as .9 were found, suggesting excellent reliability.  Construct validity coefficients as high as XX were found, suggesting good validity for lipreading across open-set and closed set test formats.

Conclusions:   The findings provide the essential initial steps required for going forward with the development of a standard clinical test of lipreading and speechreading.  Without establishing that lipreading tests are valid and reliable, a clinical assessment of lipreading and speechreading would not be useful.