An acute MI is an event of heart muscle necrosis caused by insufficient blood and oxygen supply. The incidence of acute MI or death is approximately 805,000 people in the United States alone, and accounts for approximately 20% of all causes of death [23]. Acute MI is classified based on the presence or absence of ST-segment elevation (ST elevation MI [STEMI] vs non–ST elevation MI [NSTEMI]) on electrocardiogram, while serial troponin levels differentiate a non-ST segment MI from unstable angina. Treatment includes coronary reperfusion with percutaneous coronary intervention, intravenous fibrinolytic therapy, or surgery. Depending on the bleeding risk and other patient factors, different antithrombotic drug combinations exist and are recommended in the guidelines. Further management focuses on risk reduction, including lipid-lowering drugs such as statins [24].

How many patients were admitted with a new diagnosis of an NSTEMI or STEMI, received an interventional coronary revascularization therapy, and which antithrombotic and lipid-lowering drugs were prescribed?

We identified 730 patients in 2022 with a diagnosis of acute MI. Of these, 359 patients were coded as STEMI or received a coronary intervention (Figure 2). Antiplatelet therapy was started in 71.6% (257/359), with most being on dual-antiplatelet therapy, while partly single antiplatelet therapy was administered in combination with existing anticoagulation therapy (not shown). Independent of antiplatelet therapy, most patients received a statin (227/359, 63.2%), with only a minority also receiving ezetimibe (97/359, 27%). Taken together, 72.7% (261/359) of patients received guideline-recommended therapy. Meanwhile, it should be appreciated that additional patients were considered for therapy, yet not administered due to contraindications, or received bridging therapy.

Platelet inhibitors and statins play an important role in the prevention of MI and mortality. Implementation of a clinical dashboard highlighting patients not receiving all recommended therapies can help to initiate these cost-efficient drugs and substantially improve patient care.

Ischemic stroke is a cardiovascular disease estimated to affect approximately 5.5 million patients across the globe in 2016. With approximately 116 million disease-adjusted life years lost globally, it carries a high mortality and morbidity [25]. A stroke occurs when a circulatory block results in a lack of oxygen delivery thus resulting in the subsequent death of brain tissue. In the case of an ischemic stroke, the main treatment is the resolution of the blood clot, either by drugs or mechanically by catheterization. The time from symptom onset to the resolution of the clot is one of the most important predictors of outcomes and an important indicator of improved patient care [26]. Huge efforts are made to reduce the time of the rescue chain to a minimum (“time is brain”). Along this way, the reduction of the time from presentation in the emergency room to initiation of treatment, that is, thrombolysis (“Door to Needle”) or thrombectomy (“Door to Groin”), is an important contribution of the intrahospital treatment [26]. Even though the outcome after acute ischemic stroke is generally assessed after 90 days, early neurological improvement is a surrogate indicator for a 90-day outcome. It is often defined as the percentage change in the National Institutes of Health Stroke Scale (NIHSS) score from baseline to 24 hours thereafter [27].

Are there fluctuations in the process quality, as measured by the waiting time between emergency room arrival and treatment application from the time of arrival in the emergency room until treatment? Does the time between the onset of symptoms until the start of a procedure correlate with early neurologic improvement?

In the study period, we identified 1277 patients with a stroke diagnosis treated in the department of neurology, of which 1088 were ischemic and 189 were hemorrhagic. Out of 165 patients treated with intravenous thrombolysis, 143 (86.7%) received this therapy within the guideline-recommended time window of less than 60 minutes (Figure 3). Similarly, we identified 108 patients who received thrombectomy, of which 82 (75.9%) were within the guideline-recommended time window of 90 minutes. For both therapies, our cohort of a referral stroke center has a significant number of patients presenting in an extended time window and with the need of previous coagulation testing.

Furthermore, we correlated the time from symptom onset (witnessed onset or last known-well time) to intervention with early neurological improvement (NIHSS score) from 157 patients receiving thrombolysis and 101 receiving thrombectomy (Figures 3B and 3D). Among the patients receiving thrombolysis, 63.3% (101/157) displayed a decrease in NIHSS score (ie, clinical improvement), 15.9% (25/157) showed no change, and 21.7% (31/157) exhibited an increase in NIHSS score. Among the patients who received thrombectomy, 63% (48/76) demonstrated a decrease in NIHSS score, while 7% (5/76) showed no change, and 30% (23/76) exhibited an increase in NIHSS score. In total, 25 patients were excluded from the analysis due to ongoing weaning from mechanical ventilation after thrombectomy.

Monitoring of clinical workflows can identify cases with longer-than-recommended turnaround times. While many reasons for delays exist, for example, need of previous coagulation assessment in patients on oral anticoagulation or need of perfusion imaging in extended-time windows, and not all may be modifiable, some can potentially be addressed, and workflows adjusted. Structured analysis of this data is the first step for improving the care of future patients.

In all figures, patients who underwent a single therapy (either thrombolysis in Figures 3A and 3B or thrombectomy in Figures 3C and 3D) are color-coded in blue, and patients who received a dual therapy (both thrombolysis and thrombectomy) in red.

Diabetes mellitus is a metabolic disorder with a high disease burden affecting 540 million people worldwide [28]. In hospitals, diabetes is one of the most common but undetected comorbidities increasing the length of stay and risk for nosocomial complications [29]. Identifying dysglycemia is a critically necessary step to introduce affected patients to dedicated in-hospital diabetes management and offer primary prevention strategies to those with prediabetes [29].

