Adjustment of Incidence Rates after an Estimate of Completeness and Accuracy in Registration of Acute Leukemias in a Swedish Population

• Published in: Leukemia & lymphoma Vol. 41; no. 5-6; pp. 559 - 570
• Main Authors: Åström, Maria, Bodin, Lennart, Tidefelt, Ulf
• Format: Journal Article
• Language: English
• Published: United States Informa UK Ltd 2001; Taylor & Francis
• Subjects: Acute Disease; acute leukemia; Adolescent; Adult; Age Factors; Aged; ALL; AML; Data Interpretation, Statistical; Female; Humans; Incidence; Leukemia - epidemiology; Male; Middle Aged; Neoplasms, Second Primary - epidemiology; registration; Registries - statistics & numerical data; selection bias; Survival Rate; Sweden - epidemiology; Sweden
• ISSN: 1042-8194; 1029-2403
• DOI: 10.3109/10428190109060346

Earlier studies have revealed undernotification of hematological malignancies in Swedish and other Cancer Registries. We present epidemiological data on AML, ALL and unspecified AL in adults diagnosed 1987-1992 in a well-defined population. Blast crises of CML were excluded. The Swedish Cancer Registry and Cause of Death Registry were compared and patient records reviewed for validation. When available, listings of pathology bone marrow reports and inpatient discharge diagnoses were utilized for casefinding. 260 cases of acute leukemias could be verified in a population of 663,135 adults, corresponding to a yearly incidence of 6.5/100,000. The median age of the patients was 69.2 years. 214 cases were AML, 38 ALL and eight unspecified AL. Undernotification in the Cancer Registry was found to be 15.4%, greater for AML and unspecified AL than for ALL. In addition the coding was not uniform, resulting in an incidence rate in adults of 5.3/100,000 for the Cancer Registry which is 18.5% lower than that of our study. A significant survival advantage was seen for notified patients. Combination of the Cancer Registry and Cause of Death Registry gave acceptable coverage, omitting only four patients. As the incidence of acute leukemias in our study is comparatively high, we hypothesize that underestimation of incidence and overestimation of survival are general problems for cancer registries.