Mortality Data Collection by Local Authorities Compared to a Cancer Registry

Background: Mortality data, i.e., information on whether individuals are living or dead and the causes of death, provide essential endpoints for the assessment of disease progression in clinical-epidemiological studies. In Germany, mortality data for scientific research are often available only through municipal offices, and rarely from registries. In this study, we collected and compared mortality data from these two sources.

Methods: As part of a study on the early detection of lung cancer and pleural mesothelioma, we collected mortality data on patients recruited in 2014–2015 and on control subjects in the general population. These data were obtained from registration and public health offices and (except for the control subjects) through record linkage with the Cancer Registry North Rhine–Westphalia.

Results: 448 of the 460 patients consented to follow-up via the cancer registry, and 429 consented to municipal-office–based follow up, as did 197 of the 207 control subjects. Much more effort was needed to collect data from municipal offices than through record linkage with the cancer registry. Over a 5-year period, the percentage of deaths recorded was slightly higher when assessed with data from the cancer registry (51.1% versus 49.7%). Further information on the causes of death was more frequently available through the cancer registry (98% versus 84%).

Conclusion: Follow-up data from the cancer registry were generally more informative, and required less organizational effort to collect, than data from municipal offices. Obtaining data from cancer registries is, therefore, preferable for future cancer studies, although nationwide registry structures for mortality data are still lacking.

Cite this as: Hovanec J, Stang A, Kajüter H, Darwiche K, Bölükbas S, Theegarten D, Brüning T, Behrens T: Mortality data collection by local authorities compared to a cancer registry. Dtsch Arztebl Int 2026; 123: 353–6. DOI: 10.3238/arztebl.m2026.0056

Mortality data—information on whether persons are living or dead and the causes of death—are of central importance in clinical–epidemiological studies, used to ascertain typical endpoints such as disease-specific mortality rates and survival times. On the basis of these endpoints, numerous public health measures, e.g., early detection programs, can be evaluated.

In Germany, mortality data are recorded by local authorities and communicated to the statistical office of the federal state concerned. Information on time, place, and cause of death can always be requested from the individual municipal offices. However, the options for centralized collection of mortality data at federal state level differ from state to state. In most federal states, including North Rhine–Westphalia (NRW), there are no mortality registers yielding person-related data for scientific purposes (1, 2). However, the Cancer Registry NRW (CR NRW) contains mortality data for the whole population of NRW, received from the reporting authorities and the NRW State Office for Information and Technology. CR NRW can therefore also register cancer cases on the basis of causes of death (death certificate only, DCO). CR NRW can provide mortality data—along with tumor data—for scientific studies in cancer research (3).

In the course of a study on biomarker screening for early detection of lung cancer and mesothelioma, we obtained retrospective mortality data. Follow-up data were acquired from two different sources:

• Conventionally, from resident registration and public health offices in NRW (municipal office-based follow-up)
• From CR NRW (cancer registry-based follow-up)

In the case of the cancer registry, tumor data were also derived.

This article describes and compares the processes for each form of follow-up and the mortality data thus acquired. Beyond the concrete goal of data collection, the intention was to use the comparison of the two forms of follow-up to help establish the best procedure for derivation of mortality data.

Methods

The two forms of follow-up were carried out in the course of the study “Development of methods in protein analysis for the identification of candidate markers to facilitate the (early) diagnosis of asbestos-associated tumors of the lung and the pleura” (4, 5). The patients for this study were recruited from among those treated for suspected lung cancer or pleural mesothelioma at the Ruhrlandklinik of University Hospital Essen in the period 2013–2015. In parallel, randomly selected control probands with comparable age and sex distributions (population controls) were recruited from the resident registration authority (response rate: 12%). Persons who had participated in the pilot phase of the study were not considered for the follow-ups.

In the main phase of the study, starting in March 2014, written consent was obtained from each participant at each follow-up. These consents were not tied to study participation as such. The population controls could only agree to a municipal office-based follow-up process. For this reason, the numbers of participants for the two forms of follow-up were not identical. The follow-ups were linked with the study data with the help of the study’s data trustee, which holds the requisite identification data (given name, family name, address, date of birth).

