Clinical Decision Support Systems (CDSS) were first introduced in the 1980s and have since experienced a significant increase in adoption and effectiveness within healthcare. This growth has been fueled, in part, by endorsements from US government acts, such as the Health Information Technology for Economic and Clinical Health (HITECH) Act of 2009, which allocated $22.6 billion to incentivize the implementation of health information and clinical decision support systems. These efforts aimed to promote the use of electronic health records and its associated capabilities and additional tools to improve care coordination, quality of care, and reduce health disparities. As a result, the CDSS market has experienced remarkable growth, with an estimated market value of approximately $2.1 billion USD. Experts predict that the market will continue to expand, with a projected value of $3.81 billion by 2030, resulting in a CAGR of 6.83%. These figures are a testament to the increasing importance of CDSS in healthcare, as they offer healthcare professionals timely, evidence-based recommendations and real-time access to relevant patient data, ultimately leading to improved patient outcomes.
Though CDSS can be classified as a mature and saturated market, Brooks Hill Partners has conducted extensive research on CDSS over the past several months, exploring the market landscape, emerging CDSS startups, the applications of CDSS, and industry trends. This series of posts, which we plan to release over the coming months, will provide a commentary on CDSS and their evolving role in healthcare. In this initial post, we introduce CDSS and provide an overview of their purpose and underlying technology.
Clinical Decision Support Systems (CDSS) are computer-based tools that provide healthcare professionals with actionable information and medical knowledge to aid in clinical decision making. They are designed to help clinicians make better informed decisions about patient care by providing them with real-time information and latest evidence-based recommendations. These software systems can take many forms ranging from simple reminders and alerts to complex algorithms and predictive models. Today, CDSS are often integrated into electronic health records (EHRs), computerized clinical workflows, or any similar healthcare information system to enable clinicians with easy access to patient data and clinical guidelines.
CDSS are built to improve the quality of care, increase patient safety, and enhance efficiency in healthcare delivery. To achieve this, any CDSS created will adhere to the ‘five rights’ principle that lays the foundation to any successful CDSS framework:
“delivering the RIGHT information to the RIGHT person in the RIGHT intervention format through the RIGHT channel at the RIGHT time”
By following the ‘five rights’ CDSS can provide clinicians with timely, relevant information and recommendations that can help improve screening and early detection, diagnosis, and treatment and patient outcomes.
From a technological standpoint, a typical CDSS contains three core elements: a base or data management layer, an inference engine or processing layer, and a user interface layer. The data management layer consists of patient data, clinical databases that stores information on diseases, diagnoses and medications, and clinical pathway systems in the form if-then decision trees (knowledge-based) or machine learning models (non-knowledge-based). The inference engine or processing layer serves to incorporate all elements of the data management layer by applying rules and algorithms from the knowledge or non-knowledge-based system to the clinical database whilst simultaneously taking into consideration patient data. The interface layer displays the recommendations and guidance synthesized from the inference engine – this layer can take the form of an EHR system dashboard or a mobile/web-based application.
CDSS can be broadly classified into two categories based on the type of technology the system is grounded in: knowledge-based systems and non-knowledge-based systems.
Knowledge-based CDSS are based on a set of rules and algorithms that are derived from expert medical knowledge and clinical guidelines. These rules can be made using literature-based, practice-based, or patient-directed evidence. Fundamentally, knowledge-based systems rely on formalized knowledge representations, such as ontologies, decision trees, and if-then rules to provide clinicians with recommendations and alerts based on a patient’s medical history and data on current condition.
Non-knowledge-based CDSS still requires a data source, but further relies on an additional layer of artificial intelligence, machine learning algorithms and statistical models to make predictions or recommendations. Techniques that are often incorporated in non-knowledge-based CDSS to identify patterns and correlations in patient data include data mining, clustering, neural networks, and genetic algorithms. Further, depending on the application and intended use of the CDSS, non-knowledge-based CDSS can also incorporate natural language processing and image analysis to extract meaningful information from unstructured data sources such as clinical notes or medical images.
Grounded in evidence-based medical knowledge due to formalized knowledge representations such as ontologies, decision trees, and if-then statements.
Easy to maintain and update as new medical knowledge becomes available.
Highly customizable to meet specific needs of healthcare providers and organizations.
Highly consistent which can reduce variability in clinical decision making.
Limited scalability as the number of rules and knowledge sources grows.
Time-consuming to develop due to significant reliance on input from clinical experts and knowledge engineers.
Limited to known knowledge which may not be comprehensive or up to date
Difficulty in capturing tacit knowledge – knowledge that is difficult to express or codify may not be effectively captured
Ability to handle and analyze complex large datasets.
Highly adaptable to changes in patient data and adjust recommendations accordingly.
Machine learning capabilities allow the system to continuously learn and improve over time, leading to more accurate and personalized decision support.
Lack of interpretability – CDSS that uses deep learning can be difficult to interpret and understand making it challenging to validate recommendations or troubleshoot issues.
Limited knowledge transfer if a system is trained on specific datasets – difficult in transferring knowledge to new datasets or patient population.
Reliance on data quality – systems rely on completeness and quality of data to train the system which can lead to poor or inaccurate recommendations.
Both knowledge-based and non-knowledge-based CDSS come with their advantages and disadvantages; therefore, in recent years, there has been an increasing shift in interest for the development of hybrid CDSS. These systems aim to leverage the strengths of both approaches by integrating formalized knowledge representations with machine learning and statistical models. Creating a hybrid CDSS allows clinicians access to the precision and transparency of knowledge-based systems with the flexibility and scalability of non-knowledge-based systems.
Moving on, the applications of CDSS and their use cases will be explored in the next post of this series.
Castillo, Ranielle S, and Arapad Kelemen. “Considerations for a Successful Clinical Decision Support System.” NursingCenter, https://www.nursingcenter.com/lnc/ce_articleprint?an=00024665-201307000-00003.
Straits Research. “Clinical Decision Support Systems (CDSS) Market Size Is Projected to Reach USD 3.81 Billion by 2030, Growing at a CAGR of 6.83%: Straits Research.” GlobeNewswire News Room, Straits Research, 26 July 2022, https://www.globenewswire.com/en/news-release/2022/07/26/2486236/0/en/Clinical-Decision-Support-Systems-CDSS-Market-Size-is-projected-to-reach-USD-3-81-Billion-by-2030-growing-at-a-CAGR-of-6-83-Straits-Research.html.
Sutton, R.T., Pincock, D., Baumgart, D.C. et al. An overview of clinical decision support systems: benefits, risks, and strategies for success. npj Digit. Med. 3, 17 (2020). https://doi.org/10.1038/s41746-020-0221-y
Admin. “Introduction to Clinical Decision Support System (CDSS).” Omics Tutorials, 12 Aug. 2021, https://omicstutorials.com/introduction-to-clinical-decision-support-system-cdss/.