Hartmuth Nowak

Dr. Hartmuth Nowak


Research Focus & Metrics

Academic Career Length

11Years (2016-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
10 December 2020 in BMJ
Verified
Journal Article
Open access
PDF file
HNHartmuth Nowak
Hartmuth Nowak
NZNina Zech
Nina Zech
SASven Asmussen
Sven Asmussen
TRTim Rahmel
Tim Rahmel
MTMichael Tryba
Michael Tryba
EHErnil Hansen
Ernil Hansen
Objective To investigate the effect of therapeutic suggestions played to patients through earphones during surgery on postoperative pain and opioid use. Design Blinded randomised controlled study. Setting Five tertiary care hospitals in Germany. Participants 385 of 400 patients consecutively recruited from January to December 2018 who were to undergo surgery for 1-3 hours under general anaesthesia. In the per protocol analysis 191 patients were included in the intervention group and 194 patients in the control group. Intervention The intervention comprised an audiotape of background music and positive suggestions based on hypnotherapeutic principles, which was played repeatedly for 20 minutes followed by 10 minutes of silence to patients through earphones during general anaesthesia. Patients in the control group were assigned to a blank tape. Main outcome measures The main outcome was dose of opioid administered by patient controlled analgesia or nurse controlled analgesia within the first postoperative 24 hours, based on regular evaluation of pain intensity on a numerical rating scale (range 0-10, with higher scores representing more severe pain). Results Compared with the control group, the intervention group required a significantly (P=0.002) lower opioid dose within 24 hours after surgery, with a median of 4.0 mg (interquartile range 0-8) morphine equivalents versus 5.3 (2-12), and an effect size (Cohen’s d) of 0.36 (95% confidence interval 0.16 to 0.56). The number of patients who needed opioids postoperatively was significantly (P=0.001) reduced in the intervention group: 121 of 191 (63%, 95% confidence interval 45% to 70%) patients in the intervention group versus 155 of 194 (80%, 74% to 85%) in the control group. The number needed to treat to avoid postoperative opioids was 6. Pain scores were consistently and significantly lower in the intervention group within 24 hours after surgery, with an average reduction of 25%. No adverse events were reported. Conclusions Therapeutic suggestions played through earphones during general anaesthesia could provide a safe, feasible, inexpensive, and non-drug technique to reduce postoperative pain and opioid use, with the potential for more general use. Based on the finding of intraoperative perception by a considerable number of patients, surgeons and anaesthetists should be careful about background noise and conversations during surgery. Trial registration German Clinical Trial Register DRKS00013800.

3 February 2025 in Journal of Medical Internet Research
Verified
Journal Article
Open access
GGGodwin Denk Giebel
Godwin Denk Giebel
PRPascal Raszke
Pascal Raszke
HNHartmuth Nowak
Hartmuth Nowak
LPLars Palmowski
Lars Palmowski
MAMichael Adamzik
Michael Adamzik
NBNikola Blase
Nikola Blase
Background: Digitalization is currently revolutionizing health care worldwide. A promising technology in this context is artificial intelligence (AI). The application of AI can support health care providers in their daily work in various ways. The integration of AI is particularly promising in clinical decision support systems (CDSSs). While the opportunities of this technology are numerous, the problems should not be overlooked. Objective: This study aimed to identify challenges and barriers in the context of AI-based CDSSs from the perspectives of experts across various disciplines. Methods: Semistructured expert interviews were conducted with different stakeholders. These included representatives of patients, physicians and caregivers, developers of AI-based CDSSs, researchers (studying AI in health care and social and health law), quality management and quality assurance representatives, a representative of an ethics committee, a representative of a health insurance fund, and medical product consultants. The interviews took place on the web and were recorded, transcribed, and subsequently subjected to a qualitative content analysis based on the method by Kuckartz. The analysis was conducted using MAXQDA software. Initially, the problems were separated into “general,” “development,” and “clinical use.” Finally, a workshop within the project consortium served to systematize the identified problems. Results: A total of 15 expert interviews were conducted, and 309 expert statements with reference to problems and barriers in the context of AI-based CDSSs were identified. These emerged in 7 problem categories: technology (46/309, 14.9%), data (59/309, 19.1%), user (102/309, 33%), studies (17/309, 5.5%), ethics (20/309, 6.5%), law (33/309, 10.7%), and general (32/309, 10.4%). The problem categories were further divided into problem areas, which in turn comprised the respective problems. Conclusions: A large number of problems and barriers were identified in the context of AI-based CDSSs. These can be systematized according to the point at which they occur (“general,” “development,” and “clinical use”) or according to the problem category (“technology,” “data,” “user,” “studies,” “ethics,” “law,” and “general”). The problems identified in this work should be further investigated. They can be used as a basis for deriving solutions to optimize development, acceptance, and use of AI-based CDSSs. International Registered Report Identifier (IRRID): RR2-10.2196/preprints.62704

