medtech

In Asia, 47% of countries are facing a severe shortage of Oncology healthcare professionals. As the population continues to age, the incidence of cancer is expected to increase. Oncology centres are already feeling the strain, clinic sessions often overrun, appointments are overbooked, and patient experience is suboptimal. In fact, a study in Singapore shows that 50% of patient did not obtain information from their preferred source which is their healthcare professionals.

To make matters worse, the operational workflow is highly inefficient. About 43% of oncology professionals spend more than 20 hrs per week on administrative tasks, when this time can be spent on patient care activities.

At Heysra, we are building safe, oncology-specific AI healthcare agents designed to automate administrative and simple clinical tasks using large language models (LLMs) and medical speech recognition technology. Hesyra frees up valuable healthcare resources and increases accessibility to care. This allows healthcare professionals (HCPs) to devote more time to addressing patient concerns more effectively and efficiently.

Hesyra aims to automate administrative tasks and simple clinical tasks that do not require complex clinical reasoning in order to free up time for oncology HCPs so that they can focus on caring for their patients, address the diverse needs of cancer patients, and deliver person-centred care.

We are building a multi-agentic system architecture that leverages on the strengths of LLMs and oncology domain specific rule-based algorithms. We will subsequently orchestrate these multiple agents to accomplish a diversity of tasks ranging from clinician-fronting tasks to patient-facing ones.

Hesyra is proud to have been recognized with several prestigious awards and achievements. We received the BES18SM BES Design Challenge Silver Award and were honored as a Youthtopia Mini-Grant Winner, receiving a grant to further our mission. As part of NUS Enterprise, we are incubated under the Block71 Hanger and are proud alumni of the NUS Graduate Research Innovation Programme (GRIP) Run 12. Additionally, we were recognized as one of the Top 60 teams at the MedTech Innovator APAC 2024, reflecting our commitment to innovation in healthcare.

If you’re a specialist/ general practitioner – do check us out at hesyra.ai

Approximately 1 billion people, or 15% of the world’s population, live with disabilities, often causing mobility issues. Stroke patients, with 12.2 million new cases annually, face significant challenges, making it difficult to reach light switches, fan controls, and electronic devices. This limits their independence and increases dependence on others, leading to physical and emotional strain for caregivers and family members. Healthcare providers also struggle to find suitable solutions to improve accessibility and independence.

Connecto offers a solution with a smart IoT home system that includes a custom-designed assistive mouse and intuitive software. This system empowers users to control their home environments and communicate with caregivers easily through sensors and actuators. By promoting independence and reducing the need for constant assistance, Connecto aims to improve both the physical and emotional well-being of users, enhancing their overall quality of life at home.

Connecto is the student commercial and engineering arm of Microtube Technologies, a local MedTech start-up. Connecto made it to the top 5 of the Medical Grand Challenge (MGC) 2024.

The team developed a multiplayer serious game to encourage the development of soft skills and technical skills in minimally invasive surgery. Our solution is a customised multiplayer cooperative game, ‘Protectors of Eshana’ in which 3 players take on roles that mimic the scopist, main surgeon, and assistant surgeon in an operating room, to defuse a time bomb in the game. Throughout the activity, the three players in the game must mutually communicate with each other to approach the target, encouraging the development of soft skills and extensive cooperation between players, similar to the scenario in the operating theatre.

‘Perceived effectiveness to an innovative mobile based serious game on the improvement of soft skills in minimally invasive surgery’ has been published in the Asian Journal of Endoscopic Surgery.

Project has been funded by the Games for Health Innovation Centre (ALIVE) Serious Games Grant (SGG18/SN15). Grant project titled ‘Innovative Mobile Based Serious Game Application to Improve Technical and Cooperative Skills in Minimally Invasive Surgeries’. Project was a Medical Education Grand Innovation Challenge (MEGIC) 2023 finalist and received the MEGIC seed grant.

Hospitals struggle to anticipate how long post-surgical patients will remain in ICU. Over- or under-estimating length of stay (LOS) strains beds, staffing, and budgets, disrupts operating lists and step-down flow, and leaves families without clear timelines. Existing tools such as coarse risk scores and clinician heuristics often miss procedure- and patient-specific signals. We set out to build a reliable, generalizable, and operationally simple predictor of post-operative ICU LOS.

The Philoso-P is our machine-learning solution that forecasts ICU LOS after surgery to support smarter bed planning, staffing, and family counseling. Using de-identified MIMIC-III/IV data (2001–2019) across seven surgical categories, we defined LOS as the ICU stay immediately after each surgery, merged demographics, vitals/labs, medication data, and procedure text (spaCy NER), and trained models on an 80/20 split. Framing prediction as Short (0-2d), Medium (2-7d), and Long (>7d) improved clinical usefulness despite class imbalance. Our best model—a LightGBM classifier-achieved F1-scores of 0.728 (Short), 0.548 (Medium), 0.482 (Long) and outperformed regression approaches (~0.638 vs ~0.20 accuracy). Feature importance highlighted primary ICU medications, body temperature, and oxygen saturation, readily available signals for early risk stratification.

Designed to be usable by non-technical staff, The Philoso-P helps teams anticipate demand, guide early mobilisation and discharge planning, and give families clearer expectations; future work will incorporate complications and failure-to-rescue metrics to further refine care pathways.