S86

VA L U E I N H E A LT H 2 1 ( 2 0 1 8 ) S 1 – S 2 6 8

follow-up period upon PERT initiation. Resource utilization including hospital and Objectives: Severity and frequency of symptoms are critical endpoints in clinical
emergency room (ER) visits were compared across matched adherent (defined by trials for celiac disease (CD) in order to assess response to therapy. Valid and reli-
Proportion of Days Covered (PDC) =  > 80%) vs. non-adherent cohorts. All-cause and able instruments are therefore needed to assess symptoms as a key, and frequently
disease-specific healthcare costs were compared for adherent vs. non-adherent primary, endpoint. This study evaluated evidence regarding validity and reliability
cohorts.  Results: Among 5,358 EPI patients, 91.6% had CP, 32.8% had PC and 35.9% of existing outcome measures and explored key stakeholders’ views on measuring
underwent PY. Mean adherence levels for PERT were found to be at 48%, 52%, and CD symptoms and impacts.  Methods: A targeted literature review identified PRO
52% in CP, PC, and PY patients respectively. Approximately 20% of the patients measures and studies detailing their development or validation for comparison
were shown to be adherent to PERT in one year follow-up period. The majority of against FDA evidentiary standards. Stakeholders in the US, EU, and UK, i.e. patient
patients (~70%) discontinued their PERT medications within the first year of PERT advocates (n= 1), clinicians (n= 7), and payers (n= 13), were then interviewed to under-
initiation. Matched analysis showed that adherent cohort compared to the non- stand which patient-centered aspects of CD are most important to treat, feasible
adherent group incurred lower hospitalization and ER visits. Additionally all-cause to measure as PRO endpoints, and valued for reimbursement.  Results: Several
and disease-specific healthcare costs were also shown to be higher for non-adherent PRO instruments relevant to CD were identified. Two were deemed most suitable
cohort. However, these differences were not statistically significant.  Conclusions: for measuring CD symptoms accurately: the Celiac Disease Symptom Diary (CDSD)
Management of EPI is shown to be suboptimal. Large proportion of patients was and the Celiac Disease Patient Reported Outcome tool (CeD-PRO), though each had
found to be non-adherent to their PERT regimen, and was found to be associated limitations in terms of conceptual coverage, evidence for validity, reliability, and
with higher utilization and healthcare costs. responsiveness, and readiness for implementation in clinical trial settings. In paral-
lel, clinicians and patient advocate interviews showed that GI symptoms are most
PGI25 burdensome, though extra-intestinal symptoms are negatively impactful as well.
Comparing Short Form – 6 Dimension (SF-6D) Derived Utilities and Both the disease and the gluten-free diet (GFD) can impair patients’ quality of life.
Euroqol Questionnaire – 5 Dimensions (EQ-5D) Mapped Utilities Payers emphasized the importance of measuring both symptom frequency and
Among Patients with Iron Deficiency in Inflammatory Bowel severity, particularly for GI symptoms, and the need to better understand and target
Disease (ID-IBD) Treated with Ferric Maltol (FM) Vs Placebo patients who are non-responsive to GFD.  Conclusions: With emerging thera-
Tremblay G1, Forsythe A1, Rabe A1, Jones T2 peutic options in celiac disease (CD) beyond a strict GFD, further work is needed to
1Purple Squirrel Economics, New York, NY, USA, 2Shield Therapeutics, London, UK finalize an appropriate PRO instrument to assess CD symptoms and to operational-
Objectives: We aimed to compare mapped and derived utilities for health states ize an endpoint in clinical trials in a way that is meaningful to patients, valued by
among patients with ID-IBD treated with FM versus placebo. EQ-5D-mapped utilities payers, and acceptable to regulators in demonstrating efficacy.
