Effects of electronic health records implementation on workflow and health care provider outcomes: a prospective, multi-professional cohort study Authors: Matthias Weigl, Karin Berger, Julia Mayerle, Jürgen Behr, Michael von Bergwelt-Baildon, Helmut Ostermann, Katharina Schmidt Published: 2026 ABSTRACT AbstractBackgroundProspective studies examining electronic health record (EHR) implementation in multi-professional healthcare provider samples remain scarce, particularly with respect to impacts on providers’ workflows and outcomes in real-world clinical settings.ObjectivesTo determine the prospective effects of workflow interruptions, multitasking, information overload, and burnout over time among physicians and nurses in before and after EHR adoption.Materials and MethodsMulti-methods, unit-based, and before-after-design was established to determine provider experiences before and after EHR implementation (5 months-time lag). Information on key outcome variables was assessed via standardized surveys among healthcare professionals and augmented through qualitative interviews with stakeholders to determine the quality of the EHR implementation. Prospective effects were determined within and across professions with multivariate analyses.ResultsA total of 163 physicians and nurses participated (baseline: n = 87, follow-up: n = 76). Interviews revealed different implementation strategies across professions with varying procedures, training opportunities, and resources: nurses were supported via a more structured and on-site EHR introduction process, whereas physicians were offered a more self-directed and organized process. Overall, we did not find significant differences over time for workflow variables and burnout. Moderation analyses showed 2 significant effects: information overload for nurses was significantly reduced at follow-up; reported depersonalization in physicians was higher at follow-up.DiscussionThis prospective study provides important contributions into the varying effects of profession-specific EHR adoption processes for provider workflow and burnout. EHR utilization did not per se improve workflow and provider experience.ConclusionFuture studies should investigate the interplay of profession-specific management of EHR adoption processes and respective provider outcomes.Keywords: electronic health records, electronic medical records, hospital, digitalization, facilitators, workflow, healthcare, hospital Background The adoption of digital systems in healthcare offers profound benefits for both patient outcomes and provider performance. In recent decades, the introduction of health information technology has been shown to impact healthcare safety and quality as well as provider well-being. Of particular interest has been the transition from paper-based documentation to electronic records, with widespread adoption of electronic health records (EHR). The introduction of EHR has been proposed to reduce documentation burden with potentially positive ramifications for provider satisfaction. Despite the extensive research and growing literature base on provider effects of EHR implementation into healthcare work and inherent effects on provider, the available evidence remains inconclusive, for several reasons. First, available research is inconsistent, partly due to missing prospective designs that allow reliable inferences concerning trajectories over time.1,2 Despite the abundance of research in this field, just a few prospective studies (ie, capturing pre- and post-implementation status) that allow inferences concerning actual changes over time are available.3–5 Naturalistic observations that prospectively track developments in real-world clinical environments in course of EHR implementation may provide more insightful inferences than retrospective approaches.6 Moreover, previous research demonstrated that stark contrasts of pre-implementation expectations and post-implementation experiences and outcomes may occur.7,8 Secondly, there is a lack of investigations comprehensively surveying multiple healthcare professions.9 The majority of surveys often focus on 1 health care profession, for example, nurses.10 This limits inferences with higher external validity concerning the overall impact of technology adoption for the multi-professional workforce in healthcare. Beyond methodological limitations in the literature base, previous studies showed that particularly clinical work environment factors are important when assessing work stress in course of EHR implementation.11,12 Particularly workflow factors have been in the focus of previous investigations. In frontline clinical care, alignment to clinical workflows is key to acceptance and physician well-being.13,14 Perceived efficiency of EHR is essential for physician satisfaction.4,15 Issues of poorly designed user interfaces and poor usability have been associated with increased likelihood of burnout.10 Opportunities to gain proficiency in using EHR functionalities are essential.16 Therefore, several investigations into the effects of EHR implementation and resulting challenges for healthcare providers have been drawing upon a sociotechnical perspective. Such system-based frameworks in healthcare emphasize that digital tools are not isolated but components of a larger work- and care-system composed of technologies, humans, and organizational rules working together toward shared goals.17 For example, the systems engineering initiative for patient safety (SEIPS) framework informed previous investigations into identifying the broader spectrum of critical contextual as well as provider factors during health IT implementation.18 Investigations that draw upon sociotechnical thinking assume that health IT outcomes depend not just on the technology itself, but on workflows, people, organizational context, and implementation practices.19,20 Further perspectives have emphasized the linkage between poor health IT design and unintended workflow outcomes of health IT such as alterations, blockages, and fragmentation.21,22 Moreover, provider outcomes—such as interruptions, multitasking, and information processing—unfold in a dynamic and complex system of various organizational, work, process, and provider factors.23 Another persistent shortcoming is lack of consideration of the actual implementation process in course of EHR introduction in the clinic. Successful EHR implementation in the hospital is a challenging, interdisciplinary, and collaborative process of multiple stakeholders and professions with leveraging project management techniques for coordination and consistent leadership support during the implementation process.16,24 Various implementation factors are critical to successful adoption of EHR, including sustainable efforts to facilitate day-to-day use of EHR, for example, through post-launch training.5,14 To this end, despite various promising advantages of digitalized records in hospital care, there is a notable lack of studies focused on EHR implementation, particularly regarding its effects on workflow and provider outcomes in real-world clinical settings. We therefore set out a prospective cohort study to survey the experiences of healthcare providers in course of implementation of EHRs in a hospital. Objectives Specifically, we sought to determine the trajectories of perceived workflow interruptions, multitasking, information overload, and burnout over time among physicians and nurses in course of EHR adoption. Materials and methods Study design A before-after-design with a 5-month-time lag was established to determine reported experiences before and after EHR implementation. Furthermore, this prospective evaluation used quantitative (ie, standardized surveys among healthcare professionals) and qualitative methods (ie, interviews with stakeholders). Baseline assessments started in 08-09/2021, EMR implementations were rolled out in clinics during 10-11/2021, and follow-up surveys were conducted from 02-06/2021. Prior to start, the study was approved by the Ethics Committee of the Medical University of LMU University, Munich, Germany (registry code: 21-0142). For reporting, we used STROBE Reporting Guidelines for cohort studies.25 Our reporting of interview methods drew upon the checklist of the Consolidated Criteria for Reporting Qualitative Research.26 Setting and prospective data collection The study was conducted at an academic university hospital in South Germany. Three departments (2 clinics for internal medicine, 1 surgical clinic) were surveyed. Altogether, the clinics employed throughout the investigation around 90 physicians and 150 nurses (full- and part-time). Baseline survey captured pre-implementation status, with all patient records were paper based, including all vital parameters, medication, and documentation. The paper-based records were basis of all handover procedures between professions as well as between shifts. In the course of EHR implementation, a comprehensive EHR system was introduced across all 3 clinics. A commercial EHR solution developed by a Germany-based company was acquired and implemented in a stepwise manner. The system represents a typical electronic medical record (EMR) platform and includes several key functionalities: patient data documentation, providing access to all relevant clinical information within the respective departments (including clinical notes, vital parameters, and care-related information); computerized provider order entry, with an integrated algorithm to detect potential drug interactions and incompatible orders; the presentation of diagnostic results; and clinical decision support (CDSS) to assist in the decision-making process. Additionally, the EHR offers an interface with the preexisting clinical information system. At both waves, data collection consisted of standardized surveys with paper-based questionnaires distributed via internal mail (in sealed envelopes). Follow-up surveys assessed post-implementation status with identical methods after EHR implementation in respective clinical departments. Eligible professionals were informed and recruited via internal departmental meetings, intranet, and email. Moreover, a study team member attended weekly team meetings, informed, distributed, and