Evidence and decisions

Interpreting patient-reported outcomes and quality of life

Read symptom, function and quality-of-life measures with attention to validity, meaningful change and missing responses.

Patient reported outcomes (PROs) allow research participants to provide valuable insight into their subjective experience, which cannot be captured through objective laboratory measurements alone. Examples of these include pain, fatigue, difficulties with daily activities and the burdens associated with treatments. These experiences can be documented through patient reported outcomes.

The defining characteristic of a PRO is that the report is provided by the individual experiencing the outcome themselves. If assessments are provided by third parties they are considered different types of measurements. In order to understand the type of data being used by the researcher, the reader needs to determine if the survey was completed by the participant or a proxy and if anyone assisted the participant in interpreting their answers.

Determine what is being measured

Labels like "quality of life" encompass a broad spectrum of possible experiences. For example, a quality of life questionnaire may assess physical function, emotional wellness, social interaction and other similar areas, whereas a symptom scale will often narrow down and focus specifically on a certain aspect of experience. The specific questions and method of scoring need to support the particular interpretation of the findings presented in the article.

In addition, many instruments have been developed for use in one particular context or population. Therefore, using evidence from a different context, population or language may not be sufficient when proposing to utilize an instrument in a new or alternate context. Simply changing the words, response options or methodology of administration can potentially impact the meaning of an instrument's score. Familiarity with an instrument's name is no guarantee against this occurring.

The FDA's fit-for-purpose approach connects the intended use of the measure, the individuals taking the measure, and the interpretation of their responses. This includes developing a measurement argument that supports both the study design and the intended use of the instrument.

Understand the scale before interpreting differences

Scores on different instruments do not necessarily have the same direction. For example, higher scores indicate greater function on one scale and more severe symptoms on another. Without knowing the minimum and maximum values of a scale and/or the direction of the scale, it is impossible to accurately interpret results. Scores can be represented in various ways including total scores, subscale scores, standardized scores and change from baseline. Each represents a unique perspective regarding the data collected.

For instance, assume there is a fictional functional questionnaire that has scores ranging from 0 to 100 with higher numbers representing better functioning. Two groups demonstrate improvements on this scale. one group demonstrates an average increase of 12 points while the second group demonstrates an average increase of eight points. The difference between the average increases is four points. However, simply stating that one group demonstrated an average increase of twelve points ignores potential gains experienced by the second group.

Uncertainty surrounding this four point difference is important. Furthermore, understanding the implications of this level of difference is crucial to the populations being studied. A statistical difference may not necessarily be clinically significant to participants.

Distinguish between types of measurement data

Three key terms related to measurement data include validity, reliability and responsiveness. Validity refers to empirical evidence that supports the intended interpretation(s) of scores. Reliability refers to consistency in obtaining results when testing occurs under specified circumstances. Responsiveness refers to an instrument's ability to detect true change in an individual's experience or construct over time.

Achieving success on one of these dimensions does not imply success on any or all of the remaining dimensions. For example, an instrument can produce consistent results without providing evidence for every desired interpretation. Establishing a meaningful change threshold requires justification for both the instrument and context in which it will be utilized. There is a critical difference between change experienced by an individual versus a change observed between two group averages. Therefore establishing a threshold for one should not be assumed to apply to the other. Additionally, presenting the percentage of participants meeting a justified response criterion along with an average result provides additional information that can aid in interpreting results.

Address missing questionnaires

Individuals responding to follow up questionnaires may differ from non-respondents. Factors contributing to missingness may include illness, dissatisfaction with previous evaluations, difficulty answering questions, etc. Although demonstrating a favorable average score among responders may suggest positive experiences among at least some participants, important experiences may remain unrepresented due to missing data.

To illustrate this concept; consider two study groups entering with 100 participants each however only 90 in group a and 65 in group b respond to follow up questionnaires. Any observed differences between respondents require thorough investigation. Specifically examine why data were missing and identify any assumptions made during subsequent analysis as well as any sensitivity analyses conducted. Replacing missing values with previously recorded values or only analyzing complete cases does not inherently eliminate bias.

Timing when follow up questionnaires are completed may also influence responses. Respondents may have varying degrees of stress and anxiety after appointments compared to those responding after less stressful periods. Specifying advance evaluation timeframes prior to study commencement clearly identifies planned comparisons and limits opportunities to choose evaluation time frames resulting in more favorable results.

Document results alongside other relevant outcome measures

Under the CONSORT PRO reporting guidelines, researchers should describe their hypotheses for patient-reported outcomes, relevant measurement characteristics, strategies employed for handling missing data and limitations associated with utilizing patient reported measures. Researchers can assist readers in evaluating trials through clear documentation; however it does not serve as certification of either the measure or trial itself. Symptoms and function must also be evaluated along with other relevant study outcomes including adverse events and survival as applicable. An average gain may mask large variations among individual patients' experiences. Therefore providing adequate descriptions regarding what was measured and what the study suggests, along with uncertainty, is necessary as well as documenting which domains were measured for which subpopulations, whose responses were excluded from consideration in analyses, etc.

Methodology references

  • FDA: Patient-Reported Outcome Measures, Guidance for Industry (2009)
  • FDA: Selecting, Developing, or Modifying Fit-for-Purpose Clinical Outcome Assessments, final guidance (October 2025)
  • Calvert et al. Reporting of Patient-Reported Outcomes in Randomized Trials: CONSORT PRO Extension