Introduction

Myeloproliferative neoplasms (MPN) are rare hematological malignancies with an incidence of 2–4 in 100000 person years and a median age at diagnosis of 71 years.1–4 The main MPN diagnoses encompass essential thrombocythemia (ET), polycythemia vera (PV), prefibrotic myelofibrosis (preMF), primary myelofibrosis (PMF), and unclassifiable MPN (MPN-U).5 The MPN diagnoses may be challenging to distinguish in clinical practice, due to shared features and the potential for progression from one diagnosis to another.6 Indolent diagnoses such as ET exhibit a near-normal life expectancy but in PMF substantially reduced survival is observed.4,7 For rare malignancies with prolonged survival, real-world evidence (RWE) is essential to inform clinical practice and to complement clinical trials, which are often limited by small, highly selected patient cohorts and short duration of follow-up. RWE requires granular, individual-level real-world data with good coverage, high validity, long follow-up and a possibility of crosslinking with other registers. Relevant use of real-world data in MPNs include assessment of health care utilization patterns, delineation of long-term safety of therapies, evaluation of risk factors for progression and complications like vascular events. While the large, administrative Danish registers like the Danish National Patient Register (NPR) have automated data capture, disease-specific registries like The Danish National Chronic Myeloid Neoplasia Registry (DMR) contain granular disease-specific data, mostly from manual registrations performed on site by health professionals.8–10 However, a high validity of the DMR data is crucial for the reliability of research based on the registry. Therefore, we aimed to perform a nationwide validation of clinically relevant variables in the DMR.

The Danish National Chronic Myeloid Neoplasia Registry

The DMR was established in 2010, and it captures >95% of MPN cases in Denmark (crude incidence ~ 600 patients per year) when using the Danish National Patient Register (NPR) as reference.10–12 The DMR initially comprised data from patients with MPN, chronic myeloid leukemia (CML), chronic myelomonocytic leukemia (CMML) and myelodysplastic syndrome (MDS). MPN and CML remain in the DMR, but CMML and MDS have been transferred to The Danish Myelodysplastic Neoplasia Database.13 Reporting of individual-level patient data to the DMR is mandatory for all hematology departments and therefore for all patients with MPN nationwide. Safe-guarding mechanisms for missed registrations are in place through cross-linking of DMR with the NPR, ie if a patient has an MPN diagnosis in the NPR and a registered contact with a hematology department, the relevant department will be reminded to report to the DMR if the patient’s data is not already in the database (Figure 1). Data are reported by local healthcare professionals via three electronic case report forms (eCRF); Baseline, 2-years and 5-years (Figure 1 and Table 1). Additional variables and an end of follow-up eCRF were previously parts of the DMR, but they have been phased out due to automated data capture in the DMR or other national registers. At present, the DMR still retains unique, manually captured data that are only obtainable from reviews of the local electronic health record (eHR). Therefore, this validation of the DMR was done using data from the eHR obtained through local eHR audits.10,14

Table 1 The Danish National Chronic Myeloid Neoplasia Registry - Current Variables of Baseline-, 2-Year- and 5-Year Electronic Case Report Forms (eCRFs)

Figure 1 Overview of patient data flow in the Danish healthcare system and the registration in the relevant Danish registers. Patients are referred to the Department of Hematology for MPN work-up from either primary or secondary care. At each contact with a secondary care facility in Denmark, such as a Department of Hematology, a diagnosis is required. These diagnoses are then automatically transferred to the Danish National Patient Register (NPR) along with contact details ie PIN, date, department ID. If a patient is registered with an MPN diagnosis, it is compulsory to fill in the DMR eCRF for that patient. If a patient is registered in the NPR but not in the DMR, the Department of Hematology that has registered in the contact with an MPN diagnosis is informed to fill in the details in the DMR.

Abbreviations: MPN, Myeloproliferative Neoplasm; DMR, The Danish National Chronic Myeloid Neoplasia Registry; NPR, The Danish National Patient Register; CRS, Civil Registration System; PIN, Personal Identification Number; ICD-10, International Classification of Disease version 10; eCRF, electronic case report forms.

