Harvard Pilgrim Health Care: Pilot Test of IVR Outreach Calls as a Mechanism for Collecting REL Data

Background

Harvard Pilgrim HealthCare (HPHC) has been using Interactive Voice Response (IVR) technology since 2003 to generate educational outreach calls to members who have not received necessary preventive or chronic care services within the recommended time period. Calls initially focused on flu reminders, but HPHC has since expanded outreach calls to include colorectal cancer screening, asthma and cardiovascular disease. HPHC generates more than 200,000 IVR calls in connection with clinical outreach projects. IVR technology is also used for other member communications.

IVR as a Vehicle for Collecting Race and Ethnicity Data

In June 2007, HPHC piloted the collection of race and ethnicity data in the context of IVR outreach calls to educate and encourage members to be screened for colorectal cancer. The calls began with questions about members’ screening history and plans, delivered information about screening options, and then asked about barriers to screening. After covering these items, IVR call recipients who remained on the line were then asked to describe their race and ethnicity.

Sample Script for Obtaining Race and Ethnicity Information Through Computer-Generated Outreach Calls:

  • Please tell me, yes or no, are you of Hispanic or Latino origin (such as Puerto Rican, Latin American, Mexican American or Cuban)? [IF NO]
  • Now I'm going to read from a list of other categories, including white, black or African American, Asian, Native Hawaiian or other Pacific Islander, American Indian or Alaska Native, or multiracial. Please say, "yes" after the option you feel best fits you or you can just say "please move on."
  • OK. Are you white? Black or African American? Asian? Native Hawaiian or other Pacific Islander? American Indian or Alaska Native? Multiracial? Another race that was not mentioned?

HPHC initially targeted 50,000 members for colorectal cancer screening outreach. After removing individuals who appeared in multiple call lists and accounting for individuals who could not be reached, HPHC was able to contact 22,000 individuals. Of the 20,000 members who accepted the CRC screening outreach call, 27 percent were still on the line when the query about race/ethnicity was made, representing 13 percent of the initial target population.

Results

Almost 96 percent of those who were queried about their race and ethnicity readily volunteered information for an overall yield of 11.5 percent of targeted members. Furthermore, no complaints were received from the members who were asked to provide the information. Although the percentage of members queried about their race/ethnicity could be increased by moving this query earlier in the call, the primary purpose of the call was to determine whether members had been screened and to convey important clinical messages to those who hadn't been screened (i.e., you should be screened for colorectal cancer; there are several acceptable tests; speak with your doctor about the right test for you; etc.). These messages were viewed as needing to be delivered first.

Despite the seemingly low response rate, there are several factors that still make the use of IVR for race/ethnicity data collection attractive. First, the cost of adding the race/ethnicity query to an existing IVR call is marginal since there is a one-time development cost and essentially no operational cost. There is also no cost associated with data entry, since the IVR responses are captured in electronic form. Lastly, it may be that when race/ethnicity queries are added to IVR calls that focus on less unpleasant topics or with more brief messages, a higher response rate is likely to result.

Lessons Learned

There are several benefits associated with the use of IVR technology.

  • IVR provides an opportunity to educate members as well as probe on their self-management behaviors.
  • The use of a toll-free number allows members to hear the information at a time that is more convenient for them.
  • Spoken messages may be more effective when dealing with individuals with low literacy, especially since members can ask to have statements and questions repeated as often as necessary.
  • Advances in IVR technology have also enabled calls to be conducted in Spanish.
  • Computer-generated messages may be perceived as less threatening than a personal discussion.
  • Previous IVR initiatives suggest that information reported by members through IVR is as reliable as that obtained through structured clinical interviews.

HPHC has learned several lessons in conducting the IVR colorectal screening pilot. The topic of colorectal screening is more unpleasant than many other issues and a larger number of members do not complete the entire call. The call is also longer than most as it asks a series of questions on CRC screening, plans for future screening and barriers to getting screened, explains all of the screening options, and provides information about the importance of screening. Regarding the sensitivity of indirect data collection methods, some members who self-identified as Hispanic or Latino were not correctly identified as such by geocoding and surname coding. Lastly, experience in reaching members who had been called the previous year and who had not been screened in the interim suggests that these individuals may require stronger messages and different questions to secure participation.

Footnotes:

1. Piette JD. “Interactive Voice Response Systems in the Diagnosis and Management of Chronic Disease.” The American Journal of Managed Care, 6(7): 817-827, 2000. 2. Piette JD. “Interactive Voice Response Systems in the Diagnosis and Management of Chronic Disease.” The American Journal of Managed Care, 6(7): 817-827, 2000.

Most Requested