Personalised health and patient safety


The health system is producing data at an unre- strainable speed. Data that can mean personalized therapy, patient safety, personalized cancer prognoses, better prevention and monitoring of epidemics. We show how such data can be exploited, with a series of innovative prototype projects.

Standard cancer treatment consists of a combination of drugs, at various dosages and in different order, implying a huge number of possible cocktails. We develop a new approach to therapy optimization, based on simulation of cancer growth. Also, we develop methods to predict the synergy between drugs based on cancer cell line data. This allows exploring new approaches to personalized cancer treatments.

National population based cancer registries routinely publish survival statistics. To become more relevant both to the clinician and the patient, the survival statistics should preferably be tailored to encompass more detailed information, moving towards personalized cancer statistics.

Patient safety is critical in healthcare. The amount   of data collected in healthcare is vast and rapidly expanding, including electronic health records (EHR) and health care system/organization data. We develop new data analytic methods to predict and control risk in healthcare organizations at system level.

We use data on mobile phone locations and their movements to describe movement of and contacts between people, with the purpose to monitor and predict the development of infectious diseases. Simulating various vaccine plans allows evaluating cost-benefits.


Personalized cancer statistics

National population based cancer registries publish survival statistics by cancer site, stage, gender and time period, using relative survival methods. As new clinical registries are established, more data on treatment and later events become available, in addition to information on comorbidity or income and educational level. To become more relevant both to the clinician and the patient, the survival statistics will be tailored to encompass more detailed information, in line with the tradition of risk prediction models.

Personalized cancer treatments

We develop a new mathematical, statistical and computational methodology to improve current ways of predicting treatment response for breast cancer patients. We design theoretical and computational models that capture key molecular and cellular mechanisms of the breast cancer, reproducing the effect of specific treatments used in actual clinical trials. The models are personalized to specific subgroups by calibrating the computer simulations to actual patient data. This allows personalized treatment guidance, by simulating an array of possible treatment schedules for a given patient to find the optimal regime. In addition, we develop and extend statistical learning methods for predictive modelling of cancer drug sensitivity based on large-scale in vitro screens of drugs and drug combinations.

Healthcare safety management

This project harvest information lying in the collections of high dimensional health records and administrative databases, routinely acquired in hospitals and health institutions, which carry a preventive signal relative to a potential harm to patients. This signal is used to predict the occurrence of a possible harm in an automatic way at aggregated health institution/ward level. It can also be used to define new control rules on hazards, to prevent and mitigate the risk of harm.

Telecom data for epidemics control

We study the spread of infectious diseases, by observing social mixing and mobility patterns of susceptible and infectious individuals, who are key drivers of the spatial dissemination of infectious diseases. The use of mobile phone data containing geo-temporal information from individuals gives an accurate, real-time description of population movements. This in turn allows accurate predictions of epidemic spread.

Activities 2017

  • Variable selection in highly dimensional time series regressions
  • Multi-state Markov models with frailties
  • Stochastic model of breast cancer growth
  • First pilot study of hazard/harm prediction at system level for a hospital ward
  • Mobile movement data in Norway for infectious diseases prediction
  • Drug synergy prediction in cancer cell lines by data integration techniques

 Principal Investigator  Magne Thoresen

Principal Investigator
Magne Thoresen

 co-Principal Investigator  Clara Cecilie Günther    

co-Principal Investigator
Clara Cecilie Günther