The economic burden of depression substantially impacts health and social services. Estimating system-level costs is crucial for planning resource use, as well as designing healthcare systems and services aligned with value-based healthcare principles. However, translating these insights into effective and patient-centred care plans requires an understanding of individual-level total costs, distributions and key predictors before and after a depression diagnosis. This observational cohort study used data from 17 Finnish health and social care and benefits registers during 2014–2019 to investigate resource use in patients aged 18–79 years with a depression diagnosis recorded by specialised psychiatric care in the Helsinki-Uusimaa region in 2015. The total costs pre- and post-diagnosis were estimated for each patient using two generalised linear mixed models, assessing the impact of gender, age, socioeconomic status and number of psychiatric comorbidities. This study included 9926 patients. Following a depression diagnosis, total annual costs per patient more than doubled compared with the previous year (€18,789 versus €8575). Post-diagnosis, the total annual costs per patient decreased incrementally each year; by the end of the 5-year follow-up, total costs were 39
Decision-analytic models are widely used to evaluate specialty disease-modifying antirheumatic drugs in rheumatoid arthritis, but approaches to modeling uncertainty, validation, and transparency vary. We aimed to synthesize the modeling approaches of published economic evaluations of specialty disease-modifying antirheumatic drugs for adults with rheumatoid arthritis after failure or intolerance of conventional therapies. We searched EMBASE, MEDLINE, and EconLit (up to 30 June 2026) for model-based full economic evaluations published in 2010 or later. Two reviewers screened studies and abstracted model conceptualization, scope, structure, and assumptions; model uncertainty and value-of-information analyses; model validation; and model transparency and potential conflicts of interest. Forty-four studies met inclusion criteria. State-transition models predominated (22 cohort Markov models and 17 microsimulations), with fewer hybrid decision tree plus Markov models (2) and discrete-event simulations (3). Most studies used payer or health-sector perspectives (31/44) and lifetime horizons (35/44). One-way sensitivity analysis (41/44) and probabilistic sensitivity analysis (40/44) were commonly reported, whereas value-of-information analysis was rare (2/44). Validation reporting largely reflected face validity; technical verification was reported in 5/44 studies and predictive validation in 1/44. No study provided a publicly available executable model, and reporting of data and input accessibility varied. Rheumatoid arthritis economic models commonly use state-transition frameworks and functional disability measures to project lifetime costs and health outcomes, but show persistent gaps in uncertainty analysis, validation, and reproducibility. Future models should more systematically test structural and extrapolation assumptions, strengthen validation, and improve documentation and data sharing to support policy and coverage decisions.
The Pediatric Quality of Life Inventory (PedsQL) is a widely used, generic, health-related quality of life instrument validated for children aged 2–18 years. However, because it is not preference weighted, it cannot be directly used to estimate health utilities for economic evaluations. This study constructed a value set for the PedsQL: the PedsUtil scoring system. Preferences for child health states were elicited from US adults using an online discrete choice experiment (n = 2518) and composite time trade-off survey (n = 868). Child health states were defined by the seven-dimension PedsUtil health state classification system previously derived from the PedsQL. Discrete choice experiment data were analyzed using a mixed logistic regression and the latent coefficients were anchored onto the quality-adjusted life year scale using composite time trade-off data with a linear mapping approach. Preferences varied with age. Pain was the most important dimension for ages 2–4 years and Social Functioning was most important for ages 5–18 years. Pairwise comparisons confirmed significant preference differences between the youngest and older age groups and these age-related differences were also reflected in composite time trade-off data. Consequently, age-specific scoring systems (2–4 and 5–18 years) and an age-combined scoring system (2–18 years) were developed. Utility ranges for the final scoring systems were 0.368–0.999 for ages 2–4 years, 0.427–0.951 for ages 5–18 years, and 0.410–0.972 for ages 2–18 years combined. The PedsUtil scoring system is the first value set derived for the PedsQL, enabling direct estimation of health utilities from PedsQL responses and facilitating pediatric economic evaluations.
