A qualitative study to explore the symptoms and impacts of Crohn’s disease and to develop the Crohn’s Disease Diary

This was a cross-sectional, qualitative interview study in the USA, involving 3 rounds of semi-structured combined concept elicitation and cognitive debriefing interviews (Fig. 1) conducted between March and September 2019. Before the first round of interviews, a targeted review of literature was conducted. Articles, conference abstracts, and studies on the patient experience of Crohn’s disease published between 2008 and 2018 were used to develop an initial conceptual model to describe signs and symptoms of Crohn’s disease. Based on the literature review, a draft conceptual model and preliminary 19-item daily symptoms diary was developed for the study. Items included in the 19-item diary represented key symptoms and impacts of Crohn’s disease (including extraintestinal symptoms). Response options were item specific, depending on the type of rating scale considered most appropriate for each item (e.g., 0–10 numerical scales or Likert scales). Patients were asked to recall key symptoms and impacts over the past 24 h and extraintestinal symptoms over the past 7 days.

The study was approved by an independent scientific review committee (The Copernicus Group, Cary, NC; IRB tracking number: 20190635, IRB amendment tracking number 20190635). All patients and clinicians provided oral and written informed consent prior to participating in any study-related activities. Written permission was obtained from the parents/guardians of adolescent patients and written assent was obtained from the adolescents.

Medquest Global (Valencia, CA), a third-party recruitment agency assisted with the recruitment of patients from four geographically diverse locations in the USA (New Orleans, Louisiana; Baltimore, Maryland; Los Angeles, California; and Chicago, Illinois). Patients were recruited via referrals from gastroenterologists and general practitioners. Age, sex, race, education, disease severity, and treatment sampling quotas were applied to the sample patient population to ensure a range of demographic and clinical characteristics. Sample size was determined based on the principles of conceptual saturation. It was estimated that a total of 36 patients (n = 12 for each round of interviews) would be adequate to achieve saturation and to ensure sufficient clinical and demographic diversity in each round of interviews [16‐18].

Adolescent (≥ 16 years of age) and adult patients with radiographically, histologically, and/or endoscopically confirmed Crohn’s disease, diagnosed ≤ 6 months prior to screening, who had experienced symptoms within the 3 months prior to the study were eligible. Severity of disease was based on patient-reported data collected via a 7-day diary card completed prior to study entry. Disease severity categories and detailed eligibility criteria are included in the Supplemental Materials.

Clinicians specializing in the management or treatment of adult and/or adolescent patients with Crohn’s disease on a regular basis (i.e., at least 8 patients/month) were eligible. Experience treating patients with biologic agents, fluency in spoken and written English, and access to a telephone or computer was also required.

The first and second rounds of combined concept elicitation and cognitive debriefing interviews were conducted by trained Adelphi Values interviewers via telephone and the third was face to face. Interviews were scheduled 7 to 12 days after screening. Each interview lasted approximately 90 min, with breaks offered to participants as needed. The time devoted to CE versus CD varied across the rounds, with greater focus on CE in round 1 (~ 40 min). To allow more focus on CD in rounds 2 and 3, less time was spent on CE (~ 20–25 min). In the concept elicitation portion, broad open-ended interview questions were posed in an unbiased manner to explore spontaneously reported patient experiences of Crohn’s disease. For example, interviewers asked patients to describe their experience of Crohn’s disease. These spontaneously reported symptoms likely reflect those which patients attribute most to their experience of Crohn’s disease, but to ensure that the full range of potentially applicable symptoms (as identified through literature review and clinician input) are explored, additional probing questions by interviewers are necessary. Thus, following the open-ended questions, more focused questions probed for topics of interest not spontaneously mentioned. Patient-friendly language was used when probing for symptoms. For example, patients were asked about the concept of urgency, which was commonly described as “needing to use the bathroom immediately” or “rushing to a toilet.” Similarly, to explore the concept of tenesmus, patients were asked about a “sensation of needing to go to the toilet but not passing anything.”

