Predicting healthcare costs using bio-social data
Apostolos Davillas, Professor Stephen Pudney
Healthcare costs are rising worldwