Providers generally have limited time to digest PRO information before meeting with patients and are primarily looking for trends and general comparisons in their review of ePRO reports for clinical care. Providing simple, commonly recognized graphs increases both the accessibility and efficient uptake of information presented to providers, which supports the clinical use of ePROs in busy healthcare delivery settings. In addition, in the spirit of patient engagement, when providers share ePRO graphics with a patient during a patient encounter, they need graphic formats that patients can easily understand.
It is crucial for designers to recognize the dynamics of the clinical use of PROs (time constraints, purpose, and patient sharing) and the need to balance simplicity with detail. Graphic choices generally should be simple (easy to interpret) yet presented in a way that enables users to discern trends and relative comparisons. As a result, line graphs and bar graphs are often an appropriate choice for displaying longitudinal and comparative PRO data (Table 4G). If patient engagement is a core driver of the use of ePRO graphics, pictographs (Table 4G) may be appropriate for graphics of PRO data being shared with patients.
Pie charts are another common type of graph; however, they come with some cautions when it comes to ePRO report application. While pie charts (Table 4G) may effectively show comparison in the case of a very limited number of properties, they lose their practical effectiveness as more comparators are added and additional screen space is needed to showcase the size and color of the contrasting wedges. For example, a pie chart may be an acceptable way to visually depict the proportion of people who experience a common side effect when starting a new medication. It becomes more challenging when we attempt to add details such as age, sex, or race to the pie chart to inform the provider about the patient and/or sub-population status or situation.
Caution is needed with simple, common graph types. For example, connecting PRO scores at points of time on a graph to form trending graph lines has intrinsic imprecision, particularly when patient status varies widely between the collection points. Furthermore, while trends and general comparisons may be detected with these graphics, they do not provide the level of precision and statistical detail supported by more complex graphics, such as a box plots. One must balance these cautions with the nature of the use of the graphs.
Quote from Provider Interviews:
“I think it has to be something that people have seen before and are able to interpret.”
While providers may prefer complex representations for research purposes to show distribution or provide statistical precision, it may take too much time to decipher more complex representations and statistical data during clinical encounters. Table 4G highlights graph types that may be more appropriate for precise analytical purposes.
Table 4G: Example graph types for reporting PRO data
|Graph types generally considered easy to understand for the purposes of PRO clinical application to support patient encounters
|Typically used for longitudinal PRO information; may be used for comparative PRO information with the addition of multiple lines.
|May be used for either comparative or longitudinal PRO information
|May be used for comparative PRO information
|May be used for presenting comparative information
|Graph types generally considered to be suitable for more precise PRO analytical purposes
|Typically used to show the distribution of data points.
|Floating column graph
|Typically used for comparative PRO information
|May be used for comparative or longitudinal PRO information
Some of the prevailing challenges for operationalizing these strategies are listed below, along with tactics to consider in addressing these challenges:
Challenges and Tactics Simple Graphs
- Scale of X and Y axes: The scale used to represent time (X axis) and PRO Score (Y axis) is critical to interpretation.
- Choice among simple: What if multiple graphs are “right” choices?
- Missing data points: Data is not always consistently reported by patients, which can lead to missing data points.
- Carefully consider what time and PRO score scales support the clinical context.
- Provide graph type (e.g., bar or line) presentation options.
- Represent missing data points on visualizations (may be easier on a bar graph than line graph).
For additional insight on providing appropriate graph and table visualizations on ePRO reports, see Recommended Resources related to: