Using statistical analysis tools within organizational improvement practices in clinical settings plays a key role. Data-driven and evidence-based improvement practices are valuable because they allow for test-evidenced change. One such tool is the one-way ANOVA test, which evaluates for differences between the means of more than two cohorts. In a clinical organization, customer satisfaction is an essential measure of operational effectiveness. The satisfaction factor can be measured on a hundred-point scale for each of the clinic’s employees, allowing for the collection of individual satisfaction statistics. However, as required by ANOVA, satisfaction is not a continuous variable, so this discrete scale can easily be converted to a continuous scale using standard transformations, such as logarithmic algorithms – this will be the dependent variable for the ANOVA test. In the case of the independent variable, the age of the patients can be measured. All patients could be divided into age groups: 18 to 24, 25 to 34, 35 to 44, and over 45. This would create four cohorts for which average customer satisfaction values would be measured. The figure below shows the results of an ANOVA test for dummy data in which 95 patients reported satisfaction with the services of specific (n = 8) clinic staff. Particularly noteworthy in these results is the statistical significance score: it shows that for all employees, there was no significant difference in satisfaction scores among clients of different ages. On the contrary, if the statistical significance for specific cells was lower than 0.05, it would imply that patients from a specific age group are less satisfied with the services of an individual staff member, and, therefore, this requires attention. Low patient satisfaction can cause errors in treatment or refusal to listen to physician recommendations, so studying this pattern proves valuable to the clinical company. An additional test that should be performed after the ANOVA is the Post Hoc analysis, which allows for determining the location of changes. For example, ANOVA can only report a difference in satisfaction for a particular employee among different age groups of clients (p <.05), but Post Hoc can show between which cohorts changes were found.
Reference
Laerd. (2020). One-way ANOVA in SPSS Statistics. LS.