collected completed surveys immediately afterwards. For absent team members, surveys were directly deposited in their individual mailboxes. Additionally, we conducted semi-standardized interviews with key stakeholders and professionals from surveyed departments to obtain deeper insights into the course of the actual implementation process and status of EHR utilization at follow-up. Interviews were conducted using an interview guide capturing questions concerning current work practices and expectations (for baseline interviews; see Table SA-1) as well as experiences with and status after EHR implementation (for follow-up). Key themes of the interview guide were informed by previous research syntheses of the field.19,20 A convenience recruiting sampling approach was used, where a study team member (author KS) approached clinicians and stakeholders who were assuming key roles in the implementation process or intra-departmental promotion of the new EHR software. All interviews lasted usually around 20 min and were all conducted by the 1 study team member (KS). Participants At both time points, a convenience sampling approach was established. All physicians and nursing professionals of respective clinics were invited to participate. Eligibility criteria were: (1) physician (including residents) or nursing professional, (2) professional tenure > 2 years, and (3) organizational tenure > 6 months. Data sources and measurement We surveyed nurses and physicians with standardized survey tools at baseline and follow-up. The following identical measures were perceived levels of workflow interruptions, multitasking, information overload, and burnout: Workflow interruptions were assessed with 3 items of an established work analysis tool for hospitals.27 A sample item is “I often have to interrupt my work due to electronic messages (eg, e-mail, device message).” Internal consistency was Cronbach’s Alpha (CA) = .50 (baseline)/.57 (follow-up). Workflow interruptions were assessed with 3 items of an established work analysis tool for hospitals.27 A sample item is “I often have to interrupt my work due to electronic messages (eg, e-mail, device message).” Internal consistency was Cronbach’s Alpha (CA) = .50 (baseline)/.57 (follow-up). Multitasking was measured with 2 items using a scale from another work analysis tool28; example item: “Due to digital technologies I have to work on several tasks at the same time.” CA = .86 (baseline)/.92 (follow-up). Multitasking was measured with 2 items using a scale from another work analysis tool28; example item: “Due to digital technologies I have to work on several tasks at the same time.” CA = .86 (baseline)/.92 (follow-up). Information overload was also assessed with 2 recently published survey items29; example item: “I feel that the information I receive via on-duty digital media is too much.” CA = .89 (baseline)/.91 (follow-up). Information overload was also assessed with 2 recently published survey items29; example item: “I feel that the information I receive via on-duty digital media is too much.” CA = .89 (baseline)/.91 (follow-up). All workflow items were answered on a 5-point Likert scale (1 = “not at all” to 5 = “to a very great extent”). Burnout: two scales from the German-language version of the Maslach Burnout Inventory (MBI-D,30) were used to assess burnout symptoms. Emotional exhaustion (EE) and depersonalization (DP) were measured with 3 items, respectively. Answers were collected on a 6-point Likert scale (1 = never; 6 = very frequent). Internal reliabilities were satisfactory at baseline (EE: .90, DP: .59) as well as at follow-up (.88, .81) Burnout: two scales from the German-language version of the Maslach Burnout Inventory (MBI-D,30) were used to assess burnout symptoms. Emotional exhaustion (EE) and depersonalization (DP) were measured with 3 items, respectively. Answers were collected on a 6-point Likert scale (1 = never; 6 = very frequent). Internal reliabilities were satisfactory at baseline (EE: .90, DP: .59) as well as at follow-up (.88, .81) Furthermore, the following sociodemographic information were obtained at both waves: age (categories: < 21 years, 21-30 years, 31-40 years, 41-50 years, > 50 years), gender (categories: female, male, diverse), profession (categories: physician, nurses, other), and professional tenure (categories: < 6 years, 6-10 years, 11-15 years, 15 years). Statistical methods and analyses Analyses included descriptive statistics as well as the testing for differences over time with t-test (for overall group changes over time) and univariate mean-difference tests with tests of interactions effects (for changes across professional groups over time). All analyses were conducted with SPSS 29 (IBM Inc., Chicago, US). Interview data on EHR pre-implementation expectations, the actual implementation process, and post-implementation appraisals were first transcribed verbatim. Subsequently, a qualitative analysis was conducted using thematic content analysis (ie, combination of deductive and inductive approaches).31 After multiple readings of the transcripts and accompanying notes, key