Understanding gastroenterologist preferences at the time of treatment escalation to first-line advanced therapies in ulcerative colitis: a discrete choice experiment in five European countries

Introduction: As the number of advanced therapies for moderately to severely active ulcerative colitis (UC) increases, it is necessary to understand the factors driving gastroenterologist (GE) choice in escalating patients (pts) from conventional to advanced therapy.

Objective: To quantify GE therapy attribute preferences when escalating pts to their first advanced UC therapy.

Methodology: We conducted an online, cross-sectional survey using a discrete choice experiment (DCE) design. Attribute and level selection was informed by targeted literature search and formative qualitative research with pts and clinicians. Survey responders were practising GEs experienced in treating pts with moderately to severely active UC, recruited from France, Germany, Italy, Spain and the United Kingdom (UK). Preference weights were estimated using a random parameters logit model for multiple levels of seven attributes: time to symptom improvement, probability of remission at one year, difference between probability of remission and corticosteroid-free remission, five-year risk of malignancy, annual risk of serious infection, annual risk of major adverse cardiovascular events, and mode and frequency of administration. Relative importance (RI) was calculated as the difference in preference weights between the most and least preferred level of each attribute, proportionate to all attribute differences. An extra survey section was added to understand GE treatment and prescribing practices.

Results: A total of 397 GEs were included (France n=140; Germany n=40; Italy n=40; Spain n=47; UK n=130). The most common GE-reported barriers to prescribing advanced therapies were concerns about contraindications and risks/side effects from pts (54.9%) and GEs (39.8%), perceived patient concerns about receiving injections or infusions (35.5%) and concerns about cost or insufficient reimbursement (32.2%). All DCE attributes factored into GE treatment decisions (see Table for RI and preference weights). The three most important attributes were probability of remission at one year (RI 48.4%), five-year risk of malignancy (RI 11.4%) and time to symptom improvement (RI 11.1%; [Table 1]).

Table 1 Preference weights and RI of attributes influencing advanced UC therapy choice (N=397)

	Time to symptom improvement	Probability of remission at one year	Difference between probability of remission and CS-free remission	Five-year risk of malignancy	Annual risk of serious infection	Annual risk of MACE	Mode and frequency of administration
Attribute RI,a% (95% CI)b	11.1 (8.9, 13.4)	48.4 (45.7, 51.1)	8.0 (6.1, 10.0)	11.4 (9.5, 13.2)	6.7 (4.9, 8.5)	6.8 (5.0, 8.6)	7.5 (5.4, 10.1)
Preference weight level (95% CI)c
Level 1	2 weeks 0.50 (0.31, 0.69)	20% probability -2.11 (-2.42, -1.81)	0% difference 0.27 (0.13, 0.41)	1/1000 pts 0.49 (0.35, 0.62)	1/100 pts 0.27 (0.14, 0.40)	1/1000 pts 0.27 (0.14, 0.40)	Oral pill 1–2 times daily with potential dose change 0.15 (-0.02, 0.33)
Level 2	4 weeks 0.22 (0.09, 0.34)	35% probability 0.26 (0.18, 0.34)	5% difference 0.12 (0.03, 0.21)	3/1000 pts -0.04 (-0.13, 0.05)	3/100 pts 0.02 (-0.07, 0.10)	3/1000 pts 0.01 (-0.08, 0.10)	Oral pill 1–2 times daily with the same dose throughout 0.23 (0.10, 0.35)
Level 3	8 weeks -0.31 (-0.44, -0.18)	45% probability 1.85 (1.65, 2.06)	15% difference -0.39 (-0.48, -0.29)	5/1000 pts -0.45 (-0.54, -0.35)	5/100 pts -0.28 (-0.38, -0.19)	5/1000 pts -0.28 (-0.38, -0.19)	Injection every 1–2 weeks 0.01 (-0.12, 0.14)
Level 4	12 weeks -0.41 (-0.55, -0.27)	N/A	N/A	N/A	N/A	N/A	Infusion every 4–8 weeks -0.39 (-0.52, -0.26)

The DCE model included seven attributes, each with several preference weight levels. ^aRI is calculated as the difference in preference weights between the most preferred and least preferred level divided by the sum of the differences across all attributes; estimates sum to 100%. ^b95% CIs that do not include zero indicate a statistically significant RI of an attribute. All seven attributes were statistically significantly important when selecting an advanced therapy. 95% CIs that do not overlap for pairs of attributes indicate a statistically significant difference in importance between attributes. Probability of remission at one year was statistically significantly more important than all other attributes. ^cPreference weight levels are effects coded; zero indicates the mean effect across all attribute levels. CI, confidence interval; CS, corticosteroid; DCE, discrete choice experiment; MACE, major adverse cardiovascular events; N, total number of pts; N/A, not applicable; pts, patients; RI, relative importance; UC, ulcerative colitis.

Conclusion: All attributes factored into the trade-offs GEs consider when escalating pts with moderately to severely active UC to their first advanced therapy. While risk of side effects was the most stated GE barrier to prescribing advanced therapy, probability of remission outweighed all other DCE attributes.

Publication History

Article published online:
04 September 2025