Fielding a patient satisfaction survey in a community health clinic

Objective – To pilot test the newly developed CAHPS-modified patient survey with an English-speaking population receiving services from a state-licensed community health clinic. This survey will be a component of NCQA recognition as a patient-centered medical home (PCMH).

Background – The PCMH item set was designed and added to core questions from the CAHPS Clinician & Group Survey to assess the additional domains of patient experience and care expected in a PCMH. The additional questions inquire about whole-person care, shared decision-making, and screening for additional needs such as nutrition education or help with substance use problem.

Methods – This retrospective study examines patient experiences for selected patients and parents of pediatric patients via service encounter information from our EMR system, dates ranging from July 1, 2010 through June 30, 2011. Inclusion criteria included at least two provider visits during this period and a reported language preference of English. The patient population was stratified by provider, and respective adult and pediatric samples were selected as those meeting majority provider algorithm rules. As many as 10 attempts were made to reach selected sample patients unless closed for inaccurate mailing address and inaccurate phone numbers.

Results - Our final patient sample was the following: 144 adult and 367 pediatric patients totaling 511 surveys. Adult response rates were 36.1% and the pediatric 22.2%. The active refusal rates were 13.2% and 10.8% respectively; unable to contact rates were 13.3% and 15.0%. Consistent with other studies on patient survey responses, patients rated satisfaction with identified provider higher than scores assigned for specific behaviors associated with patient-centered care. The highest domain scores were reported for provider communication and the lowest domain scores were reported for self-management support and shared decision-making. Bivariate analyzes were conducted; however, few significant associations were found. Two key associations were the following: patients 35 years of age and older reported significantly more support with self-management than younger patients, and persons with lower education levels reported receiving lower levels of comprehensive care.

Conclusions – Results were shared with the pilot research team (Harvard/AHRQ/RAND). There were many challenges conducting this pilot within a community health center environment that must be considered with future application of the survey. Despite these challenges, we had above average response rates for the English survey and have been selected to pilot-test the Spanish survey.