Winter is fast approaching and acute trust informatics departments up and down the country will be gearing up to send their daily situation reports to NHS England starting on 1 December.
For many, this will be an onerous task that will involve pulling together a daily return with information about capacity that includes bed occupancy and availability, demand including emergency department attendances and emergency admissions, and any ED closures or diverts.
“Responsibilities around sit reps creates a significant overhead for informatics departments,” says Dr Mark Davies, medical director of MedeAnalytics. Often these rely heavily on manual processes to the point that Davies refers to them as “hand cranked”.
There has to be another way to do this?
Several business intelligence suppliers now report that NHS trusts are asking whether there is a better way of doing things. They want not only to automate the daily sit rep data extract but also glean some insights that will help them predict and manage patient flows.
Barry Duke, BI Practice Lead at Emis Health says: “Trusts are increasingly using BI to provide whole system views on hospital activity by moving away from departmental datasets to hospital-wide data marts to produce integrated indicators on demand and capacity.
“To do this they are developing enterprise wide BI platforms incorporating data from multiple source systems. These systems include ED, the patient administration system, bed management and pharmacy for example, as well as HR and e-rostering systems for staffing provision.”
Belfast Health and Social Care Trust is one of a number that has been exploring the power of such systems. The trust, which sees more than 200,000 people annually in its EDs, has implemented Emis Health’s Business Intelligence platform, built on Microsoft SQL Server 2012, which connects to the reporting database within the ED system, EMIS Health’s Symphony.
Clinicians and managers now have a set of interactive dashboards to help them analyse the information they need. Their key performance indicator dashboard refreshes every few minutes, giving them near real-time visibility of service pressures, and supporting their teams through crisis management during busy periods.
The dashboards also enable users to ‘slice and dice’ data to view recent ED activity, apply filters and identify trends to help allocate resource more effectively and ultimately improve patient care.
Dr Brian Armstrong, co-director of unscheduled and acute services at the Belfast trust, says: “What we have now is effectively like an air traffic control dashboard of current emergency department performance.
“It helps clinicians and managers to take both a strategic and operational approach, using live data to improve areas such as patient waiting times within the emergency department.”
More data, from more places
The move to integrate BI in this way goes beyond the hospital boundaries, though. Duke explains: “Organisations that have robust BI platforms in place are looking at more sophisticated analytics. These include predictive modeling, based on previous trends and seasonal variations.
“They are also reaching out to seek collaboration with organisations within their health economy. Data held in primary, community and social care can be shared and fed into those models to provide holistic views on demand and capacity.
“They can then inform operational managers with predictions and alerts so that preventative measures can be taken to manage patients at risk of hospital admission in primary and community care.”
Davies warns that making this shift is going to require a complete change in thinking. “Lots of people are talking about this, but in all honesty we are in the early days of the kind of thinking that will change the approach to using BI to drive operational efficiency,” he says.
Wayne Parslow, UK managing director of MedeAnalytics, explains. “Most of the data that hospitals have is about the machinery of running the hospital. Value will only come when we have a person-centric view of data from all sides.”
Predict and survive
Very few acute trusts currently have access to that overview of individuals across a health economy.
Parslow asks: “Where is the community data? Who is living in poor quality housing without heating and therefore at risk of emergency admission? All of this feeds into demand for A&E services.”
Davies adds: “There is a really fundamental point around this. Looking at how the A&E department functions in isolation is a mistake. Hospitals are barometers of how the community is functioning.
“If you create an information architecture that is embedded around individuals, in the way we do at MedeAnalytics, then you can start to create a much more dynamic picture and start to ask questions of the data.”
Beverley Lewis, senior sales executive at Connexica, shares this vision. “With the cold. Wet, British weather comes an increased number of hospital admissions; particularly via A&E departments,” she says. “Is it possible to predict this demand and proactively manage it?
“Many NHS organisations are already planning for this and have been for a number of years. But can BI influence the success of these initiatives?”
Very much so, she says. “The NHS has a whole host of systems that are capturing data across primary, community and secondary care, so surely there must be some answers in there somewhere. NHS organisations are using this data to identify patients at risk – by analysing historic data and capturing the specific cohort of patients that did result in a hospital admission.
“Trusts can then use this information to target patients with similar criteria and conditions to define and deliver proactive care plans.”
Meanwhile, back in the ED, BI software could deliver real time views of activity that may influence patient decisions.
Lewis says: “The patient can see the number of patients waiting and make an informed choice as to the hospital they decide to visit. In circumstances were the visit may not be critical, the patient may decide to seek alternative options and relieve the pressures on their local hospitals.”
She argues that modern BI supports both scenarios and indeed Connexica is already working with one of the Welsh Health Boards to deliver just this kind of BI system.
She says: “Unlike traditional BI, new state of the art technology provides the user with the ability to perform ad-hoc analysis over any combination of data fields without the need to design and implement a fixed set of dimensions or measures.”
She agrees that it is a shift in thinking and culture and not as simple as implementing a data warehouse with BI software. “I think some people in informatics departments find the software a bit of a threat,” she says. “They are paid to write the queries. We need a change of mind set.”
Forecast: fine
Stepping back from this big vision for a moment, and trusts are starting to use predictive analytics in combination with their traditional data.
Graham James, CACI’s vice president and director of business intelligence, explains: “Trusts are using more data to look for correlations between days of the week and months of the year and, particularly in A&E, weather conditions. They are starting to use this to predict what the impact will be on services and bed requirements.”
Davies agrees. “The science of algorithms is moving on apace,” he says. “We have predictive algorithms that relate to the weather and COPD admissions, algorithms predicting falls and others predicting mental health admissions. These will have a significant impact in the near future.
“But if we continue to look at hospitals in isolation, then we will continue to treat the symptoms and not the cause.”
So for this year, many acute hospitals will be looking at the weather forecast with one eye and the ED front door with the other; and hoping to avoid a repeat of last winter’s chaos. A handful will be hoping to be ahead of the game with their live dashboards and predictive modeling in place. Next year, maybe a few more will join them.
|