statements (ie, recurring themes and meanings) relevant to the research question were identified and coded by 1 member of the study team (KS). These statements were then reviewed and interpreted in collaboration with a second author (MW), in a combination of inductive and deductive approach. All professionals’ statements were then extracted separately, grouped, and synthesized into key themes.31 Results Sample description and test for group differences at baseline Altogether, 163 healthcare professionals filled out the survey (baseline: n = 87, follow-up: n = 76); with responses from 81 physicians and 82 nurses. Among all eligible professionals and across both assessments, we obtained a response rate of 45% among physicians and 27.3% among nursing professionals. At baseline, we did not identify any differences concerning the sociodemographic characteristics between all 3 clinical departments, neither for gender (Chi2(df = 2) = 2.29, P=.32), age (Chi2(df = 8) = 5.84, P=.67), and professional tenure (Chi2(df = 6) = 3.03, P=.80). We only found a significant difference for professions at baseline, with a larger proportion of physicians in the surgical department compared to both internal medicine departments (Chi2(df = 3) = 14.37, P < .01); detailed results are available from the corresponding author. Interview results on EHR implementation and provider experiences over time Across both time points, we conducted 8 interviews with 6 professionals (2 nurses, 3 resident doctors, 1 senior consultant; see Table SA-2). Our analyses revealed the following key themes (codes with illustrative data extracts are depicted in Table SA-3). In line with the initial project plan, the EHR was to be introduced stepwise across all participating clinical departments. Before rollout, nursing staff received on-site training in small groups, and were supported on the wards by hospital IT support staff during the first days after go-live. In contrast, physicians received training via an online portal for self-directed learning, including instructional videos on EHR use. At baseline, interviewed physicians and nurses reported mixed experiences with paper-based documentation and HIS: common themes were accessibility, usability, and information clarity, including issues such as illegible notes and transmission errors, as well as perceived system reliability. High stress levels were linked to managing patient load, planning, and task allocation, especially in interprofessional collaboration. Overall, physicians rated their use of paper-based records more favorably than nurses, possibly reflecting physicians’ greater flexibility in how written information is used. Before EHR implementation, physicians expressed more skepticism than the interviewed nurses. Their main concerns were potential work and information overload and perceived system “blackouts” associated with introducing a new EHR. Only a few anticipated clear benefits, such as reduced administrative burden, improved communication and coordination with nursing, and better interoperability with the existing clinical information system. In contrast, nurses reported more positive expectations, including faster access to patient information and improved visibility of key data, as well as potential time savings and better clarification of physicians’ documentation. They also raised concerns about insufficient IT infrastructure (eg, lack of workstations and touch monitors near patient rooms) and limited software usability due to poorly designed, unintuitive user interfaces. At follow-up and after transitioning to the EHR, interview statements showed a broad range of experiences among providers during both implementation and subsequent use. Concerning the rollout, both professions reported—often retrospectively—that physicians had received inadequate training, which repeatedly caused problems when using the software. By contrast, nurses described a smooth implementation and adoption process, with successful integration into routine clinical workflows. Overall, nurses stated that their expectations were largely met. Interviewed nurses highlighted the continuous availability and improved legibility of information, as well as the resulting reduction in the likelihood of errors. Additional reported advantages included clearly defined access roles and the possibility to use the EHR during ward rounds for immediate review and documentation. Nurses also identified remaining improvement needs, particularly the extension of interoperability with other clinical IT systems. In contrast, interviewed physicians repeatedly reported increased workload, insufficient customization to individual needs, and problems with EHR usability. Work conditions and burnout over time: survey results Table 1 reports means between baseline and follow-up for key outcome variables: Table 1.Key study variables over time and test for mean differences.BaselineFollow-upTest for significance (t-test)MSDMSD T, p WorkflowWorkflow interruptions3.42.733.28.791.15, .25Multitasking3.331.053.251.07.52, .60Information overload2.831.112.65.961.14, .26Burnout symptomsEmotional