Materials and Methods

To ensure representativeness, we sampled data from patients that fulfilled the inclusion criteria within the two strata that were: (1) The five main MPN diagnoses (ET, PV, PMF, preMF, MPN-U). (2) All five geographical regions of Denmark. The inclusion criteria were: (1) Registration with a personal identification number (PIN) in the DMR. (2) Date of diagnosis in the DMR from 2017 to 2020, except for preMF, where we extended the timeframe to 2021 due to low patient numbers. The year 2017 was chosen to ensure the availability of the registration guide at the time of DMR registration. The registration guide was developed and released in 2016 to align registration practices.15 (3) Both Baseline- and 2-year eCRF registered for each patient in the DMR by the same hematology department to increase accessibility to the full eHR.

We selected eight main variables for validation based on their clinical relevance after consultation with the steering committee of the Danish MPN Study Group (Table 1). Seven were from the Baseline eCRF and one was from the 2-year eCRF. Six variables were mandatory for all patients, while two were conditional, appearing in the eCRF only if the entries indicated their applicability; for instance, the variable Molecular findings appeared only if it was entered that Molecular investigations had been conducted. Three variables, Treatment 0–2 years, Specification of cardiovascular risk factors, and Molecular findings, consisted of a total of 27 subordinate variables, and MPN diagnosis was a categorical variable with five response categories (ET, PV, PMF, preMF, MPN-U), but we report the validity of each of the five diagnoses individually. Three of the subordinate variables of Molecular findings: ASXL1, JAK2 Exon12, and Other abnormal findings were combined in one variable called Other findings due to few expected cases. A detailed overview of each selected variable is shown in the Table S1.

In the DMR Baseline eCRF, the response categories for the mandatory variables (Cardiovascular risk factors, Splenomegaly, and Molecular investigations) were positive (yes), negative (no), or unknown. But for all dichotomous subordinate variables (Treatment 0–2 years, Specification of cardiovascular risk factors, and Molecular findings) only positive (yes) was actively entered, and fields left empty were automatically imputed as negative (no) in the DMR. Therefore, the distinction between negative (no), unknown (as an active entry), and missing data was not possible for the subordinate variables. The audit eCRF was more comprehensive because the audit response categories were positive (yes), negative (no), and unknown for all dichotomous variables other than Molecular findings, where only positive (yes) could be entered as in the DMR eCRF.

All auditors were physicians experienced in hematology, and they received uniform instructions by the study team on how to perform the audit by adhering to the 2016 registration guide.15 DMR data were not blinded to auditors. Thus, specifically for MPN diagnosis, if two diagnoses were equally plausible according to the eHR at the time of DMR entry, the audit assigned the same diagnosis. Audit data were collected using REDCap.16,17

After the audit, we categorized data from the DMR and the audit so that missing data and unknown data were registered but otherwise left out of subsequent analyses. The remaining entries were used for validity assessment by comparing DMR entries to audit entries (reference). We report most variables as true positive (TP), false positive (FP), true negative (TN), false negative (FN), albeit Date of diagnostic sample, MPN diagnosis, and tobacco Packyears could only be true or false, but for the sake of uniform presentation of data, we report them as TP and FP. For tobacco Packyears, ±20% of the audit value in the DMR was deemed acceptable, and for Date of diagnostic sample, ±30 days were allowed. Subsequently, we calculated the validity metrics; positive predictive value (PPV), negative predictive value (NPV), sensitivity, specificity, and accuracy. Wilsons score interval was used to estimate 95% confidence intervals (95% CI) using the binom package for R in Visual Studio Code.18,19 Ultimately, validity metrics were categorized as low (0–0.69), intermediate (0.70–0.89), or high (0.90–1.0).

We further quantified and visualized correlations for Packyears and Date of diagnostic sample using Pearson’s correlation coefficient.20

The study was approved by the Capital Region of Denmark (R-24025261), and all data were handled according to local data protection legislation.

