Data-driven care improvement

Doctors and other care providers want to provide care that best meets the needs of the patient. But how do we determine what the best care is for the patient? To determine this, it is important to map out the standard of care: the patient, the health complaints and the outcomes of the care. This combination of data provides insight into the quality of care. But mapping is easier said than done.

After all, every patient is different. Genetic differences, differences in lifestyle, capabilities, expectations and environment make us all unique. It is these differences that determine whether we get a certain disease, how we react to therapy and how we experience the care process. It is therefore crucial to understand these factors. An important step is the structured recording of the patient's health complaints during the entire treatment in combination with the experienced course of the disease.

This is a challenge. Filling in long questionnaires is no fun for anyone. Healthcare providers already spend about 40 per cent of their time on administrative burdens. What's more, interpreting and analysing this enormous amount of data is a special skill. Are the relationships found coincidental or are they causal? Analysing the data and providing feedback on the findings to the care recipient and care provider requires specialised knowledge and experience.

At ConsultAssistent, we therefore bring together data collection, processing, analysis and feedback in order to help care with content. Using adaptive auto-anamnesis, follow-up and PROM questionnaires, ConsultAssistent collects the right data from the patient. Because of these questionnaires, patients are better prepared, doctors get to the point faster and care professionals are relieved of many administrative tasks.

After anonymisation or, if this is not possible for the study, pseudonymisation, all data from patients who have given explicit permission are placed in context and analysed for better insight into outcomes. And thus for more data-driven leads for effective and efficient care. For this analysis, we deploy descriptive statistics and dashboards and use advanced techniques such as Machine Learning and Neural Networks.

We use the descriptive analyses and dashboards to find correlations. These links are discussed with medical specialists. Valuable new descriptive analysis and dashboards become available to everyone, so that the wheel does not have to be reinvented each time. The links contain starting points for improving and personalising care standards and for improving care processes.

We use Machine Learning and Neural Networks to develop prediction models. With these techniques, we can use the large amount of data to predict which therapy is best for each individual patient. Using the data from a patient's autoanamnesis, which includes information about lifestyle, capabilities, expectations, circumstances and health complaints, the most promising therapy is determined based on the therapy that provided the best care outcomes in patients with similar characteristics.

We are currently busy developing this kind of prediction model. Proof of Concepts have already been completed. Once these have been tested and validated, the models will be used as decision support for the doctor. The doctor will be the one at the helm, making the right choices together with the patient. The technology supports doctor and patient by analysing millions of links in milliseconds and making suggestions on the basis of these, something that our human brain is simply not capable of doing.