Hospital Administrative Data to Identify NSCLC: An Algorithm Evaluation Study
Author(s)
Balzi W1, Roncadori A2, Massa I1, Danesi V3, Gentili N1, Crinò L1, Delmonte A1, Altini M4
1Istituto Romagnolo per lo Studio dei tumori " Dino Amadori" IRST-IRCCS, Meldola, FC, Italy, 2Istituto Romagnolo per lo Studio dei tumori " Dino Amadori" IRST-IRCCS, Bologna, BO, Italy, 3Istituto Romagnolo per lo Studio dei tumori " Dino Amadori" IRST-IRCCS, Meldola, Italy, 4AUSL of Romagna, Ravenna, FC, Italy
OBJECTIVES This study aimed to evaluate the performances of an algorithm developed for identifying Non Small Cell Lung Cancer (NSCLC) candidates among lung cancer patients with a diagnosis ICD-9-CM 162.x code in administrative databases. METHODS Algorithm discrimination capacity to correctly identify NSCLC among lung cancer patients was performed comparing algorithm’s results against electronic health records (EHR). Data from administrative databases were collected from a specific area of the Emilia-Romagna region for which EHR diagnosis was available. The main inclusion criteria was ICD-9-CM diagnosis code 162.x during the year 2017. Exclusion criteria included: 1) ICD-9-CM 162.x code recorded in the 3 years before 2) other malignancies. Overall accuracy, Positive (PPV) and Negative (NPV) predictive values, sensitivity and specificity, positive and negative likelihood ratios and diagnostic odds ratio were used to evaluate algorithm’s performances. Associations between possible factors potentially affecting algorithm accuracy were investigated by means of a multinomial logistic regression. RESULTS A total of 430 patients were identified as lung cancer based on ICD-9 diagnosis. Focusing on the total incident cases (N=314), the algorithm had an overall accuracy of 82.8% with a sensitivity of 88.8%. The analysis confirmed a high level of PPV (90.2%), but lower specificity (53.7%) and NPV (50%). Analysing the results of the multinomial logistic regression model, higher length of stay seemed to be associated with the correct classifications among confirmed NSCLC (increased sensibility). Hospitalization regimen and a supply of antiblastic therapy seemed to increase the level of PPV. CONCLUSIONS The algorithm demonstrated a strong validity for identifying NSCLC among lung cancer patients in hospital administrative databases and can be used as a basis for selecting patients for secondary data use studies in a real-world setting.
Conference/Value in Health Info
2021-11, ISPOR Europe 2021, Copenhagen, Denmark
Value in Health, Volume 24, Issue 12, S2 (December 2021)
Code
POSC417
Topic
Epidemiology & Public Health, Real World Data & Information Systems
Topic Subcategory
Disease Classification & Coding, Health & Insurance Records Systems, Reproducibility & Replicability
Disease
Oncology