Health Analytics

Think tank and pioneer in the field of predictive analytics and data mining.

Contact us

Our vision

We expect precision medicine to improve patient outcomes and to save lives. Within our Precision Medicine unit, the team at Elsevier Health Analytics uses healthcare data and real world evidence to deliver state-of-the-art health outcomes research and predictive analytics.

Data as a healthcare revolution

Medical knowledge is exploding. New technologies such as biomarker testing and whole genome sequencing are rapidly expanding diagnostic and therapeutic options. At the same time, the growing amount of data collected by general practitioners, specialists, hospitals, and patients themselves is dramatically increasing the need for modern data management and analytic approaches.

We see the massive growth of health data as a huge opportunity for society. This data, analyzed and combined with scientific medical knowledge, drives important insights, allows life-saving predictions, and supports physicians, healthcare professionals, and hospitals to manage the complexity of modern medicine. Our work helps physicians select and prescribe more precise treatments and to improve their patients’ outcomes.

We use the most modern statistical, machine learning and artificial intelligence methods to analyze healthcare data, combining it with medical insights text-mined from scientific publications in Elsevier’s knowledge bases (including ScienceDirect®, Pathway Studio, Pharmapendium, Reaxys, and ClinicalKey®). This combination is unique. It places us in a position to make highest quality predictions on disease risk and progression, and to suggest the best therapeutic options for patients.

  • We believe that the quality of care can be improved, at lower costs.
  • We believe that big healthcare data will revolutionize patient care.

Health outcomes research: improving healthcare systems

Elsevier Health Analytics analyzes the effectiveness of care and treatment options on big healthcare data using advanced matching algorithms to select comparison groups with equal risks that differ only in the variable of interest.

We provide answers on:

  • Incidence, prevalence, and mortality for matched groups
  • Typical patient journeys, triggers for patients’ disease progression
  • Comorbidities characteristic of a disease stage
  • Success of different medications or therapies for different patient types or groups
  • Which, possibly still unknown, side effects of treatment can be derived from the data
  • Regional variations in care
  • Compliance with medical guidelines and best practices, deviations from pathways
  • Gaps in care

Representative population models

We gain ground-breaking insights from approximately 5 million anonymized patient datasets to which we are granted access for research purposes by German statutory health insurers. This database of anonymized German claims data is a proven representative sample of ca. 7% of all statutory health data in Germany.

6 years of longitudinal data are linked and include demographics and periods of treatment, physician specialization, diagnoses (ICD codes), therapies and prescriptions (procedure and ATC codes), costs, and diagnosis-related groups (DRGs) for hospital inpatients.

Medical Graph

Elsevier Health Analytics has mined this anonymized dataset for disease trajectories using the temporal disease trajectory model. An individual patient’s data can be input into this system to estimate the statistical probability of the de-novo occurrence of ca. 1600 diseases within the next four years. The average area under the receiver operating characteristic (AUC) is ca. 0.80. The Medical Graph uses age- and gender-group incidence to calculate relative risks.

The Medical Graph is a population level representation of diseases, comorbidities, disease trajectories, and risks, including cancers.

Oncology pathways

Elsevier’s Via Oncology unit (www.viaoncology.com) has provided documented pathways in oncology for 2.5 million US patients. The data set includes detailed clinical information such as biomarker tests, tumor status, and tumor progression.

The Via Oncology data enables us to model a longitudinally-linked view of current oncology treatment, including therapy changes.

Our growing network of data sources in the US and world-wide enables us to compare treatment-models between countries and health care systems.


HIPAA, GDPR and data privacy

Data privacy and data security are our highest priority. We maintain the highest standards of data privacy during our research projects. Elsevier Health Analytics is compliant with the most stringent data privacy laws worldwide, with HIPAA and GDPR.

All our healthcare data is used exclusively for healthcare research or to support providers treating patients. We publish only summary study results, reported cohorts comprise a minimum of 100 patients, and no raw data, even if anonymized, is disclosed – ever.

Any data that we use is securely anonymized at the source, employing audited procedures. It is not possible for us to re-identify any individual, provider, or health insurance.

All team members are GDPR- and HIPAA-trained and certified annually.

We use audited trust centers and our Research Data Management in Health system to ensure that no data can be removed from our data center.


Get in touch

Highly experienced team

Our team has more than 100 years of combined expertise in statistics, advanced machine learning, and artificial intelligence on healthcare data. We have successfully conducted ca. 300 research projects for our international clients.

Our success is based on the close collaboration of our cross-domain experts: data scientists, statisticians, physicians, IT specialists, and health economists.

Contact us

Please contact us to receive more information on how Elsevier Health Analytics can support you.