Richard Smith: Using data to improve care and reduce waste in health systems


Using data to improve care and reduce waste in health systems

Annual expenditure on health care in the United States is currently $2.8 trillion, and about a third of it is wasted, says the Institute of Medicine.  The sum wasted is about five times the GDP of Bangladesh, a country of 160 million people. This is waste on a spectacular scale, and reducing it while improving the quality of care is the main aim of the information technology developed by Optum, the services part of the UnitedHealth Group, said Richard Migliori, a former transplant surgeon and Chief Medical Officer of the UnitedHealth Group. I don’t come to tell you what to do, said Migliori speaking last week to the Cambridge Health Network, but I hope to at least elicit your sympathy.

Waste in the NHS

Nobody seems to have attempted such a thorough investigation of waste in the NHS as the Institute of Medicine has done in the US, and it’s unlikely to be as high as a third as much of the US waste is driven by excess capacity—too many hospitals and perhaps too many doctors. But Britain also has too many hospitals, and some of the causes of the waste will apply in Britain. About half of the waste, said Migiori, is accounted for by “doing the wrong things and doing the right thing wrong.” We know that Britain has inappropriate care and a level of medical error the same as the US. About 7% of the US waste is missed opportunities for prevention, and Britain certainly has those. Fraud accounts for some 10% of the waste, and, as the Economist has pointed out, health care is very attractive for fraudsters because “it’s where the money is” and controls are usually much weaker than in banks, where the money used to be. Britain also has fraud, but it won’t be on the same scale as in the US. Nor will Britain have the same waste from excessive administration (25% in the US) or inflated prices, particularly for drugs (14%)

I was left thinking that one of the British health think tanks should surely do a detailed investigation into waste in the NHS, particularly when the NHS is under such financial pressure. Peter Carter, the chief executive of the Royal College of Nursing, who also spoke in the meeting, has suggested that NHS waste might be £5 billion on a now annual spend in England of about £110 billion, but his figure was not based on a detailed investigation.

The need for good data

One of the reasons that the Institute of Medicine could so such a detailed examination is because the US has much better data on what is happening in the health system. These data are a product of the claims based system of payment in the US, something that no doubt also contributes to the excessive administration costs. But famously “if you can’t measure it you can’t manage it,” and, although the data may contribute to the excess cost, its analysis and intelligent use is the key to reducing the waste.

England has attempted to gather such data with the project that came unstuck earlier this year, and another attempt will be made. Carter referred to how the programme was mishandled in his talk, and Migliori’s comment from US experience is that once you can show real patient benefit from the use of data than anxieties about confidentiality and privacy fade. But he also pointed out that UnitedHealth takes data security very seriously, storing data underground in tectonically stable areas that and exceeding the security levels achieved by financial institutions.

UnitedHealth has 6.3 petabytes of data on 149 million patients and has up to 19 years’ of data on some of them. (In case you are wondering, a petabyte is 1 000 000 000 000 000 bytes, and all the academic libraries in the US have 2 petabytes of data in total.) These data have been endlessly cleaned and provide an invaluable base for developing tools for reducing waste and improving patient care. UnitedHealth, a company with a turnover of $122.5 billion, is investing $2 billion a year in developing information tools.

As its critics observe, UnitedHealth is a for profit company listed on the New York Stock Exchange, and it has to be sure that it can get a better return on that investment that investing the money elsewhere. No doubt the high level of waste in the US health system offers bigger financial opportunities than are possible in Britain, but the point that the audience picked up is that there is a hard financial case for such investment. Yet, as some in the audience said, many British health institutions are reluctant to invest in information technology. They have been scared off in part by the failure of the care record part of the National Programme for Information Technology in the NHS but the future of the NHS—and of Britain’s continuing success in life sciences—will depend on producing better data, using it well, and investing in information technology. It is, of course, hard to invest when you are struggling to achieve financial balance, but it has to be done.

Five ways to use data in health systems

UnitedHealth uses data in five broad ways, explained Migliori, “a man who speaks in numbers” as Pam Garside, cochair of the Cambridge Health Network, observed. The first way is to simply share the data. Clinicians and managers can look at their population and know about their clinical conditions, activity in the health system, and costs. They can use the data to spot trends, improve care, see gaps in care, cut inefficiencies, and reduce costs. This brings to health care the kind of data that those managing many other organisations—for example, retailers—have had for years. My son is head of food for Byron (a hamburger chain), and he knows in real time how the burgers he develops are selling and which parts of the country like them best.

The second way to use the data is to analyse it. As an example, UnitedHealth takes evidence based recommendations of medical societies and sees which doctors are following them and which not. In this way it can measure the quality of the doctors. Next it will match the costs of doctors in particular geographic areas, and in this way it can identify doctors who are both high quality and cost effective—and one of the beauties of heath care is that high quality tends to go together with low cost because of fewer complications. The data can also be analysed to identify the patients who generate most of the activity and costs, allowing them to be offered more support.

The data can be used to improve workflows—for example, by identifying those at the highest level of risk and what needs to be done for them. It allows doctors to care for populations not just the patients who consult them.

A fourth broad way of using the data is through developing tools for coordinating care. About 10% of patients discharged from hospitals in the US are readmitted within 30 days, and for those over 65 it’s about 20%. Readmissions cause much misery and generate high costs. UnitedHealth has used data to identify those most at risk of readmission. These patient are then visited by nurse practitioners who help the patient stay in their own homes. Unsurprisingly the main reasons for readmission are not medical but social. The programme has led to a 28% drop in readmissions and a 7% increase in visits to doctors. Migliori sees the increase in visits as a good thing as care is likely to be improved. (UnitedHealth employs 1150 nurse practitioners, making it the largest employer of them in the US.) David Haslam, now chairman of NICE and a former general practitioner, asked how nurse practitioners could fix social issues, and Migliori conceded that they could do only so much—helping with food, drugs, and other remediable issues.

Research is the fifth broad way in which the data can be used, and through Optum Labs, a partnership with the Mayo Clinic, researchers can get access to the deindentified data on 149 million people, 7.8 million of whom are diabetics. Researchers have used the data to study the natural history of diseases, discovering, for example, that skin disease is often a forerunner of diabetes. It’s possible to identify control groups within the database. One study showed that the annual cost of treating a patient with chronic myeloid leukaemia is $110 000, but if they are treated with imatinib, a biological treatment that costs $36 000 a year, then the costs are $55 000 a year (including the cost of the drug), justifying treatment with a very expensive drug. This is valuable information to the patients, the payers, and of course the drug company.

Asked what would come next, Migliori said tools to support patient engagement—and showed us the tool on his wrist, which looked just like a watch and monitors lots of physiological data. The data can be gathered through mobile phones. Every United employee is offered one.

We know that after a year of treatment only about half of patients with long term conditions are taking their drugs, and Migliori was asked about using data to monitor adherence. It’s often easiest, he said, to measure adherence through the physiological effects of the drug, so some 20 000 patients with heart failure have scales that are connected to a central database. If patients are not taking their dugs then their weight will rise, and their doctors are informed. UnitedHealth is also experimenting with gathering data from an X box, which monitors movement.

Somebody from Monitor, the sector regulator for health services in England, asked about using data for pricing, fretting that pricing in the NHS is not transparent. Migliori answered that payers should pay for value, which means that a treatment is better or cheaper than other treatments or allows faster improvement. Data are clearly needed to determine the scale of the added value and so for effective pricing.

How much might waste be reduced?

So how much progress has been made in reducing waste, asked a sceptic in the audience? The UnitedHealth Center for

Health Reform & Modernization, which was directed by Simon Stevens (now chief executive of NHS England), has published a document that lists 13 changes that might reduce expenditure on health by the government programmes.  UnitedHealth has, for example, saved some $22m by  reducing hospital readmissions, and applied across the government system the savings could be $5 billion.

The main message for the NHS was that it will be essential to gather good data on what’s happening in the system in order to raise quality and reduce costs, a message that is already familiar to managers in US England but has not been transmitted effectively to the population. And Migliori’s advice is to show real benefits from the use of data and then grow data collection and use. The upside is substantial enough to overcome all but the most ardent doubters.

Competing interest: RS is employed by UnitedHealth and has shares and share options. He runs a philanthropic programme in low and middle income countries for the company and has no responsibility for UnitedHealth (now Optum) in the UK or for any of the commercial activities of UnitedHealth. He’s also the chair of and has equity in Patients Know Best, a company that links records from all parts of health and social care and gives control to patients. It has many contracts in the NHS. RS has written this blog because he is the “usual blogger” for the Cambridge Health Network.