Methods

Face-to-face interviews are conducted one-on-one with seniors, caregivers, or key informants from the provider and professional communities. These interviews can be open-ended with general questions that address a particular concern or they can use a structured questionnaire or use a combination of both methods. For structured questionnaires, we generally use sets of questions that have been tested on similar populations so their reliability and validity are established. We also develop new sets of questions (scales) for new areas of knowledge. When we do this, we establish the reliability and validity of these questions through a pilot sample or focus groups.

Focus groups are held with homogeneous groups of 3-12 people to gather ideas to understand or solve problems. The advantages of this technique over face-to-face interviews are that we interview 3 to 12 times more people in the same amount of time and they often generate more ideas because of the synergy of group discussions that prompt ideas not sparked in one-on-one interviews. These groups are video or audiotaped and the tapes are transcribed and analyzed for recurring themes across focus group types (e.g., providers and clients; caregivers and people with disabilities).

Key informant interviews are held face-to-face, in focus groups, or over the phone with individuals who, by reason of their job or past experience, have a well informed opinion about a particular problem. Often, these informants are also stakeholders in the outcome of the discussion, such as agency directors, staff, clients, or family.

Mailed surveys are a common form of data collection and have the same concerns as face-to-face interviews, in terms of the reliability and validity of the questions. Since the questionnaires are self-administered and there is no opportunity for clarification, they need to be pilot-tested with a variety of people representing both genders, the most common racial and ethnic groups, age groups of interest, and lower and higher levels of education. There are various sources of mailing addresses, including the US Post Office, which can be matched with census tracts or other geographic units. Some mailed surveys are sent with prepaid postage to a service provider who is provided directions for selecting a random sample of clients and mails the survey to them. In all cases, the surveys are returned anonymously to us for data-entry and analysis. We generally subcontract with firms that accurately scan large quantities of surveys.

Record review of public or clinical records requires identification and reliable coding of relevant data and development of electronic data collection methods. Access to clinical records usually requires strict protection of client privacy within the Health Insurance Portability and Accountability Act (HIPAA) rules.

Secondary analysis of existing data is done when an entity wants to understand its existing clients or a particular problem, such as the pricing of services, by analyzing data already collected for another reason (e.g., for billing purposes). These data can be linked to other data (e.g., client demographics) to improve the analyses. There are also many public datasets collected for other purposes which can be used to understand problems associated with health, long-term care, end of life care, caregiving, and other issues that arise in later life. Although the analyses of these public data will not solve a particular agency’s problem, they may reveal valuable trends or patterns.

Telephone surveys are conducted usually through random digit dialing of phone numbers using telephone exchanges in the geographic areas of interest. They can also be conducted with known lists of individuals with phone numbers. We sub-contract the telephoning to Florida universities that have computer assisted telephone interviewing (CATI) technology and expertise. We develop the appropriate questionnaire, oversee sample selection, and analyze the results.