exhaustion4.111.144.021.10.48, .63Depersonalization2.08.962.231.07−.94, .35Abbreviations: df, degrees of freedom; M, mean; p, significance; SD, standard deviation. Key study variables over time and test for mean differences. Abbreviations: df, degrees of freedom; M, mean; p, significance; SD, standard deviation. For the overall group, we did not find significant changes over time. All 3 workflow variables as well as EE were evaluated at a slightly lower level at follow-up, whereas DP tended to be appraised higher. Concerning group-specific influences, within an additional step we then tested for potential interaction effects between both time points and both professions. The detailed results are depicted in Table SA-4. We identified 2 significant interaction effects that are displayed in Figure 1. Trajectories for information overload (left) and depersonalization (right) across professions and time-points. First, information overload was over time significantly lower appraisals of information overload for nursing professionals than in physicians (profession × time-point: F = 6.28, df = 1, P=.013; cf., Figure 1, left). Secondly, DP trajectories over time increased in physicians over time, whereas we observed lower DP in nurses at follow-up (F = 3.94, df = 1, P = .049, cf., Figure 1, right). Additionally, we observed that reported levels of workflow interruptions were in both waves substantially higher in physicians than in nurses (main effect for profession: F = 5.70, df = 1, P = .018). For Multitasking we identified an interaction that was close to our level of significance: over time, multitasking decreased in nurses whereas physicians reported higher multitasking at follow-up (F = 3.26, df = 1, P = .073). Discussion Given the increasing adoption of EHR in hospital care, concurrent evaluations into the effects on provider experiences and burnout over time are needed. Drawing upon provider reports before and after EHR implementation, this prospective study did not show overall changes over time. Although provider-rated levels of workflow interruptions, multitasking, and exhaustion were high, utilization of EHR did not alleviate workload and provider experiences. Our findings provide different, valuable contributions to our knowledge base on effects of EHR adoption in healthcare. First, we repeatedly captured information on course, depth, and status of the implementation process including pre-transition expectations and actual progress of the EHR introduction. Our observations corroborate findings of previous investigations that revealed contrasts between pre-implementation expectations and post-implementation experiences among affected healthcare professionals.7,8 Two key insights deserve further consideration: first, the actual implementation procedure differed across both professions. In nursing professionals, EHR implementation was carried out almost as planned with multiple components, structured plan, and close support on the ward. In contrast, physicians reported reliance on self-guided training via online resources. This contradicts numerous recommendations that emphasize work-based opportunities to gain proficiency in using EHR functionalities through IT support.13,16 The consistently lower appraisals of workflow conditions at follow-up among physicians (as observed in our prospective survey), might be partly explained that low physician involvement in EHR implementation, resulting in poorer functionalities and lower EHR workflow integration among physicians.7,13 Secondly, at follow-up, mixed experiences were reported, yet both professions stated profound changes in interprofessional collaboration and communication, with, for example, increased access to patient records and clinical information for nursing professionals. This resonates well with previous findings that altered documentation practices impact teamwork and coordination practices with changing roles and responsibilities.5,9,13 Secondly, after analyzing professionals’ evaluations before and after the EHR implementation, we did not observe major changes in levels of workflow (ie, workflow interruptions, multitasking, and information overload) as well as burnout symptoms (ie, EE and DP). This aligns well with the current, yet inconsistent research base, partly due to the scarcity of prospective designs that capture pre- and post-implementation status and thus allow reliable inferences concerning actual trajectories and changes over time.1–4 Nonetheless, our results may inform future studies that aim for broader consideration of work factors and their impact on provider stress during adoption of new digital technologies in patient care with particular emphasis on alignment with clinical workflow.11,13,14 Thirdly, a novel finding of our investigation was concurrent evaluation of provider experiences across 2 major health care professions, that is, physicians and nurses. We observed 2 moderation effects that showed specific effects for profession (nurses vs physicians) for 2 provider outcomes over time. First, in nursing professionals, information overload was significantly lower over