From Dell Medical School, The University of Texas at Austin, Austin, TX.
- Objective: Satisfaction measures often show substantial ceiling effects. This randomized controlled trial tested the null hypothesis that there is no difference in mean overall satisfaction, ceiling and floor effect, and data distribution between 4 different kinds of single-question scales assessing the helpfulness of a visit. We also hypothesized that there is no correlation between scaled satisfaction and psychological status. Finally, we assessed how the satisfaction scores compared with the Net Promoter Scores (NPS).
- Design: Randomized controlled trial.
- Methods: We enrolled 258 adult, English-speaking new and returning patients. Patients were randomly assigned to 1 of 4 different scale types: (1) an 11-point ordinal scale with 5 anchor points; (2) a 5-point Likert scale; (3) a 0-100 visual analogue scale (VAS) electronic slider with 3 anchor points and visible numbers; and (4) a 0-100 VAS with 3 anchor points and no visible numbers. Additionally, patients completed the 2-item Pain Self-Efficacy Questionnaire (PSEQ-2), 5-item Short Health Anxiety Inventory scale (SHAI-5), and Patient-Reported Outcomes Measurement Information System (PROMIS) Depression. We assessed mean and median score, floor and ceiling effect, and skewness and kurtosis for each scale. Spearman correlation tests were used to test correlations between satisfaction and psychological status.
- Results: The nonnumerical 0-100 VAS with 3 anchor points and the 5-point Likert scale had the least ceiling effect (12% and 20%, respectively). The 11-point ordinal scale had skewness and kurtosis closest to a normal distribution (skew = –0.58 and kurtosis = 4.0). Scaled satisfaction scores had a small but significant correlation with PSEQ-2 (r = 0.17; P = 0.006), but not with SHAI-5 (r = –0.12; P = 0.052) or PROMIS Depression (r = –0.12; P = 0.064). NPS were 35, 16, 67, and 20 for the scales, respectively.
- Conclusion: Single-question measures of satisfaction can be adjusted to limit the ceiling effect. Additional research in this area is warranted.
Keywords: patient satisfaction; floor and ceiling effect; skewness and kurtosis; quality improvement.
Patient satisfaction is an important quality metric that is increasingly being measured, reported, and incentivized. A qualitative study identified 7 themes influencing satisfaction among people visiting an orthopedic surgeon’s office: trust, relatedness, expectations, wait time, visit duration, communication, and empathy.1 However, another study found that satisfaction and perceived empathy are not associated with wait time or visit duration, but rather with the quality of the visit.2 Satisfaction measures that incorporate many of these features in relatively long questionnaires are associated with lower response rates3 and overlap with the factors whose influence on satisfaction one would like to study (eg, perceived empathy or communication effectiveness).4 Single- and multiple-question satisfaction scores are prone to a strong right skew, with a substantial ceiling effect.5 Ceiling effect occurs when a considerable proportion (about half) of participants select 1 of the top 2 scores (or the maximum score). An ideal scale would measure satisfaction independent from other factors, would use 1 or just a few questions, and would have little or no ceiling effect.
In this randomized controlled trial, we examined whether there were significant differences in mean and median satisfaction, floor and ceiling effect, and data distribution (by looking at skewness and kurtosis) between 4 different kinds of satisfaction scales asking about the helpfulness of a visit. Additionally, we hypothesized that there is no correlation between scaled satisfaction and psychological status. Finally, we assessed how the satisfaction scores compared to the Net Promoter Scores (NPS). NPS are commonly used in the service industry to measure customer satisfaction; we are using these scores as a measure of patient satisfaction.