[WEBPAPER] Patient Reported Outcomes: how to develop it and why use it in clinical studies?

« On a scale of 1 to 10, how do you feel today? » This is one of the questions that many patients may be confronted with in a clinical trial. Indeed, the state of a patient’s health can be assessed by taking into account the opinion of patients on various themes. When objective biological markers are not available or not relevant for the goal of the study, research questionnaires can be used. These can be completed by patients themselves (Patient Reported Outcomes – PROs) or by other people such as the parents in case of infants. Thus, the questionnaires make it possible to assess the patient’s symptoms, feelings, pain, anxiety, depression, fatigue, physical abilities, quality of life, etc. They are not a simple reflection of the quality of life but allow us to understand how patients feel and react following the intake of a product for example.

Clinical health questionnaires: a useful tool

MPQ, GIQLI, GSRS, PROMIS-29(1)? This is not a password for your internet box but these acronyms stand for research questionnaires.  There are many questionnaires and ways to measure the quality of life of patients. They can be divided into two categories: 

Generic quality of life questionnaires, applicable to all clinical situations and patient populations, for example: 

• SF-36(2): the Short Form (36) Health Survey
• EQ-5D(3): the EuroQol Group 5-Dimension Self-Reported Questionnaire

Specific questionnaires, applicable to particular populations, pathology, or specific outcome dimensions, for example: 

• Fatigue Severity Scale, concerns clinical studies related to fatigue
•  IBS-SSS(4) (IBS-Symptom Severity Score) for clinical studies concerning irritable bowel syndrome. 

Validation of questionnaires

First of all, in clinical research, the used questionnaires must be validated, this requires  the author of the questionnaire to carry out a specific study for the validation of the questionnaire and to publish it into a proper journal. Furthermore, the questionnaire is validated on a certain population and therefore must be used afterwards on the same target population and be delivered under conditions similar to its validation. There are several types of validation, for example:

• Content validation: The items (questions in a questionnaire) must be consistent with what is expected. We are not going to ask questions about cancer if we want to study fatigue.
• Face validity: to ensure that the patients understand the question in the same way as the health experts. 

In addition to validity, the questionnaire must also be reliable, meaning that measures are accurate. There must be a good consistency in the answers. For instance, the scores filled in by the patient at two measurement times with a stable clinical condition must be close (precision performance of the questionnaire). Also, these scores should really match the patient’s clinical condition (trueness performance of the questionnaire). There are different methods to calculate this reliability, such as the Spearman formula, the Rulon-Guttman approach or the Guttman L2. The threshold of 70% is often considered as acceptable for reliability.

In addition, a questionnaire developed and validated in one language needs to follow a specific procedure to be validated in another language. Cultural adaptation can be necessary and translation needs to be validated for the correct use.

Adapted questionnaires

Moreover, the choice of the right questionnaire for the study topic is one of the key elements. Indeed, the questionnaire needs to be validated for the targeted population and be clinically relevant for the expected effects of the products/intervention. Some questionnaires are specifically developed to provide a score that can be used as a diagnostic score with specific predefined ranges for example, and others are used for a longitudinal analysis, but also a combination  of both objectives is possible. One clinical study can require the use of several different PROs. 

Another important aspect is the patient burden, and this must be considered to ensure the reliability of the data(5). If the effort of the patient to complete a questionnaire is too big the quality of data can be affected (less accurate answers, etc.). Depending on the questionnaires, the frequency of measurement differs, i.e. patients may respond daily, weekly or at each medical  visit. This frequency should allow capturing changes in the patient’s quality of life while avoiding overloading the patient with questionnaires, in order to avoid missing or incorrect data, which complicates data analysis.

To conclude, Patient Reported Outcomes can be very useful and relevant into clinical studies but must be selected carefully. 

References:

1. McGill Pain questionnaire / The Gastrointestinal Quality of Life Index / Gastrointestinal Scale Rating Symptom / Patient-Reported Outcomes Measurement Information System 

2. The Short Form (36) Health Survey

3. EuroQol Group 5-Dimension Self-Reported Questionnaire

4. IBS-symptom severity score 

5.  J. C. Cappelleri et al., « Patient-Reported Outcomes: Measurement, Implementation and Interpretation »