Postoperative delirium (POD) is an enormous problem occurring in 25% of surgical patients aged 60+. POD has severe adverse outcomes: 25% mortality within one month, 2x risk of nursing home admission, costs of 1-2B Euros to health insurers in Germany alone and 38% long-term dementia. There are no treatments available once symptoms arise. Instead, the focus is on prevention. Many highly effective prophylactic treatments are too costly to deploy for every patient. We are developing a novel, AI-based pre-operative risk prediction tool which highlights patients at risk before undergoing surgery.
We have secured 2M euros in investment. Works-like/looks-like prototype is complete. We have 15 LOIs from clinicians who are willing to buy our product. We’re approximately 18 months away from handing in to our notified body and approximately 2-3 years from entering the market as a fully CE marked class IIa medical device
Our market size is enormous. 86 million medium-high risk, elective surgeries on patients over the age of 60 are performed globally every year. In our beachhead market and partner hospitals the Hirslanden group, Basel University hospital and Claraspital we have access to 100,000 patients per year. We aim for a pay-per-use SaaS for 40% of patients which leads to a total addressable market of 2.6 B Euros. By 2050 the number of people 60+ will have doubled.
Problem or Opportunity
Disorientation, memory loss, difficulties in speech. If you are over sixty and about to undergo surgery, you could be at risk of developing any or all these symptoms. Postoperative delirium (POD) is an enormous problem occurring in 25% of surgical patients aged 60+. POD leads to adverse outcomes such as a 25% mortality within one month, double the risk of nursing home admission, costs of 1-2 B Euros to health insurers in Germany alone and 38% of the affected end up suffering long-term cognitive decline and dementia. There are no treatments available once symptoms arise. Instead, the focus is on prevention. Many highly effective prophylactic treatments have been developed but are too costly to deploy for every patient.
Solution (product or service)
We are a start-up developing a novel, AI-based pre-operative risk prediction tool which highlights patients at risk before undergoing surgery. This patient-centric approach increases patient value by improving health outcomes and decreasing costs over the full cycle of care. Our method to develop our tool is to collect a large dataset of patient data and use this data to create an AI based algorithm to predict risk. We are collaborating with three hospitals (Hirslanden Klinik Zurich, University Hospital of Basel, Claraspital Basel), academic partners (Cochrane Anesthesia, Cochrane Response, McMaster University), software developers and regulatory specialists.
Main competitor is “business as usual”. The market today has no incumbents. However, it has high profitability and may attract new entrants. Our strategy raises the barriers to entry significantly. Though software development and marketing of a product is not their focus or expertise, we expect PharmaImage to have a product on the market in the next 3-6 years. Currently they do not have the capital requirement for CE marking. They have performed a clinical trial and hold a database of 764 patients but require a validation trial for regulatory approval, needing approximately 2M Euros to complete their product. We have more data, are better funded and better situated than this German company.
Advantages or differentiators
Spectacular network (KOLs, EIT Health, CIMIT...), large dataset containing thousands of patients, key partnerships (Cochrane, Hirslanden, University hospital Basel, Claraspital Basel, McMaster University, Future Processing Healthcare), Geography (Switzerland – out of pocket pay is accepted, wealthy beachhead market).
As Pipra, we have competitive advantages over other companies and academics in this space. We have an advantageous geographical position. Switzerland is the perfect beachhead market with many surgeries, many wealthy private hospitals and a healthcare system which allows for profits to be made before reimbursement. This can fuel our expansion into healthcare systems which rely on reimbursement. We also have a unique network of dedicated partners. EIT Health, CIMIT and our advisors give us unparalleled access through their vast partner network. The Hirslanden hospital chain serves as both a large customer as well as an influencer within the entire Swiss and neighbouring medical field. Our network of KOLs can help ensure that our tool is well received and written into clinical guidelines. Our lean, competent and committed management team that has overcome every hurdle so far and will continue to do so. Our tool differentiates itself from all currently available POD/POCD risk assessment tools. The current solutions PRE-DELIRIC (original & int’l) and the Katznelson model assess the risk of delirum after an operation. This renders them almost entirely useless since it does not allow for adaptive changes to the pre- and perioperative pathway. There are currently no pre-operative delirium risk assessments on the market. PharmaImage are working on building a tool based on data from their BioCog study. Our solution differentiates itself from this potential future competitor in three fundamental ways. 1. Includes cardiac surgeries Cardiac surgeries have the highest incidence of POD/POCD and have been excluded from the BioCog study. Our tool will include these patients. Clinicians are most worried about their cardiac patients developing POD/POCD due to the high incidence. 2. Compatibility with clinical practice The BioCog study has focused on biomarkers which are not commonly assessed in the clinical practice or which are too expensive to run routinely for every patient undergoing surgery (MRI, omics, long neuropsychological tests, non-standard blood tests). This makes their solution time consuming and expensive. Our tool requires only data which is already routinely collected during the anaesthesia pre-consult. It is thus more compatible with the current clinical practice. 3. More reliable due to more patients We will base or software on more than ten times as many patients than BioCog. We may even include the BioCog data itself into our algorithm. This makes our tool the most robust and reliable.
