The role of clinical decision support systems (CDSS) in delivering value-based healthcare 

Vivek Patkar, MBBS, MS, MRCS, FEBS Fellow of European Board of Breast Surgery

Chief Medical Officer at Deontics

Yan Yan, MSc Energy science and technology, MBA

Global Healthcare Transformation Consultant at Roche

The role of clinical decision support systems (CDSS) in delivering value-based healthcare 

2 November 2022 | 7min

Quick Takes

  • One way CDSS can be used to support the delivery of value-based healthcare is by simulating the outcomes and costs of existing and new pathways.

  • Using real-world data as opposed to random clinical trial (RCT) data is more effective when implementing and using CDSSs.

  • Patient journey mapping is critical in CDSS to provide a holistic view of the exact outcomes and costs captured at each step.

The shift toward value-based healthcare is being made easier by using clinical decision support systems (CDSS). 

The challenge is simulating the outcomes and costs associated with delivering value-based healthcare until now.  

Healthcare executives looking at implementing bespoke clinical decision support systems (CDSS) to improve the patient experience can benefit from hearing insights from a recent pilot project carried out in the Roche Startup Creasphere Program in collaboration with Deontics, specializing in unique artificial intelligence (AI) clinical decision support systems (CDSS) for global healthcare providers.

Today we hear from Yan Yan, Global Healthcare Transformation Consultant at Roche, and Vivek Patkar, Chief Medical Officer at Deontics. 

Using CDSS to support value-based healthcare delivery

HT: How do CDSSs help support the delivery of value-based healthcare?

Yan Yan: Value-based healthcare is gaining more and more traction. However, there are not many successful projects. One of the reasons is that people are naturally risk averse. People hesitate to make a move if they do not see clear evidence of improved outcomes and costs already at the beginning.

So, just imagine if there was this type of CDSS, which could already simulate the benefits of improved outcomes and costs at the beginning. It would be much easier for people to make the decision whether or not to pursue value-based healthcare if yes, then which option is the best?

Vivek Patkar: To elaborate on what Yan Yan said, let’s look at the example high cholesterol management pathway in primary care used by the National Healthcare System (NHS) in England.

The primary care physicians will need some evidence that the additional responsibility given to them of managing high-cholesterol pathways will indeed make a positive change and will be cost-effective before they embark upon a new task. It is important to back the claim of the new pathway benefits, with evidence and data to convince physicians to start using these pathways.  As Yan has said, if the CDS systems could use the data from their own practice to simulate the new pathway and quantify costs and benefits, it will give confidence to the customers to implement and use the new pathway.

How Deontics is meeting the unmet needs of today

HT: Deontics has participated in the Startup Creasphere program, a health innovation platform connecting startups with industry-leading corporate partners to help transform healthcare. What unmet need are you trying to address in this pilot project and what are the expected outcomes of this collaboration? 

Yan Yan: With there being a lack of simulation tools to showcase the outcome and cost changes for value-based healthcare solutions, what we were trying to build during the Startup Creasphere pilot project was a proof of concept of a patient journey mapping web tool, which exactly simulates outcomes and costs based on different patient journeys. 

Vivek Patkar: Healthcare consultants need a tool that will allow them to:

  • Design interactive and executable patient journeys, pathways, and workflows
  • Create quick alternative pathway variations
  • Run these pathways in simulation mode using anonymized real-world data, and provide outcomes in terms of costs and benefits for each variation of the pathway
  • Provide their customer an option for deploying the new optimal pathway for real  

None of the current tools available in the market provide any of these functionalities, let alone all functionalities. Current pathway design tools merely draw electronic versions of paper flow diagrams. They are not executable and can not make use of any data. Deontics has built a CDS-AI engine that can bring those pathways to life, and has the ability to provide all the functionality mentioned above. 

Yan Yan: What’s more is that you can use that same platform to simulate the outcomes and the costs associated with every single step during that pathway.

In the pilot project, we chose indicators as outcomes, and we compared two different pathways. One, we called the old pathway so it’s the existing old pathway in the hospital. The second one is the new pathway where we implemented a new risk factor. From the comparison between the outcomes and the costs simulated by this tool for those two different pathways, we could showcase now by using this new risk factor generated pathway, the percentage of improvement for both outcome and costs. 

The scope was to achieve proof of concept to improve this patient journey mapping web tool. By the end of the collaboration, we’d successfully achieved our goal. 

Clinical Decision Support Systems in Healthcare

Benefits of using real-world data in CDSS

HT: What are the benefits of using real-world data in CDSS?