Results

Our sample included DMR entries regarding 372 patients (75 patients each for ET, PV, PMF, and MPN-U and 72 for preMF, Table 2). The year of diagnosis ranged from 2017–2020, with 72, 85, 84, and 107 patients in each year. Additionally, 24 patients diagnosed with preMF in 2021 were included to increase the number of that subgroup. The median age at diagnosis was 71 years (IQR: 54–88); for ET, PV, PMF, preMF and MPN-U the medians were 67, 70, 72, 72, and 72 years. Forty-eight percent of the patients were female. Four variables (MPL, Busulfan, Vorinostat, and No treatment) were present in too few cases (≤ 3) and therefore analyses were not meaningful for these variables.

Table 2 Validity Metrics of The Danish National Chronic Myeloid Neoplasia Registry (DMR) and 95% Confidence Intervals

In the DMR, Splenomegaly data were missing for 169 patients due to eCRF limitations in the period 2016–2018, and for Busulfan data were missing for 86 patients for unknown reasons. Apart from Splenomegaly and Busulfan, values for all variables were present.

Cardiovascular risk factors were actively registered as unknown in the DMR for 12 patients, but all of these had relevant information in the eHR by audit. Furthermore, for 11 additional patients Cardiovascular risk factors were marked as unknown in the audit, but in the DMR all 11 were marked as absent (no). The Specification of cardiovascular risk factors (Obesity, Hypertension, Hyperlipidemia, Smoking, Previous smoking, and Diabetes) could not be reported as unknown in the DMR, but they were unknown by audit for 16, 6, 5, 11, 25 and 3 patients, respectively. For five patients, Molecular investigations were unknown in the DMR, but relevant information was found in the eHR for all patients by audit. Variables in Treatment 0–2 years could not be reported as unknown in the DMR, and in the audit relevant information was retrievable from the eHRs in nearly all cases (> 99%).

Validity

Of the 372 patients in the DMR, 366 (98%) had an MPN diagnosis according to the eHR audit results; four patients entered as PV and two as MPN-U in the DMR did not meet the 2016 revised World Health Organization (WHO) criteria for any MPN:5 Three of the six had clonal hematopoiesis, two had secondary erythrocytosis, and one had a bone marrow biopsy with MPN-like morphology albeit not meeting diagnostic criteria. High PPV was observed for; ET 0.91 (0.82-0.95), PV 0.95 (0.87-0.98), and PMF 0.97 (0.91–1.00), and intermediate PPV for preMF 0.78 (0.67-0.86) and MPN-U 0.77 (0.67-0.85) (Table 2). Date of diagnostic sample ±30 days had a PPV of 0.97 (0.95-0.99) and only four patients had discrepancies between the DMR date and the date in the eHR exceeding six months (Figure 2A). For tobacco Packyears, we found an intermediate PPV of 0.82 (0.75-0.89) and three patients had discrepancies exceeding 10 packyears (Figure 2B). The PPV of other cardiovascular risk factors ranged between 0.75 (0.63-0.83) and 0.97 (0.85-0.99). Splenomegaly had high PPV; 0.98 (0.90–1.0), NPV; 0.93 (0.88-0.96), and specificity; 0.99 (0.96–1.0) and intermediate sensitivity; 0.84 (0.73-0.91). JAK2V617F and CALR mutations have similarly high validity metrics. Of note, in the DMR, six patients were marked as having both mutations (data not shown), but in our audit, we found only one patient to carry both JAK2V617F and the CALR mutations. Audit reported the CALR mutation only in three of these six patients and JAK2V617F mutation only in two of these patients. Several MPN treatments (Phlebotomy, Hydroxyurea, Interferon, Ruxolitinib) had high PPV (≥0.93), while intermediate PPVs were observed for Anagrelide; 0.83 (0.55-0.95) and JAK-inhibitor; 0.86 (0.49-0.97).

