An Animated Functional Data Analysis Interface to Cluster Rapid Lung Function Decline and Enhance Center-Level Care in Cystic Fibrosis
Table 2
Point-of-care scenarios and animated FPCA dashboard utility.
Scenario
Decision support offered by dashboard
Small center clinician benchmarking
Clinicians at Center 9138, one of the smaller centers, would like to examine the extent of rapid decline at their center and compare it to a larger center, for example, Center 9078. The clinicians can acquire an overview of these two centers on the static images’ link of the dashboard (Figure 6). While the late decline phenotype appears similar at the centers (green curves), there are differences between centers with respect to both early and middle decliners (red and yellow curves, respectively). The clinicians can also compare their center to other centers. Variability at each center may be examined by selecting the first link on the landing page and selecting animation for the smaller center and any comparative centers.
Patients who receive care within a single large center
For clinicians and care teams at a larger center, such as Center 9062, it may be of interest to examine trajectories of lung function for patients within their center and how they cluster according to rapid decline. Clinicians can select the graphics tile corresponding to their center using the first link on the landing page.
Identifying at-risk patients for implementation research
For researchers preparing to initiate a new algorithm to improve lung function at their center, the dashboard could be used to select patients at the highest risk of rapid decline by examining the profiles and prevalence of early decliners within a given center. Furthermore, the dashboard could be used to examine the prevalence of early decliners across centers, if the researchers are planning to implement the algorithm at multiple sites. For this case, the dashboard enhances comparisons between centers.
Individual patient monitoring
At a single center, a clinician is preparing for outpatient visits in the coming week. The dashboard could be used to track his specific outpatients and their status as early, middle, or late decliners. Those classified by the algorithm as early decliners could be tagged and additional care regimens could be implemented, for example, psychosocial assessments and checking adherence [15]. It is also possible that these patients could be offered clinical trial participation based on their trajectories.
