Home / Digital Change / Healthcare / Artificial Intelligence for healthcare
Artificial Intelligence for healthcare
Artificial Intelligence for healthcare
Use case
Digital technologies and healthcare 4.0
Digital services to citizens and dematerialization of electronic documents, folders and files, IoT Data Analytics, Telemedicine, Connected care: there are many areas of digital innovation in the world of Healthcare.
Admittedly, it is true that technology is not a magic wand make all the problems of a complex and fragmented machine like the national health system go away, but it certainly presents opportunities - uniquely modern - for evolving the system towards a more effective management in all respects. In particular, the essential health of citizens.
It’s no secret that Almaviva has been the leading protagonist in the world of Healthcare in Italy since computerization became necessary, and today it continues to thrive, taking full advantage of the potential offered by digital innovation.
A mine of data
Patients monitored throughout their clinical course and hospital "archives" produce a huge amount of data. This data is more readable and interpretable than ever before. When cross-referenced with other information - that of scientific publications or guidelines, for example - and properly analyzed, they generate advanced predictive models to support healthcare professionals' decisions.
The entry into the clinical field of Artificial Intelligence, Big Data and Machine Learning promotes extraordinary advances in precision medicine and the subsequent improvement of health care performance.
Today the road to the medicine of the future appears to be clearly mapped out: more personalized diagnosis and treatment and therefore more effective diagnosis will be provided thanks to the new knowledge extracted from the immense amount of data already available.
Self treating with an algorithm
Almaviva and Asst Vimercate
Almaviva and Azienda Socio-Sanitaria Territoriale di Vimercate are carrying out a joint project: a hospital center that operates according to standards of excellence in four locations - Carate Brianza, Giussano, Seregno, as well as Vimercate - while actively collaborating with all the actors of the Lombardy Socio-Sanitary System.
The project sees clinicians and data scientists working hand in hand to develop intelligent algorithms to interrogate to make increasingly accurate medical decisions.
Having started at the end of 2018, the project foresees the development of an innovative system for the progressive and continuous enhancement of the information heritage of the Vimercate Social and Territorial Health Agency (ASST). It is aimed at the creation of effective clinical decision support tools and solutions for predictive and prescriptive analysis in the management of social and health care processes.
The starting point was the development of predictive models to assist in the management of certain chronic pathologies which require the National Health System to use considerable and continuous resources.
The project has actively involved the departments of Diabetology, Nephrology and Internal Medicine and has now extended to the Department of Oncology and Hospital Pharmacy.
The models are based on different information sources such as blood tests, physical examinations, comorbidities, admissions, emergency room access and pharmacotherapy.
The first performance of the intelligent system
In the following diagram a brief summary of the cases developed and in progress:
| Use Case | Method | Description |
|---|---|---|
| Diabetic/Healthy | CNN+MLP | A patient is classified as diabetic or healthy based on routine blood test values. |
| Diabetic/Prediabetic/Healthy | CNN+MLP | A patient is classified as diabetic, prediabetic or healthy based on routine blood test values. |
| Diabetes Complications | Random Forest | There is a prognosis of possible onset of retinopathies, nephropathies and cardiovascular diseases in diabetic patients, within 1 year and within 3 years, based on blood tests, pharmacotherapy, comorbidities, and any previous complications. |
| Dialysis from IRC stage G4 | Random Forest | There is a prognosis of the onset of total dysfunction, and therefore the time of dialysis, within a year and a half or less, in patients with chronic renal failure in stage G4, starting from blood tests, previous transplants, in relation to treatment and comorbidities. |
| Heart Failure | XG Boost | There is a prognosis of re-hospitalization of the elderly patient, admitted for heart failure, within 1 month or within 1 year, on the basis of data from the electronic medical record, physical and blood tests during hospitalization, comorbidities, and previous hospital access. |
| Oncological Toxicity | CNN | The possible onset of toxicity due to oncological treatment is predicted on the basis of data from the electronic oncological chart (examinations, therapies). |
Data Scientists and medical staff
Accuracy between 88% and 96%
The development process was carried out in perfect synergy with the clinical staff and the professionals of the health care company's information systems. As such, it has been possible to identify the problems of interest for clinicians, develop models consistently with the domain and implement the software developed within the hospital applications already in use.
It is only thanks to the combination of Almaviva technological expertise and the clinical expertise of medical specialists that it has been possible to achieve extremely satisfactory results in terms of prognosis accuracy between 88% and 96%.
Satisfaction has also been widespread among clinicians who have requested its implementation within the electronic patient record already in use to support their daily activities: this and other implementations will be further integrated in the coming months.
Evidence based Medicine + Machine Learning
Almaviva is the forerunner in Italy
A reference point for the research, development and production of decision support AI solutions for healthcare professionals, Almaviva is developing an analysis platform to make prognoses about chronic diseases, from the risk of postoperative infections to the outcome of drug therapies.
The goal is to transform data into a new asset and lead the way in Italy in integrating Evidence based Medicine and Machine Learning for the improvement of healthcare services.
Acknowledgements
Almaviva and Asst Vimercate were awarded the Premio Innovazione Digitale in Sanità 2020, promoted by the Osservatorio Innovazione Digitale in Sanità of Politecnico di Milano, in the category "Management of clinical and care processes in hospitals".
Learn more in Almaviva Press Room