Figure 2 Correlations between The Danish National Chronic Myeloid Neoplasia Registry and the audit values of Date of diagnostic sample and tobacco packyears. (A) Shows the date of the variable Date of diagnostic sample as entered in the DMR on the x-axis and as entered in the audit on the y-axis. Four patients have discrepancies exceeding six months (B) Shows the value of the variable tobacco packyears as entered in the DMR on the x-axis and as entered in the audit on the y-axis, three patients have discrepancies exceeding 10 packyears. For both (A) and (B) each dot represents a patient that has a value both in the DMR and at journal audit, and correlations are of type Pearson.

Abbreviation: DMR, The Danish National Chronic Myeloid Neoplasia Registry.

Of the 23 variables where both FN and FP were assessed, 19 variables exhibited a higher FN than FP count, indicating that misclassification more often occurred due to positives being misclassified as negatives. Of these 23 variables, the highest FP count were for Smoking (nFP = 18, 5% of total), Other findingsmolecular (nFP = 17, 5% of total), Other treatment (nFP = 15, 4% of total), and FN counts were highest for the variables Statins (nFN = 94, 26% of total), Hyperlipidemia (nFN = 75, 20% of total), and Salicylic acid (nFN = 73, 20% of total).

The distribution of the number of variables with low/intermediate/high validity for each validity metric is shown in Figure 3. For the metric PPV, the validity distribution was 2/9/19 and for NPV 3/8/12. For sensitivity, specificity, and accuracy it was 12/7/4, 1/1/21, and 0/9/14, respectively. Overall, PPV, NPV, specificity and accuracy are skewed toward high validity while sensitivity is skewed toward low validity reflecting the overall higher FN rates in the DMR.

Figure 3 The distribution of the number of variables with low-, intermediate-, and high validity by type of validity metric in The Danish National Chronic Myeloid Neoplasia Registry. Validity metrics for each variable are categorized as low, intermediate, or high for values of 0-0.69, 0.70-0.89, and 0.90–1.00, respectively. Validity metrics are the positive predictive value: true positive / (true positive + false positive), negative predictive value: true negative / (true negative + false negative), sensitivity: true positive / (true positive + false negative), specificity: true negative / (true negative + false positive), and accuracy: (true positive + true negative) / total.

Abbreviations: PPV, positive predictive value; NPV, Negative predictive value.

Discussion

This study validates the DMR via a journal audit of the eHRs of 372 persons registered in the DMR with a date of diagnosis between 2017–2020 (preMF: 2017–2021). Data availability in the eHR was high, supporting a robust validity assessment. Strengths of this study include the validation of diverse variable types and the balanced sampling by MPN diagnosis and geographical region, ensuring national representativeness. Limitations include the restricted time frame (2017–2021), use of a different eCRF system for validation, lack of blinding, and single-auditor data entry per patient, all of which may introduce bias. However, the absence of blinding also allowed for a pragmatic evaluation of variables—particularly MPN diagnosis, which is inherently complex because one subtype may progress into another, and diagnostic criteria often rely on subjective interpretation.6

To better understand the magnitude of the diagnostic challenge, an inter-rater reliability study of the MPN diagnosis would be valuable, although it lies beyond the scope of the present study.

As diagnostic knowledge evolves, so do the diagnostic criteria and in 2016 the diagnosis of preMF was introduced.5,21,22 Patients now classified as preMF were previously included in other subtypes. These evolving definitions affect comparability across time periods and should be considered when conducting research.

The role of manual data entry should also be considered, because it may introduce errors as illustrated with five patients incorrectly registered as having both JAK2V617F and CALR mutations, but it also allows for clinical judgment of complex variables. An example of this is the variable Hypertension, which in the DMR requires either antihypertensive drug use, repeated high blood pressure measurements, or hypertension reported as comorbidity in the eHR.15 The Hypertension validity metrics (PPV; 0.93 (0.87-0.96), NPV; 0.75 (0.69-0.80), sensitivity; 0.66 (58-0.72), specificity; 0.95 (0.91-0.97)) reflect a general pattern of higher PPV/specificity and lower sensitivity, due to a high FN/FP ratio. Because the PPV is high it more likely reflects underreporting and not misclassification. In comparison, a recent study investigating the validity of hypertension defined through automatically captured prescription data versus self-reported hypertension in Denmark reported a PPV of 0.88, NPV of 0.94, sensitivity of 0.75, and specificity of 0.96.23

