Digital for all now

Saint-Joseph hospital: shorter queues in Accident & Emergencies thanks to business intelligence

Econocom 21 Dec 2015

Saint-Joseph’s is a private, non-profit hospital employing just over 2,000 people. Located on Paris’ Left Bank, the hospital specialises in short stays. In 2012, Saint-Joseph set up a Business Intelligence platform with the aim of managing the hospital more effectively and making cost savings. So how can business intelligence be applied in the healthcare sector? How has data enabled Saint-Joseph to reduce waiting times in the A&E and control costs more effectively? We found out from Olivier Boussekey, Director of St. Joseph’s IT department. 





Why did you decide to use Business Intelligence?


The hospital world is a complex one: lots of people are involved and there are a lot of regulatory and economic constraints. The only funding hospitals like ours get is from health insurance reimbursements. Everything we do is covered by an agreement signed with the national health system, which means we don’t charge any extra fees on top of what is reimbursed by social security. Consequently, when we admit a patient, we’re paid a fixed rate which is always the same, whether the patient has one, two or ten X-rays and whether they stay for two days or ten days. So we really need to control our processes and track our income and costs to make sure we break even. As we’re a non-profit organisation, at the end of the year all the hospital’s profits are invested into our premises and equipment. And if we’re in deficit, obviously we have to borrow to pay off our debts.


We have a paperless policy. Like many hospitals, we digitalised the whole administrative/invoicing side a few years ago. We then moved on to the technical equipment: medical imaging, laboratories, operating theatres, etc. In 2011, we computerised part of our treatment processes, starting with the medication circuit, which was a particularly complicated project. The ultimate aim is for professionals to handle patients without using any paper.


In 2012/13, we realised we had more and more computer data and that we could probably use that information to understand what are incomings and outgoings are and how our various departments are organised.


“The main reason for implementing this BI platform was to optimise our processes and profitability.”





How can a BI platform help you increase profitability?


“From an economic standpoint, BI helps us get a better understanding of our expenses and thus enables us to get rid of any unnecessary ones or at least, reduce their impact.”


When a patient is hospitalised, we have to perform various examinations before we decide on the treatment. Without the results of these examinations, we can’t do a thing. If the results take two days, that’s two wasted days, for both the patient who’s waiting for their treatment, and for the hospital. So the first thing we decided was that any lead-times for obtaining information had to be reduced.


The second question we asked ourselves is, are all the examinations necessary? Some young doctors tend to order lots of tests to reassure themselves, even though hospital best practices require fewer. There are also certain departments that order too many, some of which are unnecessary. In order to measure this, we need figures and lines of research so that biologists, radiologists or dispensary staff can see where examination requests come from. This of course has to be combined with medical expertise.


“The figures won’t tell us to stop the examinations. But the data we have access to will help us make the right medical decisions.”



Tell us about the deployment of the BI tools


We had data in several different systems: prescription software, imaging software, laboratory software, etc. Practically all the software vendors have a back-office for exploiting the data. Our issue was consolidating, centralising and linking up the data and different systems.


Microsoft has signed a master agreement with all France’s hospitals. For an annual flat rate, we can use almost all the Microsoft licenses. For us, there are two advantages to choosing this provider: not only does it save us paying an additional license, but the data is stored natively in Excel, a program all our users use.


To set up the data warehouse and the whole BI platform, we went through a third-party provider. Together we decided which data we needed to get from our systems, how often we needed to extract it, the aggregation and correlation levels, etc. We began with the hospital activity, i.e. the admissions and movements in the different departments and rooms, the length of hospital stays, patient demography, etc. We then gradually broadened the scope of our analysis.





What sort of obstacles did you come across when you were deploying the technology? With the staff, for example?


BI has offered possibilities to doctors and other healthcare professionals who wish to optimise management and reduce the length of hospital stays. Avoiding unnecessary examinations is in everyone’s best interest: the patient’s, the laboratory and technical staff, and the hospital in general. The problem is knowing where to start!


“BI enables us to make decisions faster and more efficiently.”


The first difficulty we came across was to do with the data models in our systems: software vendors don’t really like to explain how to extract information from a database. So we have to do our own reverse engineering on the databases we use.


Another obstacle was interpreting and qualifying the data. If we take the example of the length of A&E queues again: it’s in both the patient’s and the hospital’s interest for this to be as short as possible. The question is how to do this. The point of using BI is to decipher and understand what the wait is and manage to break it down.


The first decision was at what point you should start measuring the wait: we have no way of knowing when the patient actually arrives at the hospital, so all we have to go on is the time their admission is recorded. The patient then sees an admissions nurse who determines the degree of emergency, then there’s the first diagnosis and examinations – CT scans, MRIs, X-rays or lab tests – then another wait for the  patient to be discharged: internal documents, admission for a stay  in another hospital. It was essential that we all agree on how to break down the time.


“Not everyone necessarily understands data and there are difficulties with interpreting it.”


There are different ways of looking at A&E waiting times. From the hospital’s point of view, as long as the patient is on the premises, they’re taking up time in a cubicle and thus stopping the next patient from being treated. The patient’s point of view is different: usually, once they’ve seen a doctor, they feel relieved and taken care of. So it’s this wait that counts for them. If they then have to wait for a CT scan or an MRI, it’s not a problem. This interpretation and the way we measure, count, present and aggregate this data in the dashboards are difficult to establish.





What’s your initial verdict on this platform?


It’s pretty positive, although we haven’t yet applied it all across the board: we haven’t incorporated cost accounting into our BI tool yet. The aim is ultimately to be able to say: today, we have 15 operating rooms. If we had 16, how many surgeon hours would that be, what sort of procedures, what medical devices, what examinations, and what would the financial impact of this change be? Is it a good idea to up the number of operating theatres? Do we have enough patients to fill up the new theatre?


This is the kind of analyses and decision-making based on both an overall and detailed view of our processes, structures, organisation and costs that we’d like to achieve. And we’ve got a long way to go yet!





What advice would you give other hospitals launching a similar project?


I’m all for adopting these types of tools in the most pragmatic way possible. With the Microsoft technologies, there is minimal disruption to end-users: everyone knows Excel, so staff don’t have to learn how to use new tools or technology or concepts.


However, it’s vital with this sort of project to get business experts involved from the outset. A doctor, CFO, accountant and financial controller don’t necessarily have the same perception of the number of hospital stays.


“Using data is fine but you have to know what you’re dealing with!”


We’ve managed to get quite a few people involved. But possibly not enough, because we wanted to move a bit faster than our experts could. That meant we were present them things that they hadn’t anticipated and so they found it hard to agree that our extractions and aggregations were in line with their vision.


Also, the fact that we’re there to treat people and not to do business is a major hindrance to this sort of project: it’s not always easy to translate everything into figures. But that doesn’t mean we can’t do it: on the contrary, just because we’re looking after patients, it doesn’t mean we can’t keep an eye on our spending. When we do fewer examinations, we spend less, but we also don’t bother the patient unnecessarily – so it’s better for everyone!



==> Read our other profiles of digital makers in the healthcare sector:

Olivier de Fresnoye: using open innovation to boost cancer research

– Lionel Reichardt, Pharmageek: I prefer the concept of the modified self to the quantified self

Raphael Master, Microsoft: hospitals should industrialise their digital transformation

Talk to us and share your Digital Maker’s lesson with us!