In the cognitive debriefing portion of the interview, patients participated in a “think-aloud exercise” during which they read and responded to the items in the daily symptoms diary, providing thoughts and opinions on each item. Patients were also asked detailed questions about item relevance, item wording, instructions, recall period, response options, ease of completing the items, and usability of the diary in an electronic PRO (ePRO) format.

Following each round of interviews, feedback from the concept elicitation portion determined if content modifications to the conceptual model and the daily symptoms diary were needed. Based on feedback from the cognitive debriefing portion of the interviews, patient understanding of the instructions, items, response options, and recall periods in the daily symptoms diary was assessed and wording could be modified for clarity. Additionally, item relevance and overlap among items was assessed to determine if items should be added or deleted prior to the next round of interviews. Items retained in the final daily symptoms diary (Crohn’s Disease Diary) also underwent a translatability assessment.

In addition to patient interviews, combined concept elicitation and cognitive debriefing qualitative telephone interviews were conducted with clinicians to gain clinical insights into the patient experience of Crohn’s disease in terms of symptom presentation and impact on daily life, as well as to provide feedback for the items included in the daily symptoms diary. Broad, open-ended questions were first used to prompt spontaneous responses, followed by more focused questions to probe on topics of interest or for clarification. The interviews were conducted in two rounds corresponding to the first two rounds of patient interviews.

All interviews were audio-recorded and transcribed verbatim. Data from adolescents (n = 4 for all rounds of interviews) and adults (n = 32) were pooled, as no conceptual differences between the age groups were expected or identified during analysis. Qualitative analysis of patient and clinician interview transcripts was conducted using ATLAS.ti version 7.2013 (Scientific Software Development GmbH B, Germany).

Coding schemes were derived for the interview transcripts using the first two transcripts. These schemes were used throughout the analysis to ensure consistent application and grouping of codes across transcripts. The concept elicitation sections of transcripts were analyzed using thematic analysis. Participant quotes that pertained to symptoms, patient experience, or patient impact of Crohn’s disease, and whether these were reported spontaneously or when probed, were assigned corresponding concept codes. The cognitive debriefing sections of transcripts were analyzed using a framework approach. Dichotomous codes were assigned to each item, instruction, response option(s), and recall period to indicate relevance, understanding, or difficulty to complete. Codes were also used to indicate why each response option was chosen and how the content of each item applied to the patient experience, as well as to provide suggestions for item, instruction, or response option wording or formatting changes, and general feedback on the daily symptoms diary. An induction-abduction approach was taken to identify themes in the data [19].

To ascertain whether conceptual saturation was achieved, transcripts were ordered chronologically, divided into four groups of 9 patients each and findings from each group of interviews were compared in an iterative fashion. A conceptual saturation grid was developed to summarize results and to determine if conceptual saturation had been achieved. Saturation was deemed to be achieved when no new concepts emerged in the last (fourth) group of interviews.

A total of 36 patients (n = 12 in each round) were interviewed and all sampling quotas were met. The mean age was 40.5 years (range: 16–74 years), there were slightly more females (21/36 [58%]) than males, and most patients were White/Caucasian/European (24/36 [67%]) (Table 1). Nearly two-thirds (22/36 [61%]) of patients were categorized as having moderate to severe Crohn’s disease. Time since diagnosis ranged from 6 months to 20 years and the sample was diverse in terms of location of disease (ileum, 25/36 [69%]; proximal small bowel 22/36 [61%]; colon, 22/36 [61%]; esophagus/stomach, 27/36 [44%]).

A total of 4 clinicians (round 1, n = 2; round 2, n = 2) were interviewed. Three clinicians had > 20 years of experience treating Crohn’s disease, one had 6 years, and all reported seeing 10 to 40 patients with Crohn’s disease each month.

