Qualitative development of the PROMIS Profile v1.0-Familial Chylomicronemia Syndrome (FCS) 28

First, two Northwestern University (NU) researchers with extensive experience with PROMIS and PROM development independently reviewed concept elicitation interview transcripts and results from a previous study of FCS quality of life conducted with ten individuals with FCS in the United States (data on file, Ionis) [18]. The interviews consisted of open-ended questions about FCS symptoms, symptom frequency and severity, and impacts of FCS on daily life. The researchers met to discuss their impressions of the most important symptoms and impacts represented in the data and study report, and created a preliminary list of key symptoms and impacts for FCS.

The NU study team conducted a search in PubMed using the following terms: “(familial chylomicronemia syndrome) OR (lipoprotein lipase deficiency) OR (hyperlipoproteinemia type 1) AND (quality of life).” This search strategy was then adapted for Google Scholar. Relevant articles were extracted and reviewed by the team. The NU team also reviewed reference lists of extracted articles to ensure capture of key references. The study team reviewed the pertinent literature to confirm findings from their review of the qualitative transcripts and to consider whether there were additional symptoms and impacts central to the experience of living with FCS.

Next, the team identified existing items from the PROMIS item banks, which currently contain over 1900 items, to represent symptoms and impacts included in the previous FCS study [18] and/or in the literature review. If no PROMIS items existed for a particular concept, other HealthMeasures measurement systems (https://​www.​healthmeasures.​net/​) were reviewed for possible items.

The number of items in the pool was reduced via a series of meetings of the NU study team and joint meetings with FCS and patient-reported outcomes experts from Ionis. These meetings followed the PROMIS measure development methodology (i.e., the “item-review process”) and aimed to eliminate items that were redundant, confusing, poorly written, or did not adequately represent the symptoms and impacts identified in the patient data and literature [16, 19].

Although items drawn from PROMIS have undergone cognitive debriefing to evaluate language, comprehensibility, ambiguity, and relevance [16], we conducted cognitive interviews of the draft item set with individuals with FCS to confirm item clarity, meaning, and relevance for FCS. Participants for the cognitive interviews were recruited from a sample of individuals with FCS who participated in the 2018 study, “PROMIS®-Based Survey of Health-Related Quality of Life in Familial Chylomicronemia Syndrome.”[6]. All participants reported that they had a diagnosis of FCS, were 18 years of age or older, and lived in the United States. Participants who completed the PROMIS-Based Survey study (N = 25) were contacted via email and invited to participate in the telephone cognitive interview. Up to three recruitment emails were sent to each individual. Interested, eligible individuals completed informed consent, and the study coordinator scheduled their interview at a time that was convenient for them. Participants received a copy of the draft measure via email or postal mail prior to the interview.

Interviewers first collected sociodemographic and key disease information from the participant. Next, participants completed the draft measure and the interviewer led them through a series of questions about the measure, using a semi-structured cognitive interview guide based on the work of Willis [20] to ascertain comprehension of the measure items and the response processes. Specifically, the interviewer asked participants to: (1) describe how they arrived at their answer; (2) restate each item in their own words; (3) discuss the clarity of the item; (4) describe any questions they had about the item; and (5) indicate whether the question was relevant to their experience. Participants received a $100 USD electronic gift card for participating. Trained interviewers took detailed field notes, and interviews were audiotaped to ensure comprehensive capture of all relevant information. Cognitive interview recordings were transcribed and transcripts were de-identified. Transcripts were used to confirm field notes and to provide supporting quotations.

We identified nine key published articles about patient-reported symptoms and HRQOL in the context of FCS [3‐11]. In the prior qualitative study by Davidson and colleagues, FCS interview participants reported 16 symptoms of FCS [18]. Of these, the most prevalent/important symptoms were abdominal pain, diarrhea, brain fog, and fatigue. Commonly reported symptoms per the key literature were abdominal pain, bloating, fatigue [5, 7]. Emotional, social and cognitive impacts of FCS in the literature included anxiety, cognitive difficulties, and work and social limitations [5, 7]. Based upon our review of the key literature and the Davidson article and data, we identified ten important FCS symptoms and 12 impacts, for a total of 22 key concepts (Table 1). The following symptoms, which were mentioned by patients in the prior qualitative study, were excluded from our list of the most import symptoms: blurred vision, poor appetite, difficulty concentrating, weight loss, indigestion, muscle weakness. The 12 impacts shown in Table 1 expand upon the findings by Davidson by, for example, detailing specific impacts related to mental and emotional well-being and adding the concept of sleep disruption [6].

Next, existing items (N = 96) were identified to represent the 22 key concepts. Existing items were identified for all symptoms and for every impact except “impact of dietary restrictions,” as this concept does not fall within the scope of HRQOL as measured by HealthMeasures. Thus, this concept was excluded. Identified items representing the remaining 21 key concepts (ten symptoms and eleven impacts) originated from PROMIS, Neuro-QoL, and the Adult Sickle Cell Quality of Life Measurement System (ASCQ-Me®), which is also a HealthMeasures measurement system. An item addressing financial strain was draw from the Functional Assessment of Chronic Illness Therapy (FACIT) measurement system [21].