How many inpatients are affected by dysglycemia and are cases identified appropriately?

In total, 443,866 serum glucose and 16,180 glycated hemoglobin A_1c (HbA_1c) readings from 35,494 unique patients in 2022 were found. Further, 13,887 (39%) individuals, out of 35,494, were identified as having at least one pathologic glycemic finding. Overall, 30.5% (10,835/35,494) of the total cohort were at least hyperglycemic, while the prevalence of prediabetes and diabetes was 11.8% (4176/35,494) and 15.8% (5601/35,494; of which 3270 were identified by serum glucose, 941 by HbA_1c levels, and 1390 by both criteria), respectively (Figure 4A). Per definition, mean serum glucose and HbA_1c levels discriminated patients with normoglycemia from those with dysglycemia with a rise of mean glucose from 103.1 mg/dL to 112.3 mg/dL and 130.9 mg/dL to 167.1 mg/dL, respectively, and HbA_1c from 5.6% to 5.7% and 6% to 7.2%, respectively, in patients with normoglycemia, hyperglycemia, prediabetes, and diabetes (Figures 4B and 4D). ICD-10 code documenting case identification for dysglycemia was found in 44.2% (6138/13,887) cases among those with biochemically proven hyperglycemia, prediabetes, or diabetes. Consequently, 55.8% (7749/13,887) of patients with dysglycemia would not have been identified by ICD-10 coding alone (Figure 4C).

A dashboard visualizing in-hospital glucose metabolism data improves the dysglycemia detection rate. Future automated communication of identified cases to health care professionals may improve inpatient outcome by simplifying access by an algorithm-based referral to specialized diabetes teams as the first step of a digitalized in-hospital diabetes management program triggering dedicated diabetes care.

Sepsis is a clinically heterogeneous syndrome defined by a dysregulated host response to infection resulting in life-threatening (multi-)organ dysfunction [30]. Its primary treatment is the timely administration of broad-spectrum antibiotics. Meanwhile, with an ever-increasing prevalence of multi–drug-resistant pathogens, it is paramount to treat as targeted and shortly as possible, to avoid the selection of further resistant strains. To this end, hospital-level insights into current antibiotic use, isolated pathogens, and the prevalence of multi–drug-resistant pathogens are critical to tailoring antibiotic practice to local conditions.

Which antibiotics are currently used in patients with sepsis? Which pathogens are isolated from urine and blood cultures and how do the inflammatory biomarkers, that is, C-reactive protein and procalcitonin, change during therapy?

We identified a total of 1803 patients with a sepsis diagnosis in 2022. Out of 213 distinct isolated pathogen strains, the top 10 most frequently isolated pathogen included Staphylococcus epidermidis (773/2672, 28.9%), a frequent contaminant with low pathogenic potential, followed by Escherichia coli (479/2672, 17.9%) and Enterococcus faecium (348/2672, 13%) strains (Figure 5A). Of the 10 most prescribed antibiotic piperacillin and tazobactam (593/1593, 37.2%), a broad-spectrum penicillin, followed by carbapenems (219/1593, 13.7%) and glycopeptides (eg, vancomycin, 186/1593, 11.7%) were most common (Figure 5B). For the clinical dashboard, we also provide a patient-level view showing the change in inflammatory markers, isolated pathogens, and antibiotics prescribed (Figure 5C).

The dashboard provides a summary of the evolution of inflammatory markers and antibiotic prescriptions. It can help to identify changes in the causal pathogens and monitor antibiotic prescriptions to avoid harmful over- or undertreatment.

PC is the most common solid cancer in men in most of the Western World [31]. Various treatment options exist to achieve prolonged disease-free periods and cures. Important predictors for survival include the Tumor, Node, Metastasis stage and pathological features (International Society for Urological Pathology [ISUP] Gleason Grade Groups), which are summarized as a clinical stage (I-IV) as well as synchronous versus metachronous metastases. To assist in the diagnosis and monitor disease progression, prostate-specific antigen (PSA) is commonly used.

What is the current treatment landscape for PC by ISUP grades? What is the evolution of PSA levels before and after treatments depending on surgical resection status?

We identified a total of 269 patients with newly diagnosed PC in 2021 who received treatment between 2021 and 2022. Among these, localized tumors (T1N0M0) with low ISUP scores were most common at biopsy. The most common treatment approach was active surveillance (n=29, 72.5%), followed by surgery (n=8, 20%; Figure 6A). For 54 patients treated with radical prostatectomy, PSA values were monitored longitudinally, identifying 3 patients with PSA persistence or biochemical recurrence (Figure 6B).

Increases in PSA values after treatment indicate disease relapse or progression. Especially in the context of easy access to sensitive PSA tests, earlier detection of biochemical recurrence is warranted. Identification and monitoring can streamline treatment approaches, potentially improving patient outcomes by initiating early salvage therapy. Furthermore, with recent improvements in staging through molecular imaging with prostate-specific membrane antigen positron emission tomography or computed tomography (PSMA-PET/CT) and novel combination treatment regimens, centralized monitoring of shifts in routine practice is an essential first step to draw conclusions about the real-world effectiveness of these approaches.