Loading...
Publication
Head and neck cancer: rasch analysis in quality of life assessment
| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| PPG_29406 | 3.47 MB | Adobe PDF |
Authors
Abstract(s)
A patologia oncológica de cabeça e pescoço (POCP) representa 3% de todos os
tumores malignos diagnosticados mundialmente. A POCP inclui várias localizações como:
cavidade oral, faringe, laringe e glândulas salivares. A POCP causa impacto negativo no
doente oncológico e afecta várias funções. A avaliação da Qualidade de Vida Relacionada
com a Saúde (QdVRS) é essencial para melhor compreender e explorar a autopercepção do
doente oncológico de cabeça e pescoço.
Contudo, a QdVRS é considerada um Patient Reported Outcome (PRO) de carácter
subjectivo e multidimensional justificando a necessidade de criar medidas confiáveis, e
adequadas.
O nosso objectivo consistiu em optimizar e melhorar a avaliação da QDVRS com
recurso a modelos matemáticos. Identificaram-se as questões difíceis do EORTC QLQ C-30
e a habilidade do paciente em responder discutindo o conceito de dificuldade em contextos
de saúde.
Foi analisada uma base de dados obtida na Unidade de Cabeça e Pescoço do Instituto
Português de Oncologia do Porto. Usou-se o WinRasch software package- versão de
demonstração.
Os doentes oncológicos de cabeça e pescoço (n=75) foram selecionados no momento
de diagnóstico e agrupados pelo estadiamento da doença. Foi comparado o modelo
conceptual com a análise do Modelo de Rasch (correlação point-measure; dificuldade ou
desafio de acordo com os dados; curvas de probabilidade relacionadas com a variável
latente; infit e outfit).
A análise de Rasch suportou na generalidade o modelo conceptual do EORTC QLQ
C30 nos doentes oncológicos avaliados. Detetaram-se 3 items problemáticos (Q7, Q20, Q28).
Um total de 3 doentes demonstraram resultados fora do esperado baseado na análise.
O Modelo de Rasch aplicado a QdVRS em POCP tem aumentado a precisão e
complementado as técnicas estatísticas tradicionais. Esta abordagem optimiza os PRO e
facilita a sua incorporação na prática clínica em oncologia.
Head and neck cancer (HNC) represents 3% of all cancer cases worldwide. HNC includes several upper body locations: oral cavity, pharynx, larynx or salivary glands. HNC is the most distressing human cancer disturbing cervicofacial region and affecting several functions. Health Related Quality of Life (HRQoL) assessment in HNC is essential to better understand and explore the HNC patient’s perceptions. Nevertheless, HRQoL is considered a subjective and multidimensional dynamic Patient Reported Outcome justifying the need to build reliable, robust and adequate measures. Our objective was to optimize and improve HRQoL assessment using mathematical models identifying the EORTC QLQ-C30 questions difficulty and the patients’ ability to answer discussing the concept of difficulty in health context. A database obtained in Oporto Oncology Portuguese Institute, Head and Neck Unit, was explored. It was used the WinRasch software package - demonstration version. HNC patients (n=75) were selected at the time of diagnosis and grouped according to the stage of disease. We compare our construct theory and Rasch Model analysis (point-measure correlation; difficulty or challenge according to the data; probability curves related to the latent variable; infit and outfit). Rasch analysis supported generally the theoretical constructs of the EORTC QLQ C30 in HNC patients evaluated. RM analyses detected three problematic items (Q7, Q20, Q28). A total of 3 respondents demonstrated misfit pattern at constructs based on Rasch analysis. Rasch Model applied to HRQoL in HNC as improving precision and complementing traditional statistical techniques. This approach is improving PRO and enabling its incorporation in oncological clinical practice.
Head and neck cancer (HNC) represents 3% of all cancer cases worldwide. HNC includes several upper body locations: oral cavity, pharynx, larynx or salivary glands. HNC is the most distressing human cancer disturbing cervicofacial region and affecting several functions. Health Related Quality of Life (HRQoL) assessment in HNC is essential to better understand and explore the HNC patient’s perceptions. Nevertheless, HRQoL is considered a subjective and multidimensional dynamic Patient Reported Outcome justifying the need to build reliable, robust and adequate measures. Our objective was to optimize and improve HRQoL assessment using mathematical models identifying the EORTC QLQ-C30 questions difficulty and the patients’ ability to answer discussing the concept of difficulty in health context. A database obtained in Oporto Oncology Portuguese Institute, Head and Neck Unit, was explored. It was used the WinRasch software package - demonstration version. HNC patients (n=75) were selected at the time of diagnosis and grouped according to the stage of disease. We compare our construct theory and Rasch Model analysis (point-measure correlation; difficulty or challenge according to the data; probability curves related to the latent variable; infit and outfit). Rasch analysis supported generally the theoretical constructs of the EORTC QLQ C30 in HNC patients evaluated. RM analyses detected three problematic items (Q7, Q20, Q28). A total of 3 respondents demonstrated misfit pattern at constructs based on Rasch analysis. Rasch Model applied to HRQoL in HNC as improving precision and complementing traditional statistical techniques. This approach is improving PRO and enabling its incorporation in oncological clinical practice.
Description
Keywords
Oncologia Cancro cabeça e pescoço Patient reported outcomes Modelo de rasch Teoria de resposta ao item Oncology Head and neck cancer Patient reported outcomes Rasch model Item response theory