The Digitalisation of Hospitals: how long will startups need to wait?
The Digitalisation of Hospitals: how long will startups need to wait?
The digitalisation of hospitals is inevitable. So what’s the hold up? We’ve recently looked at drug delivery innovations and patient driven medicine, but here Gian Seehra, of Octopus Ventures’ Future of Health pod, talked to 25 industry experts to identify the five problem areas that need to be cured before the full benefits of tech can be felt.
- Dire Data
According to GE Healthcare, 90% of healthcare data is concerned with imaging technology and 97% of that goes unanalysed or unused. So the problem is not too much data, but too little. Or rather, too little unprocessed, usable data.
One of the major reasons hospitals struggle to deliver good patient care is their inability to predict demand, manage wait times and align resources. Tools that help them ‘measure stuff’ – more analysed, usable data – is the answer. We are nearly there with electronic health records and more data is being captured within the hospital setting every year. But systems need to gather data at the granular level, giving a clearer view of what works and what doesn’t. Only then can the power of technology help improve day-to-day efficiencies.
Interoperability will play a big part here but the healthcare market is unsure when this will come to fruition (3 years? 5 years?). Increasing use of SNOMED CT – a general system of terminology for electronic health records – along with a greater push within the industry to open out APIs (Application Programming Interfaces) should help with this.
Communication – between teams and stakeholders
It helps to split this issue two ways:
Healthcare pathways currently work in silos. System-based Electronic Patient Reports will help subvert gaps in communication, especially between primary and secondary care, by bringing power to the patient as the central player in the clinical pathway. Current systems do not allow this, creating points where different systems have incomplete information. Joining up this data via post or email can take time and elongate the patient’s procedures.
It is common for patients to transfer to different facilities, especially when a patient requires specialist treatment not offered locally. Unified systems and information transfer across facilities and businesses will smooth out and speed up these transfers.
Today’s hospital patient experiences multiple touch-points and different clinicians for a single treatment. In fact it’s possible that a patient may interact with up to 50 different people during a single treatment procedure. For a patient’s information to be shared and transferred successfully, the team’s communication platform must take centre stage.
Teams have been known to resort to WhatsApp groups. This of course does not provide the necessary features to make the conversations patient-centric, let alone provide the necessary security to protect patient data.
Only about 10% of information transactions occur through the electronic medical record. In some clinical units like the emergency room, where a large number of staff are physically co-located and engaged in teamwork, traditional, verbal and paper-based communication accounts for nearly all information transactions. In one study, face-to-face communication between staff represented almost 90% of all the information transactions measured in two emergency rooms.
In a retrospective review of 14,000 preventative in-hospital deaths, communication errors were found to be the lead cause, twice as frequent as errors due to inadequate clinical skill. Furthermore, about 50% of all adverse events detected in a study of primary care physicians were associated with communication difficulties.
Clinical decision support
The evolution of clinical decision support technology requires companies to focus on prevention, predictive analytics and bypassing unneeded processes.
Our portfolio company Michelson‘s product allows clinicians to see below the skin surface without the need for biopsy. There are digital medical imaging start-ups automating the tasks radiologists traditionally spend time on, such as Kheiron, which provides a breast screening product. Improving the efficiency of a radiologist by automating the screening process also improves accuracy and specificity for patients being diagnosed with breast cancer.
Predictive analytics is set to impact the clinical decision support systems. The power of AI prediction – for example, predicting when someone is likely to deteriorate or when a deadly acute condition may occur – is beginning to take its place in healthcare: DeepMind and Transformative being just two examples.
Ruling out heart attacks earlier means patients can be discharged from A&E sooner. Advancing tech will mean fewer patients need to be admitted to hospital for testing or observation. NICE estimates that using these tests with the earlier discharge protocol has the potential to save the NHS £30 to £40 million per year.
Clinical decision support tools still have a way to go and questions remain as to how they should best be integrated into the healthcare pathway. Their potential to improve clinician’s accuracy and efficiency in diagnosing illnesses is, however, hard to question.
With the movement towards value-based care in the US and globally, hospitals are having to provide more than just standard care to patients. These value-add products can be vital competitive advantages to hospitals within their market.
Being able to predict and improve the flow of patients within a hospital is an easy win when trying to digitalise hospitals. For hospital managers, being able to understand service delivery, reduce wait times and increase efficiency all the way to discharge, are key metrics for success.
