RESUMO
PURPOSE: The minimal important change (MIC) in a patient-reported outcome measure is often estimated using patient-reported transition ratings as anchor. However, transition ratings are often more heavily weighted by the follow-up state than by the baseline state, a phenomenon known as "present state bias" (PSB). It is unknown if and how PSB affects the estimation of MICs using various methods. METHODS: We simulated 3240 samples in which the true MIC was simulated as the mean of individual MICs, and PSB was created by basing transition ratings on a "weighted change", differentially weighting baseline and follow-up states. In each sample we estimated MICs based on the following methods: mean change (MC), receiver operating characteristic (ROC) analysis, predictive modeling (PM), adjusted predictive modeling (APM), longitudinal item response theory (LIRT), and longitudinal confirmatory factor analysis (LCFA). The latter two MICs were estimated with and without constraints on the transition item slope parameters (LIRT) or factor loadings (LCFA). RESULTS: PSB did not affect MIC estimates based on MC, ROC, and PM but these methods were biased by other factors. PSB caused imprecision in the MIC estimates based on APM, LIRT and LCFA with constraints, if the degree of PSB was substantial. However, the unconstrained LIRT- and LCFA-based MICs recovered the true MIC without bias and with high precision, independent of the degree of PSB. CONCLUSION: We recommend the unconstrained LIRT- and LCFA-based MIC methods to estimate anchor-based MICs, irrespective of the degree of PSB. The APM-method is a feasible alternative if PSB is limited.
RESUMO
PURPOSE: The minimal important change (MIC) is defined as the smallest within-individual change in a patient-reported outcome measure (PROM) that patients on average perceive as important. We describe a method to estimate this value based on longitudinal confirmatory factor analysis (LCFA). The method is evaluated and compared with a recently published method based on longitudinal item response theory (LIRT) in simulated and real data. We also examined the effect of sample size on bias and precision of the estimate. METHODS: We simulated 108 samples with various characteristics in which the true MIC was simulated as the mean of individual MICs, and estimated MICs based on LCFA and LIRT. Additionally, both MICs were estimated in existing PROMIS Pain Behavior data from 909 patients. In another set of 3888 simulated samples with sample sizes of 125, 250, 500, and 1000, we estimated LCFA-based MICs. RESULTS: The MIC was equally well recovered with the LCFA-method as using the LIRT-method, but the LCFA analyses were more than 50 times faster. In the Pain Behavior data (with higher scores indicating more pain behavior), an LCFA-based MIC for improvement was estimated to be 2.85 points (on a simple sum scale ranging 14-42), whereas the LIRT-based MIC was estimated to be 2.60. The sample size simulations showed that smaller sample sizes decreased the precision of the LCFA-based MIC and increased the risk of model non-convergence. CONCLUSION: The MIC can accurately be estimated using LCFA, but sample sizes need to be preferably greater than 125.
Assuntos
Medidas de Resultados Relatados pelo Paciente , Qualidade de Vida , Humanos , Qualidade de Vida/psicologia , DorRESUMO
BACKGROUND: Breast cancer is one of the most common cancers in women. Patient-reported outcome measures are used to evaluate patients' health-related quality of life in clinical breast cancer studies. This study evaluated the structure, validity, reliability, and responsiveness of the National Comprehensive Cancer Network-Functional Assessment of Cancer Therapy-Breast Cancer Symptom Index (NFBSI-16) subscales in a clinical trial featuring patients with advanced/metastatic breast cancer (aBC), and estimated NFBSI-16 meaningful change thresholds. METHODS: Data from 101 patients with aBC enrolled in a phase II trial (Xenera-1) were included for psychometric evaluation of the NFBSI-16. Subscale structure was evaluated by assessing inter-item correlations, item-total correlations, and internal consistency (cycles 2 and 5). Validity was assessed using scale-level convergent validity (cycles 2 and 5) and known-groups (Baseline). Reliability was analysed via test-retest at cycles 3-4, and responsiveness to improvement and worsening was evaluated at cycles 5, 7, and 9. Meaningful change thresholds were estimated using anchor-based methods (supported by distribution-based methods) at cycles 5, 7, and 9. RESULTS: NFBSI-16 internal consistency was acceptable, but item-total correlations suggested that its subscales and the GP5 item (side-effect of treatment) scores may be preferred over a total score. Convergent and known-groups evidence supported NFBSI-16 validity. Test-retest reliability was good to excellent for Total and DRS-P (disease-related symptoms: physical) scales, and moderate for the GP5 item. Responsiveness to worsening was generally demonstrated, but responsiveness to improvement could not be demonstrated due to limited observed improvement. Anchor-based meaningful change thresholds were estimated for DRS-P and Total scores. CONCLUSION: This study provides evidence that the NFBSI-16 has desirable psychometric properties for use in clinical studies in aBC. It also provides estimates of group- and individual-level meaningful change thresholds to facilitate score interpretation in future aBC research.