Vortragssitzung

Messung und Bewertung von Gesundheit I

Talks

Mapping from SIBDQ to EQ-5D-5L for patients with inflammatory bowel disease
Isa Maria Steiner, Hamburg Center for Health Economics, Universität Hamburg

Einleitung / Introduction

Clinical studies commonly use disease-specific measures to assess patients’ health-related quality of life. However, economic evaluation often requires preference-based utility index scores to calculate cost per quality-adjusted life-year (QALY). When utility index scores are not directly available, mappings are useful. To our knowledge, no mapping exists for the Short Inflammatory Bowel Disease Questionnaire (SIBDQ). Our aim was to develop a mapping from SIBDQ to the EuroQoL 5-Dimension 5-Level (EQ-5D-5L) index score for inflammatory bowel disease (IBD) patients.

Methode / Method

We used baseline observations of 1070 IBD patients who participated in a randomised controlled trial in Germany on the effect of introducing regular appointments with an IBD nurse specialist in addition to standard care with biologics. To do so, we applied regression and machine learning models for different sets of covariates, thereby taking differences in data availability into account. For each set of covariates, we estimated the following models: ordinary least squares, Tobit regression, an adjusted limited dependent variable mixture model and a regression forest. We selected the final models based on 10-fold cross-validation and validated these with observations from six months after baseline. To facilitate the correct application of the mapping, we embedded the final models in an online tool using R Shiny.

Ergebnisse / Results

Among all estimated models, the regression forest performed best. The external validation led to similar performance measures as those obtained from 10-fold cross-validation. Our findings suggest that the demographic variables age and gender do not improve the mapping, while including SIBDQ subscales, IBD disease type, BMI and smoking status leads to better predictions.

Zusammenfassung / Conclusion

We developed an algorithm mapping SIBDQ values to EQ-5D-5L index scores for different sets of covariates in IBD patients. The algorithm was embedded in an online tool to allow for its user-friendly application.


Authors
Isa Maria Steiner, Hamburg Center for Health Economics, Universität Hamburg
Bernd Bokemeyer, Interdisziplinäres Crohn Colitis Centrum Minden
Tom Stargardt, Hamburg Center for Health Economics, Universität Hamburg
Using the Online Elicitation of Personal Utility Functions (OPUF) approach to derive a patient-based EQ-5D-5L value set: a study in 122 patients with rheumatic diseases from Germany
Paul Schneider, University of Sheffield, UK / Universität Duisburg/Essen

Einleitung / Introduction

Decision makers increasingly recognise the importance of taking into account patient preferences in health care and policy decision making. There are multiple methods, such as discrete choice experiments or time trade-off, that can be used to elicit patient preferences and to construct health-related quality of life scores. Most of these methods, however, suffer from a major limitation: they are inefficient. Little information is obtained from each patient. As a result, data from hundreds if not thousands of participants is required to derive health related quality of life scores. This severely limits the applicability of these methods in patient populations, where recruiting a sufficient number of participants can be challenging. The objective of this study was to test a new method, called Online elicitation of Personal Utility Functions (OPUF), to derive an EQ-5D-5L value set from a relatively small sample of patients with rheumatic diseases from Germany.

Methode / Method

OPUF is a new type of online survey that combines mutliple compositional preference elicitation techniques. Central to the method are three valuation steps: (1) dimension weighting, (2) level rating, and (3) anchoring. From the responses, a personal EQ-5D-5L utility function can be constructed for each participant, and a group-level value set can be derived by aggregating model coefficients across participants. A German demo version of the OPUF survey can be accessed at https://valorem.health/eq5d5l_de (an English version is available at: https://valorem.health/eq5d5l)

Ergebnisse / Results

A total of 122 rheumatic disease patients from Germany completed the OPUF survey. Participants reported various health conditions, including rheumatoid arthritis, psoriasis arthritis, and ankylosing spondylitis, and were mostly female (n=111;91%). Our results suggest that the survey was generally well received; most participants completed the survey in less than 20 minutes. We were able to construct a personal utility function for each participant, and on the aggregate group level, we derived a plausible, logically consistent EQ-5D-5L value set. The precision of mean estimates was comparable to other valuation studies with much larger sample sizes. Resulting utility values ranged from 1 (for state 11111) to -0.32 (for 55555).

Zusammenfassung / Conclusion

Our findings demonstrate that OPUF can be used to derive an EQ-5D-5L value set from a relatively small sample of patients. Even though the method is still under development, we think that it has the potential to be a valuable tool and to complement traditional preference elicitation methods, that may enable decision makers to consider patient preferences more regularly and more thoroughly when making health policy decisions.


Authors
Katharina Blankart, Universität Duisburg/Essen
John Brazier, University of Sheffield, UK
Nancy Devlin, University of Melbourne, Australia
Unstratified and patient-specific meaningful improvement thresholds for EQ-5D index scores in patients after total hip or knee replacement: a retrospective observational cohort study
David Kuklinski, Universität St. Gallen

Einleitung / Introduction

The ultimate goal of patients undergoing total hip and knee replacements (THR and TKR) is to improve health-related quality of life (HRQoL). Patient-reported outcome measures (PROMs) can quantify and monitor HRQoL changes. Usually average PROM values are used to evaluate treatment outcomes. The aim was to calculate HRQoL thresholds, which take individual patient characteristics into account to achieve a realistic evaluation of treatment success.

Methode / Method

This is a retrospective observational cohort study. Patient-level data from 5,217 THR and TKR patients of two German hospitals were used. Anchor-based receiver operator characteristics curves were applied to estimate unstratified and patient-specific meaningful improvement (MI) thresholds for the EQ-5D-5L index score at 3- and 12-month follow-up.

Ergebnisse / Results

Improvements in EQ-5D-5L index score after surgery were dependent on individual patient characteristics. Unstratified 12-months MI thresholds differed between anchors (mobility improvement: 0.264 and 0.211; pain reduction: 0.204 and 0.203) for patients after THR and TKR, respectively. MI thresholds for both anchors and treatments were more accurate, and detected critical recovery paths better when calculated for patient subgroups. Patients with worse pre-operative EQ-5D-5L index scores had higher MI, vice versa.

Zusammenfassung / Conclusion

Calculation of MI thresholds for the EQ-5D-5L index score are time-window specific, vary with the choice of anchor and must be based on patient characteristics to allow for a more meaningful treatment outcome evaluation. Moreover, to set an appropriate anchor, patient expectations from surgery should be considered. Our patient-specific MI thresholds can be used to better compare hospitals for patient-centric quality, thus deal as input for potential pay-for-performance contracts or quality transparency platforms.


Authors
David Kuklinski, Universität St. Gallen
Carlos Marques, MSH Medical School Hamburg
Karina Bohlen, Schoen Klinik Hamburg Eilbek
Karl Westphal, Schoen Klinik Neustadt
Frank Lampe, Asklepios Kliniken Hamburg
Alexander Geissler, Universität St. Gallen
Validation of the Recovering Quality of Life (ReQoL) questionnaires for patients with neurotic, stress-related, somatoform and personality disorders in Germany
Thomas Grochtdreis, Universitätsklinikum Hamburg-Eppendorf

Einleitung / Introduction

In mental health economic research, it was recognized that generic health-related quality of life (HrQoL) measures might not be responsive to recovery. The HrQoL and the recovery process of persons with mental health conditions can be assessed with the Recovering Quality of Life questionnaires (ReQoL-20 and ReQoL-10). The aim of this study was to assess the psychometric properties of the German version of the ReQoL measures in patients with neurotic, stress-related and somatoform disorders and in patients with disorders of adult personality and behavior.

Methode / Method

This study was based on a sub-sample of patients that were recruited within a randomized controlled trial evaluating an evidence-based, stepped and coordinated care service model for mental disorders (RECOVER). The internal consistency was assessed using Cronbach’s Alpha and the test-retest reliability was assessed by the intra-class correlation coefficient. Concurrent validity was assessed by Pearson’s Correlation Coefficient and the known-group validity was assessed using the Cohen’s d. Discriminative abilities were assessed by constructing receiver operating characteristic (ROC) curves.

Ergebnisse / Results

The internal consistency of the ReQoL-20 and ReQoL-10 was excellent and good with Cronbach’s alpha of α = 0.91 and α = 0.83 for all items. The test-retest reliability of the ReQoL measures was moderate with ICC of r = 0.72 to 0.74. The concurrent validity of the ReQoL measures with the Global Severity Index (GSI) was overall high with a correlation coefficient of r = −0.70. The discriminative abilities of the ReQoL-20 and ReQoL-10 for treatment response between time points were excellent with areas under the curve of 0.85 and 0.81, respectively.

Zusammenfassung / Conclusion

For patients with neurotic, stress-related, somatoform and personality disorders, the German version of the ReQoL measures is valid and reliable for the assessment of HrQoL and the recovery process. The ReQoL measures have proven to be sensitive to change in both the GSI and measures of HrQoL, and may be therefore promising for use in mental health economic research.


Authors
Thomas Grochtdreis, Universitätsklinikum Hamburg-Eppendorf
Hans-Helmut König, Universitätsklinikum Hamburg-Eppendorf
Jürgen Gallinat, Universitätsklinikum Hamburg-Eppendorf
Alexander Konnopka, Universitätsklinikum Hamburg-Eppendorf
Holger Schulz, Universitätsklinikum Hamburg-Eppendorf
Martin Lambert, Universitätsklinikum Hamburg-Eppendorf
Anne Karow, Universitätsklinikum Hamburg-Eppendorf
Judith Dams, Universitätsklinikum Hamburg-Eppendorf