The use of DMR has so far primarily been for administrative purposes with annual reports, but four studies have utilized the DMR variables MPN diagnosis and Date of diagnosis.24–27 The data regarding CML in the DMR remain unvalidated. For MPN, we have demonstrated that several variables – including, MPN diagnosis, Date of diagnostic sample, Splenomegaly, several Specific cardiovascular risk factors, common Molecular findings and several treatment modalities exhibited high PPV, which are comparable to other Danish hematological databases.28,29 In addition, comparison with the similarly structured Swedish National Myeloproliferative Register revealed similar PPV for all comparable variables indicating similar reporting practices in a separate national MPN registry.30 To assess comparability with additional MPN registries we recommend separate validation studies for those registers.

The key additions of the DMR to future epidemiological research are the disease-specific variables. These variables enable a level of data granularity that approaches that of clinical trials. An example is the stratification of patients by mutational status, which is linked to distinct disease trajectories.31 Also, such variables can help delineate long-term safety of available therapies, reveal differences in health-care utilization patterns based on real-world practices and inform risk factors for disease progression and complications like vascular events. For rare cancers like MPN, these applications are particularly important because the clinical trials that influence clinical practice and treatment approvals often involve fewer patients and shorter follow-up than those conducted with patients with more common diseases.32 However, registry data collection processes may be less standardized and transparent than those of clinical trials as exemplified by the DMR registration guide, which was released in 2016, six years after the start of registration in the DMR and by the ongoing updates of the DMR eCRFs.15 Such changes can affect the consistency of data entry and interpretability, increasing the risk of misclassification and limiting comparability with other sources. This underscores the importance of critically scrutinizing registry documentation and data quality when using registry-based data for research.

Conclusion

When using data from the DMR, the distribution of data should be compared within the DMR for different time periods and if possible, it is advised to validate variables specifically for the planned study either internally through journal audit or externally using comparable validated data sources such as the Swedish National Myeloproliferative Register. However, we have now demonstrated that DMR holds high PPV for several clinically relevant variables in the study period (2017–2021) and that the DMR as such is a valid source of real-world data for this period. Especially, variables with high validity and clinical importance such as Splenomegaly, JAK2V617F, and CALR provide increased granularity, thereby enabling an opportunity to delineate long-term outcomes for these patients. Crosslinking these variables with other data sources may therefore provide the foundation for unprecedented precision in RWE for MPN patients.

Abbreviations

MPN, Myeloproliferative neoplasia; ET, Essential thrombocythemia; PV, Polycythemia vera; preMF, Prefibrotic myelofibrosis; PMF, Primary myelofibrosis; MPN-U, Unclassifiable MPN; RWD, Real-world data; DMR, Danish National Myeloproliferative Neoplasm Registry; eCRF, Electronic case report form; eHR, Electronic health record; PIN, Personal identification number; TP, True positive; FP, False positive; TN, True negative; FN, False negative; PPV, Positive predictive value; NPV, Negative predictive value; CI, Confidence interval; WHO, World Health Organization; EHA, European Hematology Associations.

Acknowledgments

Tarec Christoffer El-Galaly and Christen Lykkegaard Andersen are co-senior authors for this study. We thank the DMR and the Danish MPN Study Group for their advice. Finally, we thank the Danish Research Center for Precision Medicine by the Danish Cancer Society (R223-A13071) and the dedicated Danish hematologists and other health care professionals for their contributions to building and reporting to the DMR. A version of the abstract from this study was uploaded for the conference European Hematology Associations (EHA) 2025.33

Author Contributions

All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.

Funding

This study was funded by the Mica Foundation, the Danish Cancer Society (R378-A22319), The Danish MPN Study Group and The A.P Møller Foundation for General Purposes.

