Wireless privacy and security are complex technical issues due to the needs for highly distributed data storages and access controls. The emerging Wireless Body Area Network (WBAN) enabling W-Health exacerbated privacy and security concerns:
http://wiser.arizona.edu/papers/WCM2009.pdf
Data and users Authentication, Authorization, Administration, and Auditing (AAAA) are vitally important but getting insufficient in wireless communications. Cryptographic techniques including encryption / decryption schemes and algorithms, encryption key management, and emerging new methods like attribute based encryption are necessary technologies to complement AAAA.
Authentication is an easy way to address privacy. For instance, ZigBee network has a PAN ID to support authentication. It automatically encrypts the transmission.
In Wireless networks, access to the raw data packets cannot be prevented. The only way I see in the end is encryption, either negotiated explicitly during communication setup, triggered implicitly during authentication or implemented in the private data itself (transport encrypted private data over unencrypted wireless channel).
Philippe De Ridder is right on point. There is no way to prevent the interception of data packets. Wireless is basically just a radio wave broadcast so short of mitigating the strength or somehow containing the radio waves you cannot prevent an outsider from capturing the packets. I have captured packets from 1/2 mile away on some wireless networks then it is just a matter of decrypting. Either block the radio waves or make sure you are encrypting using a strong algorithm and pass-phrase combination.
The traditional way is encryption, seems like no other way if you want efficiency and you have a single broadcast channel. Of course steganography is an alternative but you mention it in the question as a non option (Chaffing and winnowing is not an option-- and this method is steganographic). Of course steganography is just another keyed based method where you code bits in variations of believable non-ciphertext content.
These are already in the question and answers.
Finally, something new: if multi frequency is available, the previous answers did not mention Frequency-hopping spread spectrum (another way to exploit crypto key = pseudo randomness but without explicit encryption). See https://en.wikipedia.org/wiki/Frequency-hopping_spread_spectrum