time compared to physicians (almost the same, however insignificant observation also for multitasking, with lower levels at follow-up for nursing professionals). Additionally, DP increased in physicians over time, whereas we observed lower DP in nurses at follow-up. Although we cannot claim causality, this prospective change might be again attributed to physician-reported limitations and disappointment in usability and efficiency of EHR, mentioned in accompanying interviews.7,12 This has been repeatedly associated with lower satisfaction and acceptance as well as burnout.4,10,15,32 Moreover, previous investigations corroborated physician experiences of lower efficiencies and limited cognitive support after EHR-implementation, eventually leading to experiences of detachment or alienation.3,5,32 Finally, we observed consistently higher interruption levels in physicians compared to nurses. This underscores that observed workflow appraisals occur in a complex sociotechnical work system comprising multiple components (beyond the technologies being used) with genuine impact on workflows and patient care.33 Limitations Despite its various strengths, our findings should be carefully interpreted in the light of various limitations. Consistent with most existing evaluations of EHR implementation, we also relied on provider-reported outcomes and subjective measures of user experience rather than objective performance indicators.22 Moreover, the absence of matching providers between baseline and follow-up assessments limits the ability to attribute observed changes directly to in-person changes over time. Since we did not match identical providers between pre- and follow-up status, inferences should therefore acknowledge limited statistical power. Moreover, we could not obtain information on panel attrition and potential selection biases. Specifically, our response rates were different across both professions, indicating that different satisfaction levels with the EHR implementation may have influenced participation in evaluation. The time lag of approximately 5 months between measurements further complicates interpretation, as previous research has demonstrated that the timing of follow-up assessments is critical; challenges and inefficiencies in EHR use often persist for extended periods, even years after implementation.7 Future studies should seek for heterogeneous measures and data sources—ranging from EHR log data to retention and usage metrics.6,34 Consequently, it is not possible to draw definitive conclusions regarding the actual implications for patient safety or satisfaction.24 Concerning our interview data, we acknowledge that the interview duration and overall numbers of interviews were limited, what may limit the richness, scope, as well as depth of the information on the implementation status and fidelity. Moreover, the interviews should collect supplementary information and serve as an illustrative complement to the quantitative study by capturing employees’ lived prior expectation and post-hoc experiences. They do not aim to be exhaustive or to provide a deep, comprehensive analysis of all aspects of EHR implementation across all clinical departments. Finally, it is important to note that EHR adoption was at different implementation stages across the 3 participating clinical departments; however, due to data protection considerations, these details were not reported. Finally, external validity and generalizability of the findings might be restricted due to the specifics of the setting (academic hospital), EHR systems being implemented, limited sample sizes, as well as reporting bias. Implications for research and clinical practice Despite its various limitations, our findings may inform future research into differential effects on documentation burden and provider experiences across different healthcare professions. Upcoming investigations should investigate how documentation burden varies across medical and nursing professions in acute care by comparing workflow patterns, time allocation, and task complexity using prospective, mixed-methods approaches. Additionally, studies should examine how these differences shape provider experiences over time, including cognitive strain, job satisfaction, and burnout, to inform targeted, concurrent evaluations tailored to specific professional roles.35 Beyond the methodological limitations of our investigation, we acknowledge that our study did not fully incorporate a comprehensive framework capable of capturing the dynamic interplay among organizational, work, process, and provider-related factors in acute care settings. Although our approach was inspired by a sociotechnical work system perspective (ie, SEIPS), we acknowledge that our captured variables did not fully capture all relevant work system factors as well as potential burdensome and stressful consequences associated with EHR introduction.11,22 Future research should therefore seek to fully adopt integrative frameworks that account for the complexities of clinical work and health IT use, as well as their inherent consequences for both providers and