Disclosure

Dr Peter Brown reports personal fees from Roche, Gilead, and BMS, outside the submitted work. Dr Kirsten Grønbæk reports grants from Janssen-Cilag (Johnson and Johnson) and Medac, outside the submitted work. The authors declare no other competing interests in this work.

References

1. Rollison DE, Howlader N, Smith MT, et al. Epidemiology of myelodysplastic syndromes and chronic myeloproliferative disorders in the United States, 2001-2004, using data from the NAACCR and SEER programs. Blood. 2008;112(1):45–12. doi:10.1182/blood-2008-01-134858

2. Titmarsh GJ, Duncombe AS, McMullin MF, et al. How common are myeloproliferative neoplasms? A systematic review and meta-analysis. Am J Hematol. 2014;89(6):581–587. doi:10.1002/ajh.23690

3. Hultcrantz M, Ravn Landtblom A, Andréasson B, et al. Incidence of myeloproliferative neoplasms – trends by subgroup and age in a population-based study in Sweden. J Internal Med. 2020;287(4):448–454. doi:10.1111/joim.13019

4. Hultcrantz M, Kristinsson SY, Andersson TML, et al. Patterns of survival among patients with myeloproliferative neoplasms diagnosed in Sweden from 1973 to 2008: a population-based study. J Clin Oncol. 2012;30(24):2995–3001. doi:10.1200/jco.2012.42.1925

5. Arber DA, Orazi A, Hasserjian R, et al. The 2016 revision to the World Health Organization classification of myeloid neoplasms and acute leukemia. Blood. 2016;127(20):2391–2405. doi:10.1182/blood-2016-03-643544

6. Baumeister J, Chatain N, Sofias AM, Lammers T, Koschmieder S. Progression of myeloproliferative neoplasms (MPN): diagnostic and therapeutic perspectives. Cells. 2021;10(12):3551. doi:10.3390/cells10123551

7. Barbui T, Thiele J, Passamonti F, et al. Survival and disease progression in essential thrombocythemia are significantly influenced by accurate morphologic diagnosis: an international study. J Clin Oncol. 2011;29(23):3179–3184. doi:10.1200/jco.2010.34.5298

8. Bak M, Ibfelt EH, Stauffer Larsen T, et al. The danish national chronic myeloid neoplasia registry. Clin Epidemiol. 2016;8:567–572. doi:10.2147/CLEP.S99462

9. Erlangsen A, Fedyszyn I. Danish nationwide registers for public health and health-related research. Scandinavian J Public Health. 2015;43(4):333–339. doi:10.1177/1403494815575193

10. Lynge E, Sandegaard JL, Rebolj M. The Danish national patient register. Scand J Public Health. 2011;39(7 Suppl):30–33. doi:10.1177/1403494811401482

11. Dansk database for kroniske myeloproliferative neoplasier - annual report. 2023. Available from: https://www.sundk.dk/media/t23faipt/dmpn-aarsrapport-2023.pdf. Accessed March27, 2026.

12. Dansk database for kroniske myeloproliferative neoplasier - annual report 2022. Available from: https://www.sundk.dk/media/akolsphf/d41a65e11eb84e2ea4996c66959eca58.pdf. Accessed March27, 2026.

13. Ostgard L, Nørgaard JM, Raaschou-Jensen K, et al. The Danish national acute leukemia registry. Clin Epidemiol. 2016;8:553–560. doi:10.2147/clep.s99460

14. Arendt JFH, Hansen AT, Ladefoged SA, Sorensen HT, Pedersen L, Adelborg K. Existing data sources in clinical epidemiology: laboratory information system databases in Denmark. Clin Epidemiol. 2020;12:469–475. doi:10.2147/CLEP.S245060

15. Registration guide DMR 01-2016.pdf. Danish Hematology Society. 2016.

16. Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N, Conde JG. Research electronic data capture (REDCap)—A metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Informat. 2009;42(2):377–381. doi:10.1016/j.jbi.2008.08.010

17. Harris PA, Taylor R, Minor BL, et al. The REDCap consortium: building an international community of software platform partners. J Biomed Informat. 2019;95:103208. doi:10.1016/j.jbi.2019.103208

18. Dorai-Raj S. Binomial confidence intervals for several parameterizations - R. 2022.

19. Wilson EB. Probable inference, the law of succession, and statistical inference. J Am Stat Assoc. 1927;22(158):209–212. doi:10.1080/01621459.1927.10502953

20. Pearson K. Note on regression and inheritance in the case of two parents. Proceedings of the Royal Society of London. 1895;58:240–242.

