Developing Actionable Survey Questions to Improve Patient Experience

Abstract

Patient experience surveys should be concise, with questions that give us the most insight into our patients’ needs and how successfully they were addressed during visits. Survey questions are most valuable if answers are actionable, connect more clearly with skill sets and identify issues which improve care.

Patient experience surveys can provide important insights into what clinicians do well, what patients’ perspectives are, and where there are opportunities to improve. These data are most valuable when it inspires initiatives which help provide a better experience for patients, caregivers and care team; strengthen patient loyalty; and improve health outcomes. For example, the quality of clinician–patient communication has emerged as a key driver of a great patient experience (1) and has led to many impactful communication skills development programs such as those modeled on the relationship-centered communication curriculum of the Academy of Communication in Healthcare (2). Patient experience surveys are only helpful, however, if patients fill them out thoughtfully. As in other industries, this is most likely if surveys are time efficient and give patients a sense that their responses will result in action to improve care (3). We suggest that the current lengthy, general surveys be revisited and questions redesigned for the most meaningful impact.

What if patient experience survey feedback could become a highly efficient and specific method of pinpointing the key elements clinicians perform well, and rapidly identify the pain points to spark corrective action? Can surveys be designed with dynamic questions which better capture the patient perspective and provide more actionable feedback?

To illustrate this idea, consider the following question from the provider section of the Consumer Assessment of Healthcare Providers and Systems (CAHPS) Clinician and Group Survey: (4) During your most recent visit, did this provider show respect for what you had to say?

The patient is asked to score the provider on a scale of: 1 = yes definitely, 2 = yes somewhat, 3 = no.

Responses to this question may provide a broad sense of whether the provider was successful at showing respect, but it does not specify behaviors to be emulated or improved upon. Now, consider this alternative:

Did you have an opportunity to mention all of the concerns you hoped to discuss at your most recent visit?

This question is also about respect and specifies a desired behavior—eliciting a complete list of concerns. The answer to this question provides a clinician with an actionable patient-centered response which can be applied in practice. Moreover, it anchors both the clinician and patient to the specific visit that both can attribute the response to. Patients may be reluctant to disclose certain underlying worries which, if not drawn out and addressed, may leave them feeling incompletely cared for. These insights, brought to scale, would provide health systems with thematic content and methods to help lift up those in need of support and highlight top performers.

Current surveys may also present a series of statements (5) which asks patients to rate their provider on a scale of very poor-poor-fair-good-very good. One such statement is:

care provider’s efforts to include you in decisions about your treatment.

A more actionable statement, graded on the never-to-always continuum, would be:

I felt like my thoughts and opinions regarding my care and treatment were welcomed and acknowledged.

This question is less about the provider’s efforts, but rather how successful they were in welcoming the patient’s input and perspective.

Patient responses collected through survey data are only helpful if they are a true representation of the patient experience, and patients are willing to complete them. It follows that surveys need to be well designed, so they capture a diverse and representative cohort of respondents. Survey questions alone, though, may not give the most complete picture. Today’s technologic landscape allows for a rapid and dynamic flexibility for engaging and interacting with patients through multiple remote modalities. It may better serve clinicians and patients to move away from a one-size-fits-all model and seek feedback formats which allow patients to share their experiences in a variety of ways (eg, narrative form, both written and verbal, to complement prewritten questions.) Natural language processing may aid in extracting the most insight possible from patients’ written and verbal responses. Research and quality improvement efforts can focus on developing methods which adapt to patients, tailor to their needs, and engage them in the most constructive ways.

It is possible, as illustrated above, to design concise surveys with questions that provide the most insight into patients’ needs and how successfully they were addressed during visits. Further, questions are most valuable if they lead to specific actions to improve care. This way, survey data can connect more clearly with skill sets, and better inform initiatives to improve them (6). Clinicians may become more engaged if survey data lead to actionable and constructive opportunities to improve skills which foster connection with patients and greater joy in practice. It can be an all-around win.

Author Biographies

Jeffrey Millstein is a primary-care physician and medical director for patient experience-regional practices at Penn Medicine.

Anish Agarwal is an assistant professor in the department of emergency medicine at the University of Pennsylvania’s Perelman School of Medicine.