Testing Ways to Display Patient-Reported Outcomes Data for Patients and Clinicians

• Download Summary
• Español (pdf)
• Audio Recording (mp3)

What was the research about?

Information that patients provide about their health problems or treatments can help them and their clinicians, such as doctors and nurses, manage patients’ care. One type of information, called patient-reported outcomes, comes from surveys that ask patients how they are feeling and how bad their symptoms are. The surveys also ask patients how well they can function in daily life. The survey results help patients and clinicians track how patients are doing over time; they also help them make treatment decisions. Also, researchers can combine survey results from many patients. These results help researchers see how well treatments work to improve symptoms and well-being.

Patients and clinicians need to understand the survey results for them to be useful in treatment decisions. How well patients and clinicians understand results may depend on the display format of the results.

The study compared three ways to show survey results. The research team wanted to learn which display formats were the best understood. The team also wanted to learn which formats were clearest and most useful.

What were the results?

Displays to help patients track illness symptoms and their health. The research team displayed a series of line graphs that showed a patient’s symptoms and side effects over time. Patients, clinicians, and researchers thought these displays were clear and easy to understand when the graphs’ lines moved up to show better health for both symptoms and function. They found graphs less clear when lines moved up to show that the patient had more symptoms or side effects. They found it helpful when graphs included a line showing the point at which scores became concerning.

Displays to help patients compare treatment options. The research team also displayed bar charts, pie charts, and other displays to help patients choose between treatment options. The displays compared how different treatments affected symptoms and function. Among these displays, patients, clinicians, and researchers understood the pie charts best and found them the clearest and most useful in comparing treatments. For line graphs, they understood the information better when the lines moved up to show better health for both symptoms and function, compared with graphs that had lines that moved up to show that the patient had more symptoms or side effects.

Displays to help clinicians compare treatment options. The research team displayed bar charts, pie charts, and line graphs to help clinicians choose between treatment options. The displays compared how different treatments affected symptoms and function. The team added symbols to show when the differences between two treatments were important. For line graphs, clinicians and researchers liked having symbols to show when there were important differences between treatment choices. They found bar charts and pie charts equally clear.

Who was in the study?

The study included 1,256 cancer survivors and 608 cancer clinicians. The study also included 747 patient-reported outcomes researchers.

What did the research team do?

With input from patients and clinicians, the research team designed different types of displays of patient survey results. The team also used a survey to see if patients, clinicians, and researchers understood each graph correctly. The team also asked them how clear and useful each display was.

What were the limits of the study?

The study included patients with cancer whom the research team could reach easily through email and social media. Results might be different for patients who have other health problems or for cancer patients who couldn’t be easily reached through the internet.

How can people use the results?

These results could help researchers develop materials that help patients and clinicians track patients’ symptoms over time and also help them make treatment decisions. Researchers could use the results to choose which display graphs to use in research articles to help doctors understand patient-reported outcomes.

Objective

To investigate ways to visually display individual and aggregate patient-reported outcomes (PRO) data to help patients, clinicians, and researchers understand and use the data

Study Design

The research team conducted an online survey of 1,256 cancer survivors, 608 cancer clinicians, and 747 PRO researchers to compare data-display formats for PRO data. PRO data are outcomes reported by patients about their health conditions and treatments. These data may include measures of symptoms, functional status, and health-related quality of life, among others. The research team worked with patients and clinicians to create the formats for the data displays. The team tested data displays for three purposes:

• To help patients and clinicians track symptoms and functioning for an individual patient over time
• To help patients compare treatment options using aggregate PRO scores from multiple patients
• To help clinicians and researchers compare treatment options using aggregate PRO scores from multiple patients

Results

Helping patients and clinicians track symptoms and functioning. Patients, clinicians, and researchers viewed three display formats for showing clinically concerning scores. Participants preferred a display format with threshold lines, saying that they found it clearer than other formats to indicate whether scores were normal or clinically concerning.

In addition, patients, clinicians, and researchers viewed one of two approaches to score directionality. Participants interpreted the information more accurately and rated clarity better when higher scores reflected better health outcomes than when higher scores indicated that more of an outcome was present.

Helping patients compare treatment options. Patients, clinicians, and researchers viewed bar graphs, pie charts, and icon arrays that displayed aggregate PRO scores from multiple patients, showing the proportion of patients whose conditions improved, stayed stable, or got worse. Participants found that pie charts were easiest to interpret accurately, and perceived them to be clear and useful.

For displaying average scores over time, participants looked at three types of line graphs. Participants interpreted line graphs of scores over time most accurately when higher scores indicated better health outcomes.

Helping clinicians and researchers compare treatment options. Clinicians and researchers viewed bar graphs and pie charts that displayed aggregate PRO scores from multiple patients, as well as the same line graphs described above. In addition, the research team tested other variations of line graphs, including displaying p values, confidence intervals, and indications of clinically important differences.

Participants interpreted the information from pie charts and bar graphs equally accurately and also rated their clarity and usefulness the same. For line graphs, participants interpreted the information more accurately and rated displays as clearer when higher scores reflected better health outcomes as opposed to scores compared with a population average, and also found them clearer when they had a graphical indication of clinically significant differences between groups. Overall, clinicians and researchers preferred the versions with confidence intervals.

Limitations

The results may not be generalizable to patients who do not have cancer or clinicians who do not treat cancer. The research team did not assess how personal characteristics affected participants’ responses to the survey.

Conclusions and Relevance

Patients, clinicians, and researchers could use these formats for PRO data displays in patient monitoring and educational materials and decision aids, and when publishing in peer-reviewed journals.

Future Research Needs

Future research could examine the effectiveness of materials using these display formats.