Identification of high risk patients
PI Health Solutions supports clinicians, health care providers and health care organizations with predictive tools and logistics to support decision-making.
Clinical/health care: For clinical and organizational decisions using web-based application software packages (Apps). To make risk predictions based on various sets of parameters, we use state-of-the-art algorithms supported by artificial intelligence for pattern recognition. A particular design property of our Apps is that the algorithms are continuously improving whenever the App is used in clinical practice (self-learning). Typically, risk predictions include clinical parameters but it may also include laboratory or imaging parameters (flexible modular principle) among other parameters. A prime example is the pre-operative prediction of post-operative cognitive deficits, which allows recommendations about how to pursue (e.g. surgery: yes/no, pre-operative patient preparation, peri-operative treatment adaption). Accordingly, this software BioCog supports the decision-making of the clinician in clinical care of the individual patient (personalized medicine). However, it also helps to implement general health care policy strategies what particular patients qualify for elective surgery. For the original development of our risk algorithms we rely on big data that have been collected as part of large multisite and international cohort studies. Accordingly, PI Health Solutions closely cooperates with highly renowned clinical institutions like the Charité Berlin (Germany) or the University Hospital Utrecht (Netherlands) as well as a number of additional clinical, academic and private institutions.
Health organizations: For decision-making, health organizations require predictive data about high risk groups on the population level. Ideally, this includes data on socioeconomic, psychosocial and health condition as obtained from self-asessement questionnaires but also objective data (e.g. laboratory parameters). However, objective data are not easily obtained on the population level – in particular when it is the aim to obtain population-representative data. In this case, both a sophisticated IT solution and an intelligent logistics for collecting blood samples is required. A topical example is the population representative Corona-BUND Study. (www.rki.de) In this case, nationwide online questioning is combined with a nationwide collection of blood samples and throat swaps which then will be analyzed in central laboratories. PI Health Solution plays a central role in this study by 1) providing the logistics for the collection of blood samples and throat swaps and 2) by contributing the expertise and algorithms to make (de-indentified) predictions whether certain population groups are at risk to acquire the virus infection and what groups in the population are likely will have a serious outcome.