The Danish Myelodysplastic Syndromes Database: Patient Characteristics and Validity of Data Records

Abstract

Background

The Danish Myelodysplastic Syndromes Database (DMDSD) comprises nearly all patients diagnosed with myelodysplastic syndromes (MDS) in Denmark since 2010. The DMDSD has not yet been used for epidemiological research and the quality of registered variables remains to be investigated.

Objective

To describe characteristics of the patients registered in the DMDSD and to calculate predictive values and the proportion of missing values of registered data records.

Methods

We performed a nationwide cross-sectional validation study of recorded disease and treatment data on MDS patients during 2010–2019. Patient characteristics and the proportion of missing values were tabulated. A random sample of 12% was drawn to calculate predictive values with 95% confidence intervals (CIs) of 48 variables using information from medical records as a reference standard.

Results

Overall, 2284 patients were identified (median age: 76 years, men 62%). Of these, 10% had therapy-related MDS, and 6% had an antecedent hematological disease. Hemoglobin level was less than 6.2 mmol/L for 59% of patients. Within the first two years of treatment, 59% received transfusions, 35% received erythropoiesis-stimulating agents, and 15% were treated with a hypomethylating agent. For the majority of variables (around 80%), there were no missing data. A total of 260 medical records were available for validation. The positive predictive value of the MDS diagnosis was 92% (95% CI: 88–95). Predictive values ranged from 64% to 100% and exceeded 90% for 36 out of 48 variables. Stratification by year of diagnosis suggested that the positive predictive value of the MDS diagnosis improved from 88% before 2015 to 95% after.

Conclusion

In this study, there was a high accuracy of recorded data and a low proportion of missing data. Thus, the DMDSD serves as a valuable data source for future epidemiological studies on MDS.

Keywords:

• myelodysplastic syndromes
• cohort
• validation
• accuracy
• database

Acknowledgment

This study was supported by grants from The Danish Cancer Society (Kræftens Bekæmpelse, grant number R223-A13094-18-S68), The Danish Research Center for Equity in Cancer (COMPAS), the Dagmar Marshalls Foundation, and the Danish Acute Leukemia Group. Kirsten Grønbæk is supported by grants from The Danish Cancer Society (Kræftens Bekæmpelse, grant no. R223-A13071 and the Danish Research Center for Precision Medicine in Blood Cancers).

Disclosure

LH declares research grants from Celgene. The authors report no other conflicts of interest in this work.