Including Patient Input In Patient-Reported Outcome Instrument Development: Examples From the MDASI-CML

Loretta A Williams, Jorge E. Cortes, Patricia S Ault, Araceli Garcia-Gonzalez, Janet L Williams and Charles S Cleeland


Abstract 2573

Introduction: The United States Food and Drug Administration (FDA) recognizes patient-reported outcomes (PROs) as acceptable measures of treatment benefit and risk in medical product clinical trials. The FDA requires that patient input be included in the development and testing of PRO instruments. We have adopted a three-step process for the development of multi-symptom PROs that includes patient input in each step. This method is being used for the development of the M. D. Anderson Symptom Inventory (MDASI) for Philadelphia-chromosome-positive chronic myeloid leukemia (CML). The MDASI is a PRO measure of symptom burden, defined as the combined impact of all disease-related and therapy-related symptoms on one's ability to function as one did before the onset of disease or therapy, in patients with cancer. The core MDASI includes 13 symptom severity items and 6 interference items rated at their worst in the last 24 hours on a 0–10 scale, with 0 meaning no symptom or interference and 10 meaning as bad as can be imagined or complete interference.

Patients and Methods: The first step in the process was qualitative interviews with 35 patients with CML about their symptom experiences. Symptoms were extracted by descriptive exploratory analysis from interview transcripts. Step 2 was grading of the relevance of the symptoms (0 to 4 scale) from Step 1 by an expert panel that included professional care providers as well as patients with CML and their family caregivers. Symptoms that received a mean relevance rating of ≥ 3 were added to the 13 symptom items and 6 interference items of the core MDASI for validation in Step 3. One hundred and sixty patients with CML are completing the experimental MDASI-CML, which will undergo psychometric validation and item reduction. The first 30 patients in Step 3 completed a cognitive debriefing interview about their experience of completing the MDASI-CML.

Results: Patient characteristics are in Table 1. No personal information was collected on the expert panel members.

Table 1.

Patient Characteristics

Thirty-nine symptoms (13 core and 26 CML-specific) were extracted from the Step 1 qualitative interviews and rated by the expert panel. Four physicians, 5 nurses, 3 patients, and 3 family caregivers returned ratings. Six of the 26 CML-specific items had mean relevance ratings of ≥ 3. The experimental MDASI-CML includes the 13 core symptoms, 6 CML-specific symptoms, and 6 core interference items (Table 2).

Table 2.


During the cognitive debriefing, over 80% of patients reported that the MDASI-CML items were not at all difficult to complete or understand, were completely comfortable to answer, and were not repetitive, and that the 0–10 scoring system for rating severity of symptoms and interference with daily activities was very easy to use. Nine patients listed 12 additional symptoms that should be included, but each symptom was only mentioned by 1 patient and had already been eliminated because of low relevance ratings by the expert panel.

Conclusions: Inclusion of patient input at each step of PRO development ensures that the instrument measures what is important to patients and enhances content validity. It further ensures that the measure is easy to understand and complete. Psychometric validation of the MDASI-CML is proceeding.

Disclosure: No relevant conflicts of interest to declare.