The frequency with which patients reported symptoms spontaneously versus probed is shown in Fig. 2. Patients (N = 36) reported a total of 25 symptoms, broadly categorized as 12 core symptoms (bowel and abdominal symptoms), 6 extraintestinal symptoms, and 7 additional symptoms (Fig. 2). Every patient reported that they experienced abdominal pain, loose or liquid bowel movements, high or increased frequency of bowel movements, and bowel movement urgency related to Crohn’s disease, either spontaneously or after being probed. Abdominal pain was reported spontaneously by all patients. Most reported loose or liquid bowel movements (33/36 [91.7%]) and high or increased frequency of bowel movements (27/36 [75.0%]) spontaneously, whereas the majority reported bowel movement urgency only after being probed (22/36 [61%]). Examples of how the most frequently reported symptoms were described by individual patients include:

Although most symptoms were reported with similar frequency across disease severity groups, tenesmus, back pain, thirst/dehydration, weakness, joint pain, joint stiffness, and general body pain were reported more by patients with moderate to severe disease (Fig. 3). Fever and throat tightness were reported only by patients with moderate to severe disease. Examples of quotes from individual patients with moderate to severe disease include:

The 3 symptoms that patients experienced most often (abdominal pain, loose or liquid bowel movements, and high frequency of bowel movements) were also reported to be the most bothersome and most important to change in response to treatment (Fig. 4).

Patients also reported the impact of their disease on HRQoL. All patients, either spontaneously (33/36 [92%]) or probed (3/36 [8%]), reported that activities of daily living (ADL) were impacted by Crohn’s disease symptoms. Most patients (34/36 [94%]) reported emotional impacts, most commonly feeling worried or anxious about urgently needing to have a bowel movement. Work/school impacts were reported by 33/36 (92%), including the need for frequent bathroom breaks. Most patients (32/36 [89%]) reported that their social functioning was impacted, particularly when taking vacations or going to restaurants, and that the quality of social relationships was impacted (11/32 [34%]), including feelings of isolation and effects on sexual relationships. Physical functioning impacts were reported by 28/36 (78%), with difficulty carrying out sports/exercise reported most frequently. Most patients reported emotional, work/school, social, and physical impacts spontaneously (79% − 88%). Sleep was also affected for most (92%), but only 42% reported this spontaneously. Examples of quotes from individual patients describing the impact of their disease on HRQoL are as follows:

Two-thirds of patients spontaneously reported additional impacts that did not fit neatly into the general HRQoL categories above. Weight loss was reported by 12/24 (50%) (most frequently attributed to loss of appetite [5/12, 42%]), 8/24 (33%) reported fecal incontinence as an impact of bowel urgency, and 7/24 (29%) reported the need to lie down or rest when symptoms were severe. A small number of patients spontaneously reported rectal discomfort due to loose stools (5/24 [21%]), while others reported financial impacts (4/24 [17%]), including the cost of purchasing products required due to symptoms and lost income due to time off work.

The conceptual saturation grid was reviewed following the last (fourth) group of interviews. All symptoms emerged spontaneously within the first two groups of interviews and all impact domains emerged spontaneously in the first group of interviews, indicating that saturation was achieved.

Based on the initial conceptual model from the literature review, the concept of a “flare” is well recognized. Although not consistently defined, in clinical practice it is generally understood to mean a period during which disease symptoms are more active. Twenty-six patients were asked to describe their general experience of a flare. Some patients (5/26 [19%]) defined a flare as a general worsening of symptoms, some (12/26 [46%]) reported that abdominal pain increased, and others said that more frequent/severe loose or liquid bowel movements (5/26 [19%]) and more frequent bowel movements (4/26 [15%]) were indicative of a flare. Small numbers of patients (1 − 2 for each symptom) reported that the emergence of new symptoms such as bloating, fever, cold sweats/sweating, tiredness/fatigue, back pain, nausea, loss of appetite, skin rashes, or dehydration denoted a flare.

