Under Estimation and Over Estimation in Gastric Cancer Incidence Registry in Khorasan Provinces

Gastric cancer is a disease which the cells forming the inner lining of the stomach start to divideuncontrollably, forming a mass called a tumor. Patients with gastric cancer from low- facilityprovinces like North and South Khorasans may diagnose and registered in full- featuredprovinces like Razavi-Khorasan; this causes misclassification error. The presence of this errormakes the registry systems inaccurate and unreliable for estimating the burden of cancer andpolicy making. Two approaches are recommended for reducing the effects of misclassification;the first is using a small validation sample and the second is a Bayesian analysis which providessubjective prior information for the subset of the parameters to correct the statistics. Data for thisstudy extracted from Iranian annual of national cancer registration report in 2008. The agestandardized rate due to gastric cancer [ICD-10; C16] were expressed as rate per/100,000population for male and female of North, South and Razavi Khorasans. To correct themisclassification effect, a Bayesian approach was used with Poisson count regression and betaprior. The reported expected coverage of cancer incidence for Razavi-Khorasan was 155.5% andexceeds than what expected, whereas the North and South Khorasans have just observedrespectively 34.8% and 41.4% of their expected coverage. The results of the Bayesian analysisindicated that there was about 34% misclassification in gastric cancer incidence registry fromNorth and South Khorasans in Razavi-Khorasan. In planning for resource allocation, authoritiesshould consider that, low incidence of gastric cancer in North and South Khorasans, may be theeffect of misclassification and it is needed to allocate them more health facilities and improve theiraddress registration accuracy using National ID, electric bill, etc.

کلیدواژه ها:

Gastric Cancer ، Misclassification Error ، Bayesian Method ، Incidence ، Cancer Registry