Preparing Sample Files Before Collecting CAHPS Survey Data

Before a CAHPS survey goes into the field, survey sponsors should confirm that their vendors have taken the following steps to prepare the sample files.

Remove duplicate names: While the sample frame can include multiple individuals from the same household, the final sample for the survey should not include more than one person (adult or child) per household.

Assign a unique identification (ID) number: Once the sample has been selected, the vendor assigns a unique identification (ID) number to each sampled person to track survey responses during data collection. To ensure confidentiality, this unique ID number should not be based on an existing identifier such as a Social Security number, a patient ID number, or an employee ID number. 

Confirm contact information: The information that is most critical to the success of data collection is:

• Accurate and complete names for patients.
• Contact information appropriate for the mode of administration (i.e., addresses for mail surveys, phone numbers for phone administration, and email addresses for online administration).

If sample frames do not include complete and accurate contact information, it may be necessary to combine information from two (or more) sources, such as health plan or insurance administrative records and contact information from the medical group or provider office. When information from two sources differs, survey sponsors and their vendors should consult with each other to decide which sources of information are most accurate and should be used. This can be a complex, multistep process that requires time and rigorous quality control. In addition, because the survey sponsor may be responsible for some elements of this process and the vendor for others, it is important to carefully coordinate this process.

Maintain confidentiality: Privacy assurances are central to encouraging respondent participation. Survey vendors should have standard procedures in place for maintaining the confidentiality of respondents' names and minimizing the extent to which identifying information, such as names and addresses, are linked to the actual survey responses. Survey vendors should train employees about confidentiality and require that they sign statements ensuring that they will not reveal the names of respondents, or any results linked to specific individuals.

There are several opportunities during the survey process to explain to respondents that their responses are kept confidential. These include the advance and cover letters and interviewer assurances during phone interviews.