BETA.HEALTH innovation grants aim to boost innovation projects at Danish hospitals.

Over 15 percent of the Danish population suffers from sleep-related conditions, such as sleep apnea and narcolepsy. These conditions are diagnosed using polysomnography (PSG), a mostly manual process for which current need exceeds capacity. U-Sleep, a highly accurate machine-learning model, automatically detects and assigns initial scores to sleep stages in PSG data. This stands to reduce hands-on time by 83–89 percent, requiring an average of just 20 minutes for spot-checking and validating instead of hours to manually score each test.

Intranasal medicine plays a crucial role in treating chronic rhino sinusitis and hay fever. Although using nasal spray seems straightforward, most people aim the nozzle incorrectly. This prevents the medicine from reaching the area of the nasal cavity where it can alleviate symptoms – and can also lead to complications like nosebleeds and septal perforations. Rhinomate adds a simple aiming mechanism to standard nasal spray bottles to help ensure correct administration.

Primary brain tumours are among the most challenging forms of cancer to diagnose and treat. The subtype ‘glioblastoma’ (GBM) is the most predominant – and also among the most inaccessible for diagnostic biopsy. The project team is validating a novel biomarker that could support clinicians in evaluating treatment response and whether there's tumour regrowth – and enable earlier identification of glioma patients likely to have GBM.

Functional disorders and health anxiety affect 2% and 10% of the population, respectively. 'Digital psychological treatment' (DPT), which allows patients to access treatment online with the support of a remote therapist, is proven as effective as face-to-face treatment for these conditions. Compared to traditional methods, DPT is more cost-effective, scalable and has a broader reach. The project team has developed a digital platform with 10 DPT applications targeting various patient groups.

Lithium is the first-line treatment choice for patients with bipolar disorder. To ensure patient safety, the lithium blood concentration must be checked regularly and precisely 12 hours after the patient's last dose – and maintained within a narrow interval. This method is often inaccurate due to incorrect timing. The project team has developed a formula to accurately calculate this concentration outside of the 12-hour window.

This project aims to solve capacity issues related to the management of motor symptoms in patients with Parkinson's disease. The digital platform combines sensor-based motion capture with algorithms to automate data analysis. This enables technical staff to perform the multiple follow-ups required per yer throughout each patient's lifetime. This frees up neurologists to focus on other tasks while the system's data analysis capabilities enable clinicians to tailor medication doses with greater precision.

Chronic kidney disease (CKD) is associated with systemic acid retention, which begins slowly and is undetectable in blood samples during its early stages. By the time it’s detectable, the resulting kidney damage will be progressing rapidly. The team is developing a device that can perform a urine analysis that measures acid retention long before it can be detected by a blood test, which could help to stabilise the course of the condition.

Chronic kidney disease affects 850 million people worldwide. Many will undergo surgical construction of an arteriovenous fistula (AVF) necessitating complicated self-monitoring to avoid permanent damage from unnoticed dysfunctions. 30–50% of all AVFs become permanently damaged due to unnoticed dysfunctions, requiring construction of a new AVF. This project team has developed a wearable device to automatically monitor AVF function.

Emergency departments are one of the most unpredictable areas of hospital operations. Hospitals struggle to anticipate optimal staffing levels, mostly basing their decisions on intuition and a loose reliance on past patterns. Praemostro.com is a machine-learning-enabled system that can reliably forecast demand in high-variability settings such as hospital emergency departments, thereby improving management’s ability to schedule adequate staff.

Oral cavity cancer is the second most common head and neck cancer in Denmark. It can be difficult to accurately identify a tumor's boundaries during surgery, and inadequate surgical margins for oral cancer are estimated to be as high as 44 percent. 3Sonic enables surgeons to immediately evaluate the surgical margins of resected tumors using a 3D ultrasound visualisation. This improved precision can reduce the need for repeated treatment and significantly improve patient outcomes.

Infertility affects up to 20 percent of couples trying to conceive. Current go-to treatments for infertility rely on extensive tests, hormone stimulation and invasive procedures – all of which are physically and emotionally exhausting as well as costly. BloomBaby is a bridge between infertility and invasive clinical treatments, offering an alternative to ‘suffering with infertility’ or ‘suffering from fertility treatment’. It gives couples a viable chance of conceiving naturally.

Immunodeficient children are currently treated with allogenic hematopoietic stem cell (HSC) transplants, which introduce severe and potentially life-threatening complications that persist throughout their lifetime. This project aims to use the patient's own gene-edited HSC to circumvent the need for a stem cell donor, and mitigate the severe side effects associated with allogenic transplantation.

Major depressive disorder (MDD) affects over 300,000 people in Denmark each year – about 10% of the population. Current first-line treatments, which include psychopharmacology and psychotherapeutics, are only effective for one-third of patients.  The Nexus headband leverages recent advances in neuroimaging and neuromodulation technology, aiming to bring effective and scalable neuromodulation therapies and objective health biomarkers to patients and clinicians.

Endoscopic evaluation of the severity of disease in ulcerative colitis patients is subjective, resulting in significant variance between doctors – and even between evaluations by the same doctor. ENACT is an AI-enabled system that analyses live video frames of endoscopic procedures, providing real-time classification scores of the disease severity with 89% accuracy. This improves the speed and certainty with which doctors can evaluate each case, enabling early intervention and optimised treatment.

BED is a serious condition that affects up to 300,000 people in Denmark, and it's considered to be the most widespread eating disorder in the Western world. This project will conduct the largest BED treatment study to date – and the first to compare individual, group and technology-based psychotherapy interventions. The end goal is to develop four treatment manuals and establish their efficacy for various subgroups.

Up to 30% of hospitalised patients experience severe complications that increase their risk of permanent injury and death. The current standard of manual NEWS-based monitoring processes places unnecessary demands on clinicians' time without improving outcomes. WARD-CSS saves time while providing 24/7 vital-sign monitoring that improves the standard of care – and makes patients feel safer.

Each year, 5,000 people in Denmark suffer out-of-hospital cardiac arrest (OHCA) – 4,400 of them do not survive. While the number of registered AEDs in Denmark has increased over the past 20 years, only 8.7% of patients are defibrillated before an ambulance arrives. That's because most incidents happen at home, where an AED isn't available. This project aims to produce an affordable and easy-to-use AED specifically for use in private homes.

Diabetes is on the rise, and with it, cardiac autonomic neuropathy (CAN) – a severe and possibly fatal manifestation of dysfunction of the autonomic nervous system. The ProCVT Smart-Sheet is a wireless ECG monitor intended for sleep recordings, which reduces the influence of external factors while helping clinicians to detect CAN and quantify the severity of dysautonomia.

Half of the 30,000 patients admitted to Danish ICUs each year require mechanical ventilation because they can’t breathe on their own. For patients who are expected to require mechanical ventilation for an extended period, a tracheostomy is often placed. trachAI uses performant explainable AI software to help physicians assess the likelihood that a patient will require mechanical ventilation for longer than a week.

80% of surgically treated middle-ear pathologies are caused by Eustachian Tube Dysfunction (ETD), which is often diagnosed only after complications have developed. This new medical device facilitates quicker and easier diagnosis of ETD by streamlining the testing process and using an algorithm to interpret and summarize the data.

Organ transplants save many lives but the body can attack the foreign organ, resulting rejection. This project uses organ-derived cfDNA to directly measure organ damage, which the team hopes will help to predict organ rejection much earlier and with higher precision than other known markers – enabling better optimisation of immunosuppressive therapy.

This automatic dose-planning solution consists of software-driven, site-specific ‘treatment templates’ that automate a large part of the standard radiotherapy planning process associated with cancer treatment. This frees up time and resources, enabling treatment planners to focus on particularly challenging cases.

Involuntary muscular activity causes repeated displacement of ultrasound probes during percutaneous heart valve procedures, causing stress and poor ergonomics for cardiologists and prolonged anesthetic exposure for patients. EchoVice essentially acts as a 'third hand' to stabilise the probe during these procedures.

Gait training is a fact of life for many patients with orthopedic or neurological conditions, but it's often tedious. HearWalk captures and synthesises motion-related data into real-time feedback that’s communicated via auditory stimuli, which has been shown to increase focus and improve patient compliance.

The Novel Energy Harvester System Technology for Self-Powering Intracardiac Pacemakers (NEST) creates a new generation of pacemakers that charge themselves through 'motion energy' every time the heart muscle moves.

New technology can enable surgeons to visualise complex anatomical structures while preparing for surgery. This project aims to bring 3D technology to clinicians who can benefit from a 3D representation of certain anatomy.

The Solnedgangkassen is designed to enable more patients to receive palliative care at home instead of in hospital. The box contains a standardized assortment of the necessary medicine, instructions and information to help caregivers provide a high standard of in-home care for patients in the late-terminal phase.

Pathology AI aims to address the future shortage of pathologists by simplifying and speeding up the diagnostic process using digital image analysis. In this project, machine vision observes the same area as the pathologist, accumulating 'experience' based on the pathologist’s input. 

The TOLAC App uses data-driven software to mine these extensive datasets and create personalised risk profiles for women in specific clinical situations to better predict the probability of a successful vaginal birth after a prior caesarean delivery.

The fMRI fingerprint is an optical fiber technology that allows electrostimulation during fMRI scans. This tool help will help clinicians predict the patient’s response to expensive therapies.

DoseTracker is a real-time software-enabled tool that monitors the dose of radiation given to moving anatomy during cancer treatments. It calculates and compares the administered radiation dose with the planned dose and provides a complete overview of clinically relevant metrics to help clinicians make informed decisions.

One in four people has fatty liver disease without knowing it – and this condition can be asymptomatic for fifteen to thirty years. Evido’s algorithm-based solution can detect advanced fatty liver disease with 87% accuracy based on a series of standard blood samples that are already widely used in primary care today. 

Dermloop offers AI-augmented training, clinical feedback, and management of skin lesions suspicious for cancer. With approximately 400,000 skin-related referrals in Denmark each year, this improved diagnostic flow stands to significantly reduce societal costs while ensuring that almost all skin cancers are diagnosed and treated with minimal delay.  

An upwards of 15% of surgical or injured patients experience avoidable complications, such as infections, thrombosis events or organ failure. This project uses AI-enabled analysis of real-time data (e.g. chart notes, lab results, vital-sign registrations) for individual patients to predict 18 different post-surgical complications.

Clinical deterioration is a frequent and unfortunate occurrence among hospital patients. +priokritisk is software that uses a model based on explainable (xAI) to analyse selected parameters from patients' electronic medical records (EMRs) to predict those who are at risk of developing sepsis.

Acute hospitalisation requires a special effort to involve patients in their own treatment and ensure that they're aware of their disease's progression – or the development of new symptoms requiring attention. This project aims to improve intervention after acute myocardial infarction by introducing home-monitoring and other support for patients who have undergone highly specialised cardiac treatment and who will continue rehabilitation.

Today, only a very limited subset of antibiotic-resistant bacteria is sequenced. Hence, we only see the tip of the iceberg when it comes to bacterial outbreaks as there is also definite transmission with fully antibiotic-susceptible bacteria. This project aims to change that by sequencing a larger range of bacteria than is done today.