The digital revolution has transformed almost every industry including the healthcare sector. Taking the cue, Siemens Healthineers has invested heavily in digital transformation and emerging technologies.
Analytics India Magazine got in touch with Dileep Mangsuli, executive director at Siemens Healthineers, to understand more about healthcare and digital transformation.
AIM: What are Siemens Healthineers’ most noteworthy services and products?
Dileep Mangsuli: We have 65 AI-enriched products, many of which have won prestigious awards for their innovativeness.
AI-Pathway Companion is a software application that facilitates personalised and standardised diagnosis and treatment decisions along disease-specific care pathways using augmented intelligence, data integration, and insights from cohort analytics. It supports physicians in decision-making and complex disease management by facilitating diagnosis and therapeutic decision-making along disease-specific care pathways that enable a reduction in unwarranted variations in care, inappropriate treatment plans and improve adherence to guidelines.
From a tech perspective, AI-Pathway Companion intelligently integrates longitudinal patient data and co-relate insights from imaging, pathology, lab, and genetics ( both structured and unstructured data) into a unified FHIR-based data model with the support of AI and smart data integration techniques with a clear focus on managing complexity and data overload.
AI-Rad Companion aims to advance precision medicine and to transform care delivery in diagnostic imaging. The AI-Rad Companion’s overall approach is to provide analyses of medical images with the highest degree of automation possible to improve workflow efficiency, clinical outcomes, and patient experience in radiology departments. So, the AI-Rad Companion results are optimally integrated into the radiologist’s standard clinical workflow without unnecessary disruption.
AI-Rad Companion uses artificial intelligence to support the clinical decision-making process by automating the analysis and quantification of clinical imaging data and the presentation of results and report generation.
AIM: How does Siemens Healthineers leverage AR/VR to deliver healthcare solutions?
Dileep Mangsuli: Immersive technologies like AR/VR have a firm place in the future of healthcare by improving the quality of care, optimising the diagnostic experience, and increasing employee productivity.
They allow surgeons to practice a wide variety of surgical procedures and minimally invasive interventions in silico. These immersive technologies enable personalised therapy planning, wherein imaging-based 3D simulations of the individual target organ are created in silico. Depending on the surgery type, the best approach is chosen based on 3D-printed models and/or virtual reality (VR) models that are explored using VR gear. They also help to standardise and partially automate surgical and interventional procedures.
We have a patient experience app that dramatically improves the MRI patient experience with augmented reality (AR). Healthcare providers can install the solutions in their environment and display them directly on the smart mobile device.
We are also working on photorealistic 3D visualisation technology Cinematic Rendering available as an app for the HoloLens 2. This will expand the range of applications for this technology. Using a mixed-reality headset instead of viewing the clinical images on a 2D monitor when preparing for surgery gives surgeons a realistic 3D overview of the surgical area. This could make it easier to select the right operating room (OR) strategy and thereby increase the accuracy of the surgery.
AIM: What does digitalisation in healthcare entail? Why is it critical?
Dileep Mangsuli: The future of healthcare is digital. Digitalising healthcare is the key enabler for expanding precision medicine, transforming care delivery, improving patient experience, which are essential strategies for healthcare providers to increase value by achieving better outcomes.
Digitalising healthcare involves managing data as a strategic asset, integrating accurate data from fragmented sources, creating a holistic understanding of patients, and improving enterprise visibility. It empowers data-driven decisions by leveraging analytics, benchmark insights, and using digital companions to simplify workflows and enable decision support. Digitalisation connects care teams and patients for better coordination and knowledge sharing, thereby strengthening integrated care across the health system.
With digital technologies, medicine is more precise due to improved diagnostic accuracy and reduced variations, thereby quickly arriving at the right diagnosis and treatment.
In the future, doctors will use a patient’s digital twin to try out different treatments digitally and choose the alternative that is likely to achieve the best outcome. A patient’s digital twin is a personalised, multiscale computational model built from the patient’s data. It will harness relevant data so doctors can digitally compare the predicted outcomes of different surgical procedures and the impact of medication.
