What Patients Tell Us about Primary Healthcare Evaluation Instruments: Response Formats, Bad Questions and Missing...

Instruments have been developed that measure consumer evaluations of primary healthcare using different approaches, formats and questions to measure similar attributes. In 2004 we concurrently administered six validated instruments to adults and conducted discussion groups to explore how well the instruments allowed patients to express their healthcare experience and to get their feedback about questions and formats.

We held 13 discussion groups (n=110 participants): nine in metropolitan, rural and remote areas of Quebec; four in metropolitan and rural Nova Scotia. Participants noted critical incidents in their healthcare experience over the previous year, then responded to all six instruments under direct observation and finally participated in guided discussions for 30 to 40 minutes. The instruments were: the Primary Care Assessment Survey; the Primary Care Assessment Tool; the Components of Primary Care Index; the EUROPEP; the Interpersonal Processes of Care Survey; and part of the Veterans Affairs National Outpatient Customer Satisfaction Survey. Two team members analyzed discussion transcripts for content.

While respondents appreciated consistency in response options, they preferred options that vary to fit the question. Likert response scales functioned best; agreement scales were least appreciated. Questions that average experience over various providers or over many events diluted the capacity to detect critical negative or positive incidents. Respondents tried to answer all questions but stressed that they were not able to report accurately on elements outside their direct experience or in the provider's world. They liked short questions and instruments, except where these compromise clarity or result in crowded formatting. All the instruments were limited in their capacity to report on the interface with other levels of care.

Each instrument has strengths and weaknesses and could be marginally improved, but respondents accurately detected their intent and use. Their feedback offers insight for instrument development.