patients.21,36 Concerning implications for practice, our findings may contribute to management of adoption processes and respective human-factors improvements to facilitate EHR workflow implementation in health care organizations.36 Our observation that EHR implementation did not meet entirely physicians’ needs was eventually associated with increased DP deserves careful consideration: burnout among hospital physicians is a serious public health concern because it undermines clinicians’ mental health, reduces work capacity, and can drive physicians to leave practice entirely. It also threatens patient care by impairing decision-making, communication, and the quality and safety of care delivery, while worsening already strained healthcare workforce shortages.37 Conclusion Scientific investigations and prospective evaluations on EHR adoption in hospitals are necessary to broaden our understanding how electronic systems affect provider workflows and outcomes. Moreover, investigations across different professions also provide valuable and evidence-based insights that may guide future implementation decisions. This prospective, multi-methods investigation studies the trajectories of perceived workflow interruptions, multitasking, information overload, and burnout over time among physicians and nurses in course of EHR adoption. Supplementary Material Acknowledgments Data collection has been part of the Doctoral thesis requirements of Ms. Katharina Schmidt at Medical Faculty, Ludwig-Maximilians University, Munich, Germany. We thank all healthcare professionals for their participation in this study. This publication was supported by the Open Access Publication Fund of the University of Bonn. Contributor Information Matthias Weigl, Institute for Patient Safety, University Hospital, University of Bonn, 53127 Bonn, Germany; Institute and Clinic for Occupational, Social and Environmental Medicine, LMU University Hospital, Munich, 80335, Germany. Karin Berger, Department of Medicine III, LMU University Hospital, Ludwig-Maximilians-University Munich, Munich, 81377, Germany; Comprehensive Cancer Center (CCC Munich LMU), LMU University Hospital, Ludwig-Maximilians-University of Munich, Munich, 81377, Germany; Bavarian Cancer Research Center (BZKF), Munich, Germany; German Cancer Consortium (DKTK), Munich, Germany. Prof. Julia Mayerle, Bavarian Cancer Research Center (BZKF), Munich, Germany; Department of Medicine II, LMU University Hospital, Ludwig-Maximilians-University Munich, Munich, 81377, Germany. Prof. Jürgen Behr, Department of Medicine V, LMU University Hospital, Comprehensive Pneumology Center, German Center for Lung Research (DZL), LMU, Munich, 81377, Germany. Prof. Michael von Bergwelt-Baildon, Department of Medicine III, LMU University Hospital, Ludwig-Maximilians-University Munich, Munich, 81377, Germany; Comprehensive Cancer Center (CCC Munich LMU), LMU University Hospital, Ludwig-Maximilians-University of Munich, Munich, 81377, Germany; Bavarian Cancer Research Center (BZKF), Munich, Germany; German Cancer Consortium (DKTK), Munich, Germany. Prof. Helmut Ostermann, University Hospital LMU, Munich, 81377, Germany. Katharina Schmidt, Department of Radiation Oncology, Barmherzige Brüder Hospital Regensburg, Regensburg, 93049, Germany. Author contributions Matthias Weigl (Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Visualization, Writing—original draft, Writing—review & editing), Karin Berger (Conceptualization, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Writing—original draft, Writing—review & editing), Julia Mayerle (Conceptualization, Funding acquisition, Resources, Writing—review & editing), Jürgen Behr (Conceptualization, Funding acquisition, Resources, Writing—review & editing), Michael von Bergwelt-Baildon (Conceptualization, Funding acquisition, Resources, Writing—review & editing), Helmut Ostermann (Conceptualization, Funding acquisition, Investigation, Project administration, Supervision, Writing—review & editing), and Katharina Schmidt (Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Validation, Writing—original draft, Writing—review & editing) Supplementary material Supplementary material is available at JAMIA Open online. Funding Institutional budget, no external funding. Conflicts of interest The authors declare that they have no competing interests. Data availability Anonymized data can be obtained from the corresponding authors upon reasonable request. Ethics approval and consent to participate Approval was obtained by the Ethics Committee of the Medical University of LMU University, Munich, Germany (registry code: 21-0142). Declaration of generative AI in scientific writing During the preparation of this work, the authors used ChatGPT to check for language and to optimize readability and wording of the manuscript. After using this tool, the authors reviewed and edited the content as required and take full responsibility for the content of the publication. References Associated Data Supplementary Materials Data Availability Statement Anonymized data can be obtained from the corresponding authors upon reasonable request.