RESUMO
BACKGROUND AND METHODS: Using a simple model, this retrospective study evaluated the cost-effectiveness of different diagnostic strategies used for pretherapeutic detection of patients with disseminated or locally nonresectable upper gastrointestinal tract malignancies (UGIM). Of 162 consecutive UGIM patients referred for treatment, 73 (45%) had disseminated or locally nonresectable disease, and these patients were eligible for evaluation. RESULTS: The noninvasive diagnostic strategies (computed tomography [CT] with ultrasonography [US] and endoscopic ultrasonography [EUS]) had a low procedure cost, but a diagnostic strategy based on CT with US or CT with US and laparoscopy was not cost-effective. The inclusion of endoscopic or laparoscopic ultrasonography seemed necessary to the provision of a cost-effective strategy because both techniques had a high diagnostic accuracy combined with a low cost. A change in diagnostic strategy from CT with US to CT with US and EUS resulted in a net saving regarding the cost of each additional nonresectable patient detected, but this strategy still required up to 20% futile explorative laparotomies. CONCLUSIONS: The combination of endoscopic and laparoscopic ultrasonography was cost-effective and had no complications in this study. We use this strategy as our standard in the pretherapeutic evaluation of UGIM patients.
Assuntos
Análise Custo-Benefício , Endossonografia/economia , Neoplasias Gastrointestinais/economia , Laparoscopia/economia , Tomografia Computadorizada por Raios X/economia , Adulto , Idoso , Idoso de 80 Anos ou mais , Feminino , Neoplasias Gastrointestinais/diagnóstico , Neoplasias Gastrointestinais/cirurgia , Humanos , Laparotomia , Masculino , Pessoa de Meia-Idade , Cuidados Paliativos , Reprodutibilidade dos Testes , Estudos RetrospectivosRESUMO
This note discusses the adjustment of cost measurement to account for learning. By learning is meant improvements in productive efficiency resulting from use of a (health) technology in routine practice. A recently proposed method is shown to be potentially misleading. Alternatively, it is suggested that the total cost is decomposed in treatment cost and learning cost. Furthermore, if there is uncertainty about the long-run unit cost, learning will reveal the true cost. A method to adjust the learning cost for the value of this information is illustrated.