SaaS business model. Price per service 75 Euro. Break even 2023. Annual revenues with modest market penetration (Millions Euro): Yr 2 0.09 Yr 3 1.95 Yr 4 3.90 Yr 5 7.80 Yr 6 18.95 Yr 7 40.86
Software as a service. Free basic version which recommends and generates demand for the full version. Short term out of pocket expense (82% of interviewed patients would spend 75 Euros on the tool), medium to long term goal is reimbursement. We start by targeting hospital chains.
Money will be spent on
Regulatory pathway to CE Mark and FDA once our current funding runs out.
Offer for investor
Equity, or convertible note, to be discussed. Negotiable.
We have identified a key risk in uptake of our tool. To be used as standard of care we aim to be written into guidelines. We have recruited several KOLs as advisors and will use KOLs, ethics groups and professional societies as channels to influence guidelines.
Incubation/Acceleration programs accomplishment
Winners of EIT Health wildcard competition. Mentored by Harvard’s CIMIT (Consortia for Improving Medicine with Innovation and Technology)
Won the competition and other awards
Winners of EIT Health wildcard competition 2019
We will demarcate our market and create barriers to entry based on three primary strategies: 1. Trade secret & patenting 2. High capital requirement for new entrants 3. High customer switching cost
1. Trade secret & patenting Intrinsically, an AI-based model is more difficult to copy than a statistical model. Further, we will protect our algorithm by storing it safely in our cloud and not physically installing it in hospitals. Finally, we will patent a portion of our software, thus protecting it further. These three protective mechanisms make it practically impossible to copy our algorithm directly. The raw data itself can be used to build a new algorithm. However, we will have several safeguards protecting the patient data. The database we will curate and own will be kept safely and securely. We are working with our partners to ensure (with agreements and incentives) that their data cannot be used for competing products. This strategy of protecting our data as a trade secret means that new entrants to the market will have to gather their own data, which leads to the next strategy. 2. High capital requirement A substantial investment is needed to enter the market. This is due to the requirement of a clinical trial and the considerable expense of regulatory approval. While academic groups may complete clinical trials assessing risk of POD/POCD, typically they are unlikely to take the time and source the capital required to pursue regulatory approval and commercialisation of a risk assessment tool. On the other hand, companies attempting to move into this market have to face high costs for clinical trials. The number of patients used to build the tool is a crucial indicator of reliability. It is used as a marketing device in our comparables ClinRisk and CardioExplorer (both cardiac risk assessments). Our long term strategy for ensuring high capital requirements for new entrants to the market is based on increasing the required patient numbers to compete with our tool. Increasing patient numbers exponentially would be possible by using a self-learning algorithm. However, conventional self-learning and data collection methods are not permitted under the current regulatory and ethics policies. Our way around this problem is to leverage our unique position. Once our tool has been accepted in clinical practice, we can plan a large, multicentric and decentralised trial. A key advisor to Pipra is Prof. Reto Stocker, the head of the Zurich ethics committee. Approval from the Zurich ethics committee, in practice, guarantees approval from all other committees in Switzerland. Each clinician using our tool will be offered to enrol their patients in exchange for being listed as co-author in the resulting publication. This is a strong incentive, especially since the amount of work required is minimal. After a 2-3 year data collection period, a follow-up version of our product can be built based on 1-2 orders of magnitude more patients (200k+ patients). A number, which is not reachable using conventional clinical trial approaches. 3. High customer switching cost Our main task, after completion of the software, is to implement it into the clinical pathway. Our solution will be streamlined to fit into current processes. This is a large task but also has an advantage. Once a hospital chain has chosen our product, it will unlikely change to a different system. The value of using our tool, compared to no tool is enormous. However, switching from our well-integrated tool to a different system incurs substantial costs (especially time) and won’t add much value. We will ensure that our tool is written into clinical guidelines and becomes standard of care leveraging our network of KOLs.