Clinician-reported data were generally similar to patient data regarding which symptoms were the most bothersome and most important to treat. In addition, clinicians emphasized the importance of recognizing fatigue as a symptom of Crohn’s disease. Consistent with results from patient interviews, clinicians reported that Crohn’s disease greatly impacts HRQoL.

Based on patient and clinician concept elicitation results, the conceptual model, previously informed by the literature review alone, was updated (Fig. 5). The updated model includes additional relevant and important symptoms reported during the concept elicitation portion of the interviews that were not included in the original model. These included patient-reported symptoms such as fever, constipation, headache, cold sweats/sweating, mucus in stool, and throat tightness; and clinician-reported symptoms such as oral ulcers, vision issues, and vitamin absorption issues. Tenesmus, joint stiffness, and skin rashes, reported by both patients and clinicians, were also added. There was more of a disparity between patient- and clinician-reported HRQoL impacts. Patient-reported impacts added to the model included ADLs such as reluctance to leave the house and difficulty driving, as well as emotional, physical, and work/school impacts not reported by physicians (e.g., stress, difficulty sitting, and difficulty completing work tasks).

The preliminary 19-item daily symptoms diary was based on the literature review and the initial conceptual model. Development of the final version of the Crohn’s Disease Diary was in accordance with industry guidelines and includes items generated from the target population [9, 13]. Following three rounds of combined concept elicitation and cognitive debriefing interviews, a total of 17 items (using patient-friendly language) and 2 instructions (highlighting the recall period of the subsequent items) were retained in the final version of the Crohn’s Disease Diary (Table 2). Response options are item specific, including numerical frequency rating scales, severity scales, yes/no options for presence/absence of a symptom, or “select all that apply.” The first 11 items assess core symptoms and impacts of Crohn’s disease with a 24-h recall period, including bowel movement frequency, loose or liquid stool, blood in stool, volume of blood, urgency, tenesmus, abdominal pain, bloating, nausea, fatigue, and sleep disturbance. The remaining items have a 7-day recall period and pertain to extraintestinal manifestations, including items to assess the presence and severity of joint pain and stiffness in different locations.

Most patients (86%–100%) demonstrated an understanding of the instructions provided for completing the Crohn’s Disease Diary. In general, patients reported that the items retained following 3 rounds of interviews were easy to understand. When asked if they would be able to complete the Crohn’s Disease Diary daily for one year, 23/27 (85.2%) reported that this would be easy to do and that it would be possible to recall their symptoms over the 24-h and 7-day recall periods. Overall, the electronic format of the diary was deemed acceptable.

Revisions were made iteratively to items in the Crohn’s Disease Diary following each round of interviews, with consideration of patient understanding, conceptual overlap among items, and the translatability of items (following the third round of interviews). An example of a change made to improve understanding was to clearly define “loose or liquid stools.” Due to conceptual overlap of terms, the abdominal pain item was modified to clarify that “abdominal pain” included cramping. Item relevance was also considered after each round of interviews. For example, although few patients reported blood in bowel movements or tenesmus within a given 24-h recall period, these items were retained since most patients experience these symptoms over a longer period of time (i.e., during a clinical trial in which a PRO measure is used). The recall period for the item assessing sleep disturbance due to needing a bowel movement was updated to “in the past 24 h” from “during the night” to be inclusive of alternate work/sleep patterns (as with shift workers), but no other changes were made to recall periods as they were well understood by patients and considered to be appropriate. A nausea item was added following a patient suggestion in the first round of interviews, and an item pertaining to passing gas was removed from the diary due to patient feedback and concern about the feasibility of patients accurately recalling the frequency/severity of this symptom. Clinicians also suggested a lack of clinical importance in assessing gas. Additionally, items relating to difficulty sleeping and general body pain were removed due to redundancy and ambiguity, respectively.