21. Khoury JD, Solary E, Abla O, et al. The 5th edition of the world health organization classification of haematolymphoid tumours: myeloid and histiocytic/dendritic neoplasms. Leukemia. 2022;36(7):1703–1719. doi:10.1038/s41375-022-01613-1

22. Swerdlow SH, Campo E, Harris NL. WHO Classification of Tumours of Haematopoietic and Lymphoid Tissues. 4th ed. IARC Press; 2008.

23. Bonnesen K, Schmidt M. Validity of prescription-defined and hospital-diagnosed hypertension compared with self-reported hypertension in Denmark. Clin Epidemiol. 2024;16:249–256. doi:10.2147/clep.s448347

24. Christensen SF, Svingel LS, Kjærsgaard A, et al. Healthcare resource utilization in patients with myeloproliferative neoplasms: a Danish nationwide matched cohort study. Eur J Haematol. 2022;109(5):526–541. doi:10.1111/ejh.13841

25. Kristensen DT, Kisbye Øvlisen A, Hjort Kyneb Jakobsen L, et al. Use of statins and risk of myeloproliferative neoplasms: a Danish nationwide case-control study. Blood Adv. 2023;7(14):3450–3457. doi:10.1182/bloodadvances.2023009784

26. Kristensen DT, Øvlisen AK, Jakobsen LHK, et al. Metformin use and risk of myeloproliferative neoplasms: a Danish population–based case-control study. Blood Adv. 2024;8(16):4478–4485. doi:10.1182/bloodadvances.2023012266

27. Svingel LS, Christensen SF, Kjærsgaard A, et al. Labor market affiliation of patients with myeloproliferative neoplasms: a population-based matched cohort study. Acta Oncologica. 2023;62(10):1286–1294. doi:10.1080/0284186x.2023.2251670

28. Ostgard LS, Norgaard JM, Severinsen MT, et al. Data quality in the Danish national acute leukemia registry: a hematological data resource. Clin Epidemiol. 2013;5:335–344. doi:10.2147/CLEP.S48411

29. Arboe B, El-Galaly TC, Clausen MR, et al. The Danish national lymphoma registry: coverage and data quality. PLoS One. 2016;11(6):e0157999. doi:10.1371/journal.pone.0157999

30. Nationellt kvalitetsregister för MPN - Validering. Available from: https://cancercentrum.se/download/18.3de350c219545ea25bd241b8/1741111577379/valideringsrapport-mpn-registret_2023-08-22.pdf. Accessed March27, 2026.

31. Grinfeld J, Nangalia J, Baxter EJ, et al. Classification and personalized prognosis in myeloproliferative neoplasms. N Engl J Med. 2018;379(15):1416–1430. doi:10.1056/nejmoa1716614

32. Duijnhoven RG, Straus SM, Raine JM, de Boer A, Hoes AW, De Bruin ML. Number of patients studied prior to approval of new medicines: a database analysis. PLoS Med. 2013;10(3):e1001407. doi:10.1371/journal.pmed.1001407

33. Patel DA, Kannik K, Chanchiri I, et al. Validity of the Danish national chronic myeloid neoplasia registry. Abstract from European hematology associations (EHA) 2025. 2025. Available from: https://library.ehaweb.org/eha/2025/eha2025-congress/4162169/dustin.andersen.patel.validity.of.the.danish.national.chronic.myeloid.html?f=listing%3D3%2Abrowseby%3D8%2Asortby%3D1%2Amedia%3D1. Accessed March27, 2026.