7 July 2025 in JMIR Medical Informatics
Verified
Journal Article
Open access
GGGodwin Denk Giebel
Godwin Denk Giebel
PRPascal Raszke
Pascal Raszke
HNHartmuth Nowak
Hartmuth Nowak
LPLars Palmowski
Lars Palmowski
MAMichael Adamzik
Michael Adamzik
NBNikola Blase
Nikola Blase
Background: Artificial intelligence (AI)–based systems are receiving increasing attention in the health care sector. While the use of AI is well advanced in some medical applications, such as image recognition, it is still in its infancy in others, such as clinical decision support systems (CDSS). Examples of AI-based CDSS can be found in the context of sepsis prediction or antibiotic prescription. Scientific literature indicates that such systems can support physicians in their daily work and lead to improved patient outcomes. Nevertheless, there are various problems and barriers in this context that should be considered. Objective: This study aimed to identify opportunities to optimize AI-based CDSS and their integration into care from the perspective of experts. Methods: Semistructured web-based expert interviews were conducted. Experts representing the perspectives of patients; physicians; caregivers; developers; health insurance representatives; researchers (especially in law and IT); and experts in regulation, market admission and quality management or assurance, and ethics were included. The conversations were recorded and transcribed. Subsequently, a qualitative content analysis was performed. The different approaches to improvement were categorized into groups (“technology,” “data,” “users,” “studies,” “law,” and “general”). These also served as deductive codes. Inductive codes were determined within an internal project workshop. Results: In total, 13 individual and 2 double interviews were conducted with 17 experts. A total of 227 expert statements were included in the analysis. Suggestions were heterogeneous and concerned improvements: (1) in the systems themselves (eg, implementing comprehensive system training involving [future] users; using a comprehensive and high-quality database; considering usability, transparency, and customizability; preventing automation bias through control mechanisms or intelligent design; conducting studies to demonstrate the benefit of the system), (2) on the user side (eg, training [future] physicians could contribute to a more positive attitude and to greater awareness and questioning decision supports suggested by the system and ensuring that the use of the system does not lead to additional work), and (3) in the environment in which the systems are used (eg, increasing the digitalization of the health care system, especially in hospitals; providing transparent public communication about the benefits and risks of AI; providing research funding; clarifying open legal issues, eg, those related to liability; and standardizing and consolidating various approval processes). Conclusions: This study offers several possible strategies for improving AI-based CDSS and their integration into health care. These were found in the areas of “technology,” “data,” “users,” “studies,” “law,” and “general.” Systems, users, and the environment should be taken into account to ensure that the systems are used safely, effectively, and sustainably. Further studies should investigate both the effectiveness of strategies to improve AI-based CDSS and their integration into health care and the accuracy of their match to specific problems. International Registered Report Identifier (IRRID): RR2-10.2196/62704

13 July 2024 in Annals of Intensive Care
Verified
Journal Article
Open access
PDF file
LPLars Palmowski
Lars Palmowski
SLSimone Lindau
Simone Lindau
LHLaura Contreras Henk
Laura Contreras Henk
BMBritta Marko
Britta Marko
AWAndrea Witowski
Andrea Witowski
Tim RahmelTim Rahmel
Tim Rahmel
In sepsis, initial resuscitation with fluids is followed by efforts to achieve a negative fluid balance. However, patients with sepsis-associated acute kidney injury (SA-AKI) often need diuretic or renal replacement therapy (RRT). The dilemma is to predict whether early RRT might be advantageous or diuretics will suffice. Both the Furosemide Stress Test (FST) and measurements of the urinary biomarkers TIMP-2*IGFBP-7, if applied solely, do not provide sufficient guidance. We tested the hypothesis that a combination of two tests, i.e., an upstream FST combined with downstream measurements of urinary TIMP-2*IGFBP-7 concentrations improves the accuracy in predicting RRT necessity. In this prospective, multicenter study 100 patients with sepsis (diagnosed < 48h), AKI stage ≥ 2, and an indication for negative fluid balance were included between 02/2020 and 12/2022. All patients received a standardized FST and urinary biomarkers TIMP-2*IGFBP-7 were serially measured immediately before and up to 12 h after the FST. The primary outcome was the RRT requirement within 7 days after inclusion. 32% (n = 32/99) of SA-AKI patients eventually required RRT within 7 days. With the FST, urine TIMP-2*IGFBP-7 decreased within 2 h from 3.26 ng2/mL2/1000 (IQR: 1.38–5.53) to 2.36 ng2/mL2/1000 (IQR: 1.61–4.87) in RRT and 1.68 ng2/mL2/1000 (IQR: 0.56–2.94) to 0.27 ng2/mL2/1000 (IQR: 0.12–0.89) and non-RRT patients, respectively. While TIMP-2*IGFBP-7 concentrations remained low for up to 12 h in non-RRT patients, we noted a rebound in RRT patients after 6 h. TIMP-2*IGFBP-7 before FST (accuracy 0.66; 95%-CI 0.55–0.78) and the FST itself (accuracy 0.74; 95%-CI: 0.64–0.82) yielded moderate test accuracies in predicting RRT requirement. In contrast, a two-step approach, utilizing FST as an upstream screening tool followed by TIMP-2*IGFBP-7 quantification after 2 h improved predictive accuracy (0.83; 95%-CI 0.74–0.90, p = 0.03) compared to the FST alone, resulting in a positive predictive value of 0.86 (95%-CI 0.64–0.97), and a specificity of 0.96 (95%-CI 0.88–0.99). The combined application of an upstream FST followed by urinary TIMP-2*IGFBP-7 measurements supports highly specific identification of SA-AKI patients requiring RRT. Upcoming interventional trials should elucidate if this high-risk SA-AKI subgroup, identified by our predictive enrichment approach, benefits from an early RRT initiation.

