The calls were short, typically lasting no more than two to three minutes. About 40 percent of calls made reached their target participant; messages were left for another 45 percent of calls.
While most of the calls were constructed around simple questions requiring a 'yes' or a 'no' response, Dr. Vollmer said the calls had "a lot of branching logic" to them. They were, for instance, able to explore level of symptom control and barriers to compliance.
For someone late in refilling his or her prescription, Dr. Vollmer noted that the computer might ask, "Many people with asthma don't take their controller medicines when they feel well. Is that ever true for you?" Those answering yes would then hear a brief message about the difference between relievers and controllers and why the latter should be taken every day.
Alternatively, Dr. Vollmer noted, the computer might ask if the participant's asthma wakes him/her up at night or interferes with daily activities. For someone who said yes, the computer might respond, "With regular use of ICS, this doesn't need to be the case."
During the call, participants could choose to refill their prescription through Kaiser's automated service. They could also choose to speak to a real person.
Part of the success of the calls, Dr. Vollmer noted, was how "natural" the voice of the caller sounded. Although those called were told they were speaking with a computer, the voice was a human one that captured the intonation and rhythm of speakers from the geographic areas of the HMO members.
Dr. Vollmer and his colleagues hope to conduct a similar study on patients with cardiovascular disease, reminding
|Contact: Keely Savoie|
American Thoracic Society