Seasonal differences complicate matters: every winter the NHS needs thousands of extra hospital beds during the colder weather. Overall, the minority of patients (10%) stay in hospital for more than seven days, yet they use 65% of all beds. Redesigning patient flows to make sure patients are given fast high quality care and are discharged efficiently can be an easy win for hospitals.
In the area of A&E, the four hour target for patients to be seen, diagnosed, treated and discharged has never been universally reached. (According to the Nuffield Trust, 126 out of 137 major emergency departments missed the four hour waiting time and the government mandate in 2018/19 put this on hold as the NHS is drastically below target at 84%). Companies such as Draper&Dash provide hospital analytics and bench-marking that could help improve these metrics.
Preventative medicine, hospital admissions and re-admissions
The digitalisation of hospitals goes beyond the hospitals themselves. Tech looks set to have huge impacts in the prevention of disease and the maintenance of good health.
A self-care model in which people are able to diagnose and treat themselves so they need not visit a hospital in the first place is an important part of healthcare systems’ future strategies. Unpredicted, short notice admissions represent about 65% of hospital bed days in England (34 million bed days and 4.75 million emergency admissions in 2007/8*) and there is an estimated £13bn that could be saved here**.
Foundation year doctors (FYDs) write most hospital discharge communication, although they have minimal training in this skill. Poor quality discharge summaries increase the risk of adverse events and re-hospitalisation.
Our investments into Medisafe and Big Health, helping improve medication adherence and mental health respectively, are two examples of technology that is helping keep people away from hospitals in the first place.
Education and systems that incentivise patients to self-care for their conditions can significantly reduce hospital admissions. A study by Tapp et al (2007) showed a significant reduction in hospital admissions for adult patients with Asthma due to better education. This will always be a contentious issue, but real-time evidence of improvements and self-managed action plans are beginning to produce results.
Post-hospital care, preventing readmission, is another area where tech will play a significant part. In the US, nearly 20% of Medicare beneficiaries are readmitted to hospitals within 30 days of discharge, costing $41.3 billion. It’s estimated that preventable hospital readmissions are costing insurance providers, such as Medicare, more than $17 billion annually. Medication adherence, greater knowledge and at-home diagnostic and treatment tech could affect these statistics dramatically.
What are the problems and risks for startups?
Startups face a number of hurdles in breaking into the healthcare realm – more so than almost any other sector.
The complexity of the healthcare system: its logistics, barriers to entry and compliance structures make it hard to penetrate and integrate into. UK start-ups having to go into separate trusts as a procurement strategy defers innovation within hospital systems and so is not the right way to scale. Looking to organisations in the US such as the Mayo Clinic and the Cleveland Clinic, which operate highly active R&D centres, could provide models for innovation in digitalised hospitals in the future.
A historic lack of structure and culture to promote new technologies: value-based care is the NHS’s long term strategy and a current focus in the US, so culture and infrastructure should change to show this. Fear of the new will be outweighed by the need for innovation.
The risks of failure: these can be, literally, life and death. In few other sectors are the stakes as high. ‘Fail fast’, the entrepreneur’s usual dictum, cannot apply here, once technologies are adopted.
Tunnel vision: tech for tech’s sake, not the user’s is a danger here. A broad contextual understanding is always needed. A focus on the end-user – the patient, nurse, doctor or support worker – should be primary.
The digitalisation of hospitals is just beginning. Progress is necessarily cautious, but the signs indicate that as digitalisation happens, our definition of ‘hospital’ is set to open up. This will almost certainly include the home and the patient’s own ability to participate in their own diagnosis, treatment and disease-prevention.
*(Hospital Episode Statistics 2007/8)
** (NHS Connecting for Health report, 2010)
With thanks to:
Luke Gompels, Josh Allison, Vikram Palit, Gurprit Pannu, Ben Bridgewater, Martin Lees, Terry O’Neil, Harpreet Sood, Stephen Cone, Wayne Shirt, Laura Mountain, Rory Shaw, Janet Morgan, Leanne Summers, Ella Coiera, Chris Papadopoulos, Gilad Evrony, Anne Lo
Want some help asking the right questions to help your business to succeed? Get in touch with our experienced team at Octopus Ventures.
Image credit: Michael Browning