A series of investigator meetings were held to reduce the item pool. Item retention decisions were based on identifying items that best fit the FCS symptoms and impacts and provided a variety of response options for key concepts (e.g., frequency, intensity, and interference). When there were comparable choices, preference was given to items from the PROMIS 29 + 2 Profile v2.1. The PROMIS 29 + 2 has been included in other FCS clinical trials, using items from the measure would therefore reduce response burden and ensure inclusion of items that have been tested and used extensively [22, 23]. The team reviewed the following information for each concept: definition of the concept, priority of the concept for inclusion in the measure, article/data supporting the concept, draft items and the source, time frame, and response option for each draft item. The team discussed which item(s) best fit each concept, as well as whether alternate items should be sent to cognitive interviewing. For example, five items in the pool covered the concept of abdominal pain. Of these, three items were retained. The item, “How often did you have discomfort in your belly?” was dropped in favor of items that used the word “pain” as pain was more consistent with the literature and data. The item “How much did belly pain bother you?” was dropped in favor of items on belly pain frequency, worst pain, and interference in day-to-day activities. Using this general approach, the 96 items were reduced to 32 items. These 32 items became the draft measure presented to participants in the cognitive interviews. Additionally, the concept “negative career impact” was dropped as the concept was outside the HRQOL of HealthMeasures, leaving 20 key concepts.

Twelve individuals responded to our email request. Of these, two declined to participate, one was ineligible, and one did not attend their scheduled interview. Cognitive interviews were completed with eight unrelated individuals with FCS. The cognitive interview sample characteristics (N = 8) are shown in Table 2. The sample was primarily female (n = 6, 75.0%). All participants indicated that they followed an FCS diet at least some of the time. Six participants (75.0%) were experiencing symptoms of FCS at the time of the interview. The mean number of self-reported acute pancreatitis attacks over the last 5 years was ten (range 0–50).

Face validity was assessed with the following interview question: “Please take a moment to look over the questions again. Do these questions, in your opinion, capture your experiences with FCS?” All 8 participants answered yes to this question.

Content validity was assessed in two ways. For each item on the measure, participants were asked, “What kinds of things did you think about when you answered the question?” Responses were analyzed to determine whether participants were interpreting the items in ways that were consistent with intended item meanings. None of the items were found to be misinterpreted by participants.

For 30 of the 32 draft items, every participant said the meaning of the item was clear. For the item, “I have trouble thinking clearly,” participant 002 said they thought the item could be made clearer by specifying trouble thinking due to FCS. For the item, “I have trouble doing all of my usual work (include work at home),” participant 004 thought that the meaning of work from home was confusing because so many people are working from home due to the COVID-19 pandemic. The participant suggested using the phrase “household chores” or “domestic things” instead of “work from home.” No changes were made due to these comments.

The draft measure included 5 items related to difficulty thinking (Table 3). Participants were asked which of the items best fit their FCS experience. The most frequently chosen item was, “My thinking has been slow,” which was chosen by four participants (50.0%). The next most commonly selected item was, “I have been able to concentrate,” which was selected by three participants (37.5%). The item, “I have been able to remember to do things, like take medicine or buy something I need,” was selected least often (n = 1, 12.5%).

The draft item set included two items regarding worry about one’s health (Table 4). Participants preferred the item, “I worry that my condition will get worse” (seven participants vs. five participants). Four participants did not prefer one item over the other and noted that useful, unique information could be obtained from each item.

Following cognitive interviews, the investigator team met three times to finalize the measure. During these meetings, the team reviewed the 20 key concepts and cognitive interview results. Twenty-eight items were retained for the FCS measure and four items were removed. For each item removed, alternate items were preferred by the study team and/or cognitive interview participants (Table 5). When possible, items from the PROMIS 29 + 2 measure were utilized so as to reduce participant burden in situations where the PROMIS 29 + 2 is being used in research or clinical settings.

The PROMIS Profile v1.0—familial chylomicronemia syndrome (FCS) 28 (PROMIS FCS 28) is shown in Table 6. The measure contains the most important symptoms for FCS (16 items, first column), including abdominal pain (three items), pain (two items), fatigue (two items), and brain fog or cognitive difficulties (three items). Given the importance of pain to FCS patients, the PROMIS FCS 28 assesses the frequency of, intensity of, and interference due to abdominal pain. Additionally, key impacts of FCS are including in the measure (12 items, second column), such as worry (one item), impacts on social activities (two items), physical activity (two items), and productivity (two items).

While the broad applicability of a PROMIS-based measure increases its reach and relevance, developing a condition-specific measure from PROMIS does present certain limitations. For example, whereas PROMIS domains include a wide-range of content, we were limited to the existing item wording and response options. However, cognitive debriefing showed that the chosen items worked well for FCS—their meaning was clear, relevant for FCS, and the response options worked well. Future work can evaluate the extent to which assessment of dietary intake and career impact add additional value. Additionally, the investigator team did not include anyone living with FCS. As such, the methods utilized may not have targeted the needs and priorities of FCS patients as optimally as possible. Another limitation is that the sample was relatively homogenous and the response rate to recruitment attempts was approximately 50%. Although participants were demographically similar to those who participated in past studies of patients with FCS, it is possible that results may have been impacted by selection bias, thus limiting generalizability. Finally, the psychometric properties of the PROMIS FCS 28 were not evaluated. Given the rare nature of FCS, the scientifically strong development of this measure represents notable progress toward the appropriate assessment of patients living with this disease and will enable improved understanding of the patient experience. Additionally, as outlined above, this approach for adapting PROMIS measures has yielded measures with strong psychometric properties in other populations, suggesting the same will be true of patients with FCS. However, it will nonetheless be important to evaluate the reliability and validity of this condition-specific tool in patients with FCS.