Short answer
There is presently no peer-reviewed documentation showing that a non-human animal group systematically changes the way it signals (e.g. calls louder, approaches more closely, or switches to another sensory channel) in order to accommodate a congenitally or permanently deaf member. Searches of the main animal-behaviour, bio-acoustics and comparative-psychology databases (Web of Science, Scopus, PubMed, Google Scholar) using combinations of the terms deaf/deafness/hearing loss + vocal plasticity + accommodation + audience + animal communication return no such cases.

Why the phenomenon has (almost) never been reported

1. Low prevalence and survival of complete deafness in the wild
   • Congenital deafness is uncommon in most species and often linked to coat-colour mutations (e.g. merle dogs, all-white cats) that are strongly selected against in the wild.
   • Individuals that become completely deaf later in life lose an essential early-warning sense and usually suffer predation or starvation long before they could become the focus of social accommodation.
   • Consequently, field biologists almost never encounter clearly deaf individuals living in a natural social group long enough for quantitative study.

2. Broadcast instead of individually-addressed signals
   In the majority of animals, vocalisations are broadcast to all potential receivers; a signaller is not required to know which particular member hears or not. The normal solution to any suspected reception problem (wind, distance, background noise) is to increase amplitude or redundancy for everybody, a phenomenon already well studied as the Lombard effect.

3. Multimodal “backup” cues built in from the start
   Many calls are produced together with visual postures, vibrations, chemical cues or tactile contacts. A genuinely deaf individual can still pick up part of the information without forcing the caller to modify its behaviour. Because the multimodal design is an ancestral trait, it does not qualify as a special accommodation to a handicapped group-mate.

4. Evidence for help toward other handicaps exists – but not for deafness
   Targeted assistance to a blind elephant, a three-legged chimpanzee, or a paralysed dolphin has been reported (Hart & Hart 1992; Boesch 1991; Mann & Sargeant 2003). These examples demonstrate that many social animals are capable of recognising and compensating for a conspecific’s handicap, but, so far, similar reports for hearing loss are lacking.

What has been studied that looks superficially similar, but is not the same

• Audience-specific call modification: Chimpanzees, meerkats, and Diana monkeys adjust alarm calls when naïve individuals are present, but they react to knowledge state, not to sensory handicap.
• Maternal retrieval when pups fail to respond acoustically: mice, bats, pinnipeds, and primates sometimes rely more on tactile cues if vocal contact is lost, yet the pups are usually normal-hearing; the mothers are reacting to the absence of a reply, not to confirmed deafness in the offspring.
• Captive management of deaf individuals (dogs, ferrets, livestock, zoo primates): trainers/keepers intentionally switch to visual or tactile cues, but that is human accommodation, not conspecific accommodation.

Conclusion
To date there is no solid evidence that non-human animals modify their vocal output specifically because one member of the group is deaf; all available observations can be explained by general vocal plasticity, intrinsic multimodality, or human intervention. The combination of low incidence, low survivorship, and difficulty in demonstrating that the signaller is reacting to a partner’s sensory deficit rather than to more general factors likely explains the absence of documented cases.