13 October 2020 in Cells
Verified
Journal Article
Open access
PDF file
BKBjörn Koos
Björn Koos
EMEva Lotta Moderegger
Eva Lotta Moderegger
KRKatharina Rump
Katharina Rump
HNHartmuth Nowak
Hartmuth Nowak
KWKatrin Willemsen
Katrin Willemsen
Tim RahmelTim Rahmel
Tim Rahmel
Mitochondrial DNA (mtDNA) plays a vital role as a damage-associated molecular pattern in sepsis being able to shape the immune response. Since pathogen recognition receptors of innate immune cells are activated by demethylated DNA only, we set out to investigate the amount of DNA methyltransferase 1 (DNMT1) in mitochondria and the extent of mtDNA methylation in a human endotoxin model. Peripheral blood mononuclear cells of 20 healthy individuals were isolated from whole blood and stimulated with lipopolysaccharide (LPS) for 48 h. Subsequently, DNMT1 protein abundance was assessed in whole cells and a mitochondrial fraction. At the same time, methylation levels of mtDNA were quantified, and cytokine expression in the supernatant was measured. Despite increased cellular expression of DNMT1 after LPS stimulation, the degree of mtDNA methylation slightly decreased. Strikingly the mitochondrial protein abundance of DNMT1 was reduced by 50% in line with the lower degree of mtDNA methylation. Although only modest alterations were seen in the degree of mtDNA methylation, these strongly correlated with IL-6 and IL-10 expression. Our data may hint at a protein import problem for DNMT1 into the mitochondria under LPS stimulation and suggest a role of demethylated mtDNA in the regulation of the inflammatory immune response.

28 August 2025 in Infection
Verified
Journal Article
Open access
PDF file
TBThilo Bracht
Thilo Bracht
MWMaike Weber
Maike Weber
KKKerstin Kappler
Kerstin Kappler
LPLars Palmowski
Lars Palmowski
MBMalte Bayer
Malte Bayer
HNHartmuth Nowak
Hartmuth Nowak
Background: Sepsis therapy is still limited to treatment of the underlying infection and supportive measures. To date, various sepsis subtypes were proposed, but therapeutic options addressing the molecular changes of sepsis were not identified. With the aim of a future individualized therapy, we used machine learning (ML) to identify clinical phenotypes and their temporal development in a prospective, multicenter sepsis cohort and characterized them using plasma proteomics. Methods: Routine clinical data and blood samples were collected from 384 patients. Sepsis phenotypes were identified based on clinical measurements and plasma samples from 301 patients were analyzed using mass spectrometry. The obtained data were evaluated in relation to the phenotypes, and supervised ML models were developed enabling prospective phenotype classification. Results: Three sepsis phenotypes were identified. Cluster C was characterized by the highest disease severity and multi-organ failure with leading liver failure. Cluster B showed relevant organ failure, with renal damage being particularly prominent in comparison to cluster A. Time course analysis showed a strong association of cluster C with mortality, while patients in cluster B were likely to change the cluster until day 4. The plasma proteome reflected the clinical features of the phenotypes and showed gradual consumption of complement and coagulation factors with increasing sepsis severity. Supervised ML models allowed the assignment of patients based on only seven widely available features (alanine transaminase (ALT), aspartate transaminase (AST), base excess (BE), international normalized ratio of thrombin time (INR), diastolic arterial blood pressure, systolic arterial blood pressure (BPdia, BPsys) and activated partial thromboplastin time (aPTT)). Conclusions: The identified clinical phenotypes reflected varying degrees of sepsis severity and were mirrored in the plasma proteome. Proteomic profiling offered novel insights into the molecular mechanisms underlying sepsis and enabled a deeper characterization of the identified phenotypes.