Municipal office-based follow-up

The municipal office-based follow-up began in 2021—i.e., 6 years after the end of the recruitment phase—with determination, by inquiry of the resident registration office, whether each individual participant was still alive or had died (date and place of death). The relevant registration office was ascertained from the place of residence held by the data trustee. The registration offices were then asked in writing to supply so-called extended registration information that included the required data on date and place of death. In a second step (2022–2023), data on causes of death were requested from the public health offices for the persons identified as deceased by the resident registration offices. In each case, the public health office responsible was that for the place of death specified by the resident registration office. The registration and public health offices were contacted in writing; if no response was forthcoming, a second written request was sent, followed if necessary by one or more telephone calls.

Cancer registry follow-up

An application to carry out the cancer registry follow-up was lodged with CR NRW. In 2024, following approval, the list of suitable participants was provided via the corresponding CR NRW software and record linkage was performed (3), using encrypted identity data. The mortality data provided included information on whether the person was still alive or had died (with date of death) and the cause of death. The tumor data comprised the causative incident primary lung tumors and pleural mesotheliomas (ICD-10 codes C34, C45, D02.2, D38.1, D38.2, D38.4 or ICD-O-3 code C34) since the start of the study (as of 27 February 2024), plus all “other incident” cancers including in-situ carcinomas—including those before study start. The population controls were not included in the cancer registry follow-up.

Analysis

The analysis phase began with description of the participants in both forms of follow-up. To render the mortality data from the two different sources comparable, mortality figures for individual 5-year periods were determined, starting from the time of study participation. For the same periods we established the proportion of persons who had died of lung cancer or pleural mesothelioma. The data from both follow-ups were merged to yield an overall picture. For those patients who were cancer-free in the study period, the tumor data from the cancer registry were used to identify any new cancers subsequent to the study. Entries in the cancer registry up to the end of 2021 were taken into account.

Results

The Figure shows the original study population and the cohorts available for the follow-ups. A total of 667 members of the study population (460 main-phase patients and 207 population controls) were eligible for follow-up. Consent to municipal office-based follow-up was granted by 626 participants (429 patients and 197 population controls), while 448 patients consented to cancer registry follow-up. Table 1 shows the composition of the follow-up cohorts. The table reflects all available data, without any restriction to specific time periods. Among the total 626 participants who consented to municipal office-based follow-up, the resident registration offices supplied mortality data in 602 cases (96%). For the remaining 24 patients there was either no response or no entry in the register.

Figure

Flowchart of the study population

Table 1

Study population and follow-up analyses*1

Among the 252 deceased participants according to registration office data, in 36 cases (14%) no death certificate was provided by the public health office. Overall, the response rate for municipal office-based follow-up was 90%.

Of the 448 patients with consent to cancer registry follow-up, 368 were included in the cancer registry database due to cancer (n = 349) and/or death (n = 285). Participants not entered in the database were classified as not deceased/free of cancer (n = 80) (Table 1). The majority of persons included in the database were affected by lung cancer/pleural mesothelioma (60 still alive, 199 deceased). A total of 90 persons (23 still alive, 67 deceased) were already registered as having lung cancer/pleural mesothelioma or another form of cancer before the beginning of the study.

Table 2 shows the mortality and causes of death in the first 5 years after study inclusion, with the period determined individually for each participant, combining data from both follow-ups. The overall mortality of the patients with lung cancer, mesothelioma, or other cancer entities within 5 years lay between 62% and 82%. Mortality was considerably lower in participants with benign disease or no findings (12%) and in the population controls (4%). The number of patients documented as dying within 5 years was slightly higher for the cancer registry follow-up than for municipal office-based follow-up (229 of 448 [51.1%] versus 213 of 429 [49.7%]). Restriction of the patient cohort to those who had consented to both forms of follow-up showed that 93% of the deaths (207 of 223) were recorded in both. Ten deaths were established by the cancer registry follow-up but not by the municipal office-based follow-up, while six cases were documented only by means of the municipal office-based follow-up.

Lung cancer was by far the predominant cause of death among patients with lung cancer and accounted for a certain number of deaths in the other groups of participants (Table 2). Information on causes of death in those who died within 5 years was more frequently present in the cancer registry data (98% versus 84%).

Table 2

Five-year mortality and causes of death: combined data from municipal office-based follow-up and cancer registry follow-up

The cancer registry data showed that five patients who were free of cancer at the time of the study died of lung cancer/pleural mesothelioma and a further six of other cancer entities.

Discussion

After a study to identify candidate biomarkers for lung cancer and pleural mesothelioma, two different forms of follow-up were carried out. Data sourced from CR NRW yielded information both on mortality and on tumor entities. Compared with municipal office-based follow-up, the observed relative number of deaths among the patients was slightly higher in the cancer registry data. Information on whether persons were still alive or had died, plus causes of death, was available for 90% of participants on municipal office-based follow-up, whereas causes of death were included in almost cases on cancer registry follow-up. However, it must be assumed that the cancer registry database’s standardized data processing and data transfer procedures, and also the record linkage, are less vulnerable to error than a high number of individual requests for information from local authorities with differing, possibly still manual processes for data classification. In contrast with the record-linkage procedures at the cancer registry, conventional municipal office-based follow-up involves considerable investment of time and effort in contacting the resident registration and public health offices.

Although CR NRW does not constitute a mortality register, it was able to provide mortality data for the study reported here. This illustrates how scientific research can benefit from centralized registry structures for mortality data (1). Follow-up via CR NRW can therefore be recommended for future cancer research studies. In many other German federal states, however, the mortality data do not allow record linkage to individual study participants. Especially multicenter studies covering two or more states would benefit from a nationwide mortality register.

Ethics committee/consent
All study participants gave written consent. The study received advice from the ethics committee of the Faculty of Medicine, Ruhr University Bochum (registration number 4552–12).

Conflict of interest statement
KD is a member of the German Respiratory Society.

The remaining authors declare that no conflict of interest exists.

Manuscript submitted on 13 January 2026, revised version accepted on 27 March 2026

Translated from the original German by David Roseveare

Corresponding author
Dr. rer. medic. Jan Hovanec

References

1. Luttmann S, et al.: Bundesgesundheitsblatt Gesundheitsforschung Gesundheitsschutz 2019; 62: 1500–9

2. Giersiepen K, et al.: Bundesgesundheitsblatt Gesundheitsforschung Gesundheitsschutz 2004; 47: 451–6

3. www.landeskrebsregister.nrw/krebs-in-nrw/datenlieferungen/

4. Brik A, et al.: Dis Markers 2020; 2020: 4176376

5. www.dguv.de/ifa/forschung/projektverzeichnis/ff-fp_0339.jsp (last accessed on 17 February 2026).

e1. Luttmann S, Eberle A, Kibele E, Ahrens W: Perspektiven für ein bundesweites Mortalitätsregister: Erfahrungen mit dem Bremer Mortalitätsindex. Bundesgesundheitsblatt Gesundheitsforschung Gesundheitsschutz 2019; 62: 1500–9

e2. Giersiepen K, Brünings-Kuppe C, Lehmann C: Der Bremer Mortalitätsindex. Bundesgesundheitsblatt Gesundheitsforschung Gesundheitsschutz 2004; 47: 451–6

e3. Landeskrebsregister NRW gGmbH: Krebsregisterdaten für das Gesundheitswesen. www.landeskrebsregister.nrw/krebs-in-nrw/datenlieferungen/ (last accessed on 20 December 2024).

e4. Brik A, Weber DG, Casjens S, et al.: Digital PCR for the analysis of MYC copy number variation in lung cancer. Dis Markers 2020; 2020: 4176376

e5. Deutsche Gesetzliche Unfallversicherung e. V. (DGUV): Einsatz proteinanalytischer Verfahren zur (Früh-)Diagnose asbestassoziierter Lungen- und Pleuratumoren: Projekt-Nr. FF-FP 0339. www.dguv.de/ifa/forschung/projektverzeichnis/ff-fp_0339.jsp (last accessed on 17 February 2026).