Yan Yan: Using real-world data really provides reliability to CDSS. Take the patient journey as an example in consulting. In a lot of cases, patient journeys are just mindmaps in consultants’ heads. We draw them out based on conversations we had with healthcare professionals (HCP) and based on some guidelines, to what extent, do they reflect reality? We don’t know. By using real-world data, we improve the reliability of the patient journeys, and in the end, the reliability of the CDSSs.

In our pilot project, we had two stages. In the first stage, we were using 12 paper cases created by Deontics. In the second stage, we used anonymized real-world data. Imagine if you are a C-suite in a hospital and I’m trying to convince you to pursue the new patient pathway. On one hand, I show you results based on 12 paper cases, and on the other hand, I show you results based on 500 anonymized real-world data. Which one would you believe in? Which one has more reliability?

Vivek Patkar:  Most CDSSs use Clinical Practice Guidelines (CPGs) as their primary knowledge source. CPG recommendations are based on the data from published Randomized Controlled Trials (RCTs). RCTs represent a very small patient population. For instance, it is estimated that less than 5% of adult cancer patients enroll in cancer clinical trials.1 Most of the patients are treated or managed outside RCTs. So, the conclusions or expected benefits can’t be easily extrapolated and generalized for real-world data.

Usually, you will get different results when you run the same pathway, the same experiment with real-world data because many patients will leave or change the treatment halfway and the hospital won’t have the resources at the level par with clinical trials. The results will also differ from hospital to hospital or organization to organization because of the differences in their patient population. Deprived areas often have high-risk patient populations and the magnitude of benefits may be significantly higher for the same intervention compared with low-risk populations. By using organizations’ own real-world data, Deontics tool aims to provide realistic and more accurate estimates of costs and benefits. 

The role of patient journey mapping in CDSS

HT: What is the role of patient journey mapping in clinical decision support systems?

Yan Yan: In our pilot project, patient journey mapping is a key component of this type of CDSS. I talked at the beginning that different value-based healthcare solutions result in different pathways, hence different outcomes and costs. Only by deeply understanding the patient journeys could we simulate the different outcomes and costs and thus give decision support.

Vivek Patkar: For the adoption and optimal usage of any CDSS solution, it should fit well around the natural patient journey/ clinical workflow. Detailed patient journey mapping exercise, provides a backbone to design your services. It also helps identify any dependencies,  redundancies, and opportunities to streamline the workflow. Patient journey mapping is the first step in designing CDSS solutions and should be an integral part of the technology.

Key takeaways for healthcare executives creating a bespoke CDSS solution to improve patient care delivery

HT: What key takeaways would you give to executives working towards creating bespoke CDSS solutions to help improve patient care delivery or ways of working? 

Yan Yan: The key takeaway I would give those executives is to deeply understand the user profile of this type of CDSSs. Who will use the tool? Is it clinicians, hospital C-suite, patients, or healthcare consultants? For example, in the case of the consultant, you need to understand what are the Jobs-To-Be-Done, and what are the pains and gains for healthcare consultants. They are different if you would design such a solution for patients or clinicians. 

Vivek Patkar: I would recommend the following for healthcare executives creating a bespoke CDSS solution. First, bespoke CDSS solutions should be underpinned by sound science and safety-critical engineering.

Second, the technology must use industry-accepted interoperability standards for connectivity to the other IT systems otherwise the solution will not be scalable. Preferably the CDSS technology should be domain and EMR-agnostic. The technology should be validated in peer-reviewed literature to prove its robustness.

To find out more about future trends of clinical decision support systems in healthcare, check out part 1 of our interview with Vivek here.

Vivek Patkar, MBBS, MS, MRCS, FEBS Fellow of European Board of Breast Surgery is Chief Medical Officer and one of the founders of Deontics Ltd, a medtech company in the field of AI and Clinical Decision Support. Vivek was originally trained as surgical oncologist at TATA Memorial hospital, the largest tertiary cancer hospital and research centre in India, after graduating in medicine from the University of Bombay in 1995. Vivek has developed and piloted a number of clinical decision support applications for cancer as well as other indications and led a team of researchers at the University College London medical school (Royal Free Hospital) which worked on improving the quality of decisions made in cancer multi-disciplinary team meetings using clinical decision-support. Most recently Vivek and collaborators at Guys & Kings received an NHSx and NIHR AI in health and care award in 2020.

Yan Yan, MSc Energy science and technology, MBA is a Global healthcare transformation consultant at Roche diagnostics international, where she helps hospitals in different locations in the world to transform to new ways of working (e.g. VBHC, digitalization and remote-monitoring etc). Before joining Roche, she has worked in different industries (e.g. IoT and power generation) globally for 7 years. She holds a master degree of energy science and technology from ETH Zurich and a full-time MBA from University of Sydney.


  1. Unger et al. (2016). Am Soc Clin Oncol Educ Book, 35. 185-198