were found in previous studies on other diseases to estimate higher utilities than
those derived through SF-6D.  Methods: Health state utilities were taken from PGI28
AEGIS-1 and AEGIS-2, two trials examining FM in the treatment of IDA-IBD (ulcera- The American Neurogastroenterology and Motility Society
tive colitis and Crohn’s disease, respectively), where outcomes were measured over Gastroparesis Cardinal Symptom Index-Daily Diary (ANMS GCSI-DD):
the first 12 weeks in a randomised double-blind trial, with a 52-week open-label Assessing the Content Validity in Patients with Idiopathic or
extension. The SF-36 questionnaire was used to determine general quality of life. Diabetic Gastroparesis
Attribution of scores to specific health states was based on haemoglobin levels Revicki DA1, Gleeson S1, Speck R1, Puelles J2, Kuo B3, Camilleri M4, Parkman HP5
collected concurrently with each assessment. With haemoglobin level (g/dL) as a 1Evidera, Bethesda, MD, USA, 2Takeda Pharmaceuticals, London, UK, 3Massachusetts General

surrogate for ID, the health states used were moderate (8.0-10.9), mild (11.0-12.9), Hospital, Boston, MA, USA, 4Mayo Clinic, Rochester, MN, USA, 5Temple University, Philadelphia,
and stable (> 12.9) ID. SF-36 scores were converted to EQ-5D-mapped utilities, while PA, USA
the same scores were used to derive SF-6D utilities using the Ara and Brazier algo- Objectives: Gastroparesis is a condition involving delayed gastric emptying when
rithms, respectively. These scores were then compared across the different health no mechanical obstruction is present. The ANMS GCSI-DD was developed to meet
states. A random-effect regression model was used in both algorithms to account FDA recommendations for use as Patient Reported Outcome (PRO) endpoints in
for haemoglobin at baseline, ferritin level, and crossover in the open-label extension gastroparesis studies, including therapeutic trials. The purpose of this study was
phase.  Results: 443 observations were available for analysis. Utilities favoured to determine if the appropriate symptoms are included in the instrument and to
ferric maltol over placebo, and improved as haemoglobin level increased from assess the content validity of the revised ANMS GCSI-DD in idiopathic (IG) and
moderate to mild to stable ID in both placebo and FM groups (Utilities for moder- diabetic gastroparesis (DG) patients. Methods: Patients diagnosed with IG or
ate, mild and stable health states: EQ-5D in placebo 0.784, 0.787, 0.808 vs FM 0.822, DG were recruited by five clinical sites in the United States for a cross-sectional,
0.825, 0.847; SF-6D in placebo 0.732, 0.738, 0.750 vs FM 0.750, 0.757, 0.768). EQ-5D qualitative study involving one-on-one in-person concept elicitation and cognitive
mapped utilities were higher than SF-6D derived utilities in all health states for debriefing interviews. Concept elicitation included open-ended questions to elicit
both groups.  Conclusions: Among patients with ID-IBD, utilities increased with patients’ symptoms and impacts of gastroparesis, while cognitive debriefing was
haemoglobin levels, with greater increases with ferric maltol than with placebo. designed to assess the comprehensiveness of the ANMS GCSI-DD and clarity of
EQ-5D-mapped utilities were higher than SF-6D-derived utilities. the instructions, items, and response scales. The interviews were audio-recorded
and transcribed. Transcripts were analyzed using a content analysis approach with
PGI26 ATLAS.ti software.  Results: Of 25 patients interviewed, 15 (60%) had IG and 10
A Review of Patient-Reported Outcomes in Nonalcoholic (40%) DG. Consistent with other gastroparesis studies, mean age was 42.3 years
Steatohepatitis (NASH) (range: 20-70 years), and most patients were female (n= 19, 76%) and Caucasian
Trivedi B, Cheng R (n= 19, 76%). Patients endorsed the following signs and symptoms as relevant and
ZS Associates, Thousand Oaks, CA, USA important to their condition: early satiety (n= 25, 100%), post-prandial fullness (n= 25,
Objectives: Nonalcoholic steatohepatitis (NASH) is characterized by fat build 100%), nausea (n= 22, 88%), upper abdominal pain (n= 18, 72%), vomiting (n= 15, 60%),
up in the liver and results in inflammation and liver cell damage. NASH is and bloating (n= 11, 44%). Patients confirmed that the ANMS GCSI-DD content was
rapidly becoming one of the top causes of cirrhosis, hepatocellular carcinoma and comprehensive and reflective of their gastroparesis experience. Overall, patients
indication for liver transplantation. Patient-reported outcome (PRO) measures are found the instrument’s instructions, recall period, items, and response options clear
an important component to assessing disease impact and therapy response in and easy to understand.  Conclusions: Based on patient feedback, no changes to
patients with NASH. The objective of this research is to identify commonly used the ANMS GCSI-DD were recommended. The study results provide evidence sup-
PROs in NASH, evaluate their development, and assess their utility in various types porting the content validity of the ANMS GCSI-DD for clinical trials and clinical care
of studies.  Methods: A literature review using PubMed and a search on clinical- among IG or DG patients.
trials.gov was conducted to identify PROs used in NASH studies. These PROs were
evaluated against the FDA’s PRO Guidance criteria with regard to: (1) content valid- PGI29
ity, particularly patient input in concept elicitation, item generation, and cognitive Burden of Chronic Hepatitis C (CHC) from the Patient Perspective:
debriefing; and (2) psychometric testing, particularly construct validity, internal A Literature Review and Conceptual Map
consistency reliability, test-retest reliability, and responsiveness. Each criteria was Nwankwo C1, Corman SL2
assessed and given a rating. Finally, an attempt was made to make recommenda- 1Merck
& Co., Inc., Kenilworth, NJ, USA, 2Pharmerit International, Bethesda, MD, USA
tions on which PROs should be used in different types of studies.  Results: Only Objectives: The burden of CHC is typically expressed in terms of its impact
one NASH-specific PRO publication currently exists: the NASH-CHECK. The NASH- on the healthcare system and to society. The aims of this literature review and
CHECK has only gone through qualitative validation and needs further psychometric assessment were to characterize the burden of CHC from the patient perspective
analyses. Other PROs used in NASH are generic liver-related measures such as the and to create a conceptual map of factors that influence this.  Methods: A tar-
Chronic Liver Disease Questionnaire, the EORTC QLQ-HCC18 (EORTC Quality of Life geted Medline search was conducted for the years 2010-present using the terms
Questionnaire - Hepatocellular Carcinoma / Primary Liver Cancer Module), and the hepatitis C, burden, quality of life, symptoms, employment, and costs. Prospective,
LDQOL | Liver Disease Quality of Life Questionnaire. Generic VAS scales have also retrospective, and cross-sectional studies that reported on the humanistic and/
been used to measure condition- specific symptoms such as itching. Other PRO or economic burden of CHC from the patient’s perspective were included. Upon
measures have been appropriately validated but not in specifically in the NASH identifying and summarizing key themes and supporting data, a conceptual map
population.  Conclusions: Most PROs used in NASH are borrowed from other of factors that contribute either positively or negatively to patient burden was
diseases or are general PROs and do not address the specific symptoms or concerns constructed.  Results: Themes consisting of an interplay of humanistic burden
of NASH patients. As such, there is need for further NASH-specific PRO development. (consisting of emotional and psychosocial impact and health-related quality of life
(HRQoL)) economic burden (including work productivity, associated indirect costs,
PGI27 and out-of-pocket (OOP) costs) and treatment impact were identified. Resulting
Concepts and Instruments for Patient-Reported (PRO) Assessment in in a conceptual map that indicates that stigmatization can lead to social iso-
Celiac Disease: Literature Review and Stakeholder Perspectives lation, symptoms of depression and anxiety, social isolation, and poor HRQoL,
Clifford S1, Taylor A2, Gerber M3, Devine J1, Cho M1, Walker R4, Stefani I4, Fidel S4, which in turn impact work productivity and willingness to access care. Impaired
Drahos J3, Leffler D3 work productivity leads to indirect costs, which can further worsen already poor
1ICON Plc, South San Francisco, CA, USA, 2Takeda Development Centre Europe, London, UK, socioeconomic status. Higher OOP costs in patients who are already economi-
3Takeda Pharmeuticals International, Cambridge, MA, USA, 4ICON plc, LONDON, UK cally disadvantaged not only further worsens their economic situation but can