AIM: Tell us about digital twins in healthcare
Dileep Mangsuli: The digital twin of an individual is a great vision of the future and can take personalised medicine to the next level. It is a near-real-time virtual replica of a physical object that helps detect issues before they occur, predict outcomes more precisely, and design and build better products. The healthcare sector is just starting to capitalise on this powerful tool. The advances in underlying computational models can be used to create an artificial intelligence (AI)–powered, individualised biophysiological model of the patient to improve medical technology and make it more precise, preventive, and personalised.
In healthcare, a digital twin brings digital data from various sources together with AI, machine learning, data analytics, and computational models that can “interrogate” the data to provide answers to a wide range of clinical questions. The digital twin is updated during the person’s lifetime from the vast amount of data available, including lab results, medical images, examination findings, clinical decisions, behavioural information, and/or genomic data. It is more than a fancy anatomical model; it integrates biological information at both the macro and the micro-level, incorporating the patient’s cellular, molecular, genetic, and clinical information.
A key aspect is the consideration of physiology. When building a patient’s virtual representation, collecting only data that can be observed is not sufficient; we must also consider how things work in the human body. Additionally, this type of modeling takes place at different scales. It may be at a cellular level, simulating biomolecular processes, such as proteins that make up cells and even genes; at the organ level, modelling the heart, liver, or lungs; at the system level, simulating the circulatory system; or at the pathology level, modelling tumor growth, or the mechanics of a failing knee.
When diagnosing patients, clinicians could use digital twin technology to consider patient morphology and patient physiology. By knowing a patient’s genetic and molecular makeup in advance, doctors can better determine whether a particular medication is likely to help and the appropriate dosage to prescribe.
Surgeons could use the technology to navigate a digital visualisation of an organ before operating on it. For example, a cancer surgeon would be able to precisely evaluate how a tumor is positioned in relation to healthy tissue. Similarly, an orthopedic surgeon could create a digital twin to understand the topography of a complex fracture better.
Digital twins also provide the ability to simulate, in advance, how an organ will respond to a procedure or treatment, giving physicians the ability to predict the effectiveness of specific interventions, better understand potential side effects, and accelerate necessary operations while avoiding unnecessary ones.
Finally, predictive modelling based on digital twins could alert physicians to health problems before they manifest, thanks to data gathered across a patient’s lifetime, providing profound insights into ageing and health. This can further advance chronic disease management and population health by integrating behavioural data and life conditions with more general population health data.
AIM: What is the data problem in healthcare?
Dileep Mangsuli: Healthcare organisations now have an abundance of health data aggregated from numerous sources such as electronic health records (EHRs), medical imaging, genomic sequencing, payor records, pharmaceutical research, wearables, and medical devices.
This big healthcare data has considerable potential to improve patient outcomes, predict outbreaks of epidemics, gain valuable insights, avoid preventable diseases, reduce healthcare delivery cost, and improve the quality of life in general.
However, the difficult task is deciding on the allowable uses of data while preserving security and the patient’s right to privacy. Big data, no matter how useful for the advancement of medical science and vital to all healthcare organisations’ success, can only be used if security and privacy issues are addressed.
Adopting big data analysis in healthcare has also lagged behind other industries due to challenges such as privacy of health information, security, siloed data, and budget constraints. Moreover, access to health data for research, in addition to its use for diagnosis and treatment, creates a conflict between health data anonymisation and public interest/medical progress.
Since Hippocrates (approx. 400 BC) and his famous Hippocratic Oath, trust and privacy have always been the foundation in the relationship between patient and physician. It is therefore essential to ensure the security and privacy of sensitive healthcare data. As a result, we see an increased focus on data privacy and information security protection across the globe to protect patient data.
AIM: What is the hiring process for data scientists at Siemens Healthineers?
Dileep Mangsuli: We have a five-step hiring process for data scientists, comprising multiple rounds of technical interviews. We assess the candidate’s basic understanding in data science, as well as the relevance of their work experience in the first telephonic interaction. This is followed by a thorough evaluation of the candidate’s skills, data science fundamentals, analytical and architectural skills, and implementation and deployment expertise.
We also assess the conceptual understanding of mathematical topics, followed by a focus on the role and specific project’s technical skills. Right from machine log analytics to NLP techniques based on the requirements, we provide use cases to candidates and expect them to come up with complete solutions, explaining the rationale behind design decisions, choice of approaches, and the selection of tools. They are also assessed on their programming skills for specific project requirements.