Value in Health最新文献

Objectives: Surgical decisions for ulcerative colitis are complex and preference-sensitive. This study aimed to assess patient preferences for surgical treatments, quantify preference heterogeneity, and examine individual-specific preferences to inform decision-making.

Methods: Patient preferences were elicited using a discrete choice experiment (DCE). A rigorous selection process involving focus groups and interviews with clinicians and patients resulted in seven key attributes. Each task included two unlabelled surgical alternatives and a medication opt-out. A D-efficient fractional factorial design was generated. The survey was pilot tested using 'think-aloud' interviews. Data were analysed using multinomial and mixed logit models, with conditional mean coefficients used to estimate individual-specific choice probabilities.

Results: Three hundred and fifty patients completed the survey. Results showed significant preference heterogeneity for most attributes. The preference for the medication opt-out revealed a multimodal conditional distribution, clustering patients who strongly preferred, were indifferent to, or disliked medication. The interaction term 'planning to have children' fully explained the preference heterogeneity in the fertility attribute. A gender interaction term showed that male patients had a stronger negative preference for a stoma. Choice probabilities showed individual differences; some patients had a 97.98% probability of preferring medication, while others had only a 0.09% probability, instead showing a high preference for surgical options.

Conclusions: This study demonstrates the value of using conditional distributions to examine preference heterogeneity. Simpler models failed to reveal the wide range of preferences present in the data. Conditional choice probabilities can be used to better understand how different patients make treatment decisions.

Objectives: All EuroQol instruments include a visual analogue scale (VAS), except the EQ-HWB, which captures broader aspects of wellbeing beyond health, relevant to populations providing and receiving care. This study explored the performance of a potential VAS for EQ-HWB.

Methods: The EQ-HWB VAS was developed by adapting the EQ VAS from the EQ-5D-5L, replacing "health" with "health and wellbeing" and "today" with "last week" to align the recall period with EQ-HWB. A cross-sectional survey was undertaken with an age- and sex-representative sample of the Hong Kong general population (n=1,262). Descriptive characteristics, ceiling and floor, convergent and known-group validity, and explanatory power were assessed. The performance of the EQ-HWB VAS was directly compared with the EQ VAS.

Results: The mean EQ VAS and EQ-HWB VAS scores were 76.7 and 73.5 (p<0.001). The ceiling was 5.2% for the EQ VAS and 3.6% for EQ-HWB VAS. The EQ-HWB VAS showed a very strong correlation and agreement with EQ VAS. Both the EQ VAS and EQ-HWB VAS were able to distinguish between most known groups. The EQ-HWB VAS showed larger effect sizes for groups such as caregiving, perceived income level and mental health problems. EQ-HWB-9 items explained more variance in EQ-HWB VAS scores than EQ-5D-5L dimensions did in EQ VAS scores.

Conclusion: This study provides the first evidence of the measurement properties of a "health and wellbeing" VAS for EQ-HWB, demonstrating its validity. Further research is recommended in different languages and countries to inform decisions about the addition of a VAS to the EQ-HWB.

Objective: The Dementia Village (DV) model is emerging as an alternative to traditional nursing homes (NHs) for individuals with Alzheimer's disease and related dementias. This study evaluates the impact of the French Dementia Village, Village Landais Alzheimer Henri Emmanuelli (VLAHE), on healthcare utilisation among older adults with dementia.

Methods: This study employs an indirect comparative cohort design using data from two sources: a longitudinal observational study of VLAHE residents and an ad hoc cohort from the French Nationwide claims database of NH residents. To improve comparability, the Matching-Adjusted Indirect Comparison (MAIC) method was applied. Weighted statistical tests and logistic regression models were used, yielding odds ratios with 95% confidence intervals. The outcomes analysed include hospital admission rates, hospital stay duration, mortality, and place of death over a two-year follow-up period between January 2021-December 2022.

Results: VLAHE residents had fewer hospitalisations and shorter stays (10.5 vs. 15.2 days, p = 0.0004) than comparable NH residents, particularly for acute care. Quarterly hospitalisation rates were consistently lower (6.4% vs. 11.1%), while no significant differences were found for psychiatric or rehabilitation stays. Fewer VLAHE residents died in hospital (9% vs. 31%), suggesting improved end-of-life care (OR: 5.75, 95% CI [1.75-18.90]).

Conclusions: The VLAHE model may reduce hospitalisations and enhance end-of-life care through a structured, patient-centred approach. However, due to a potential residual clinical differences between groups and limitations of the matching method, further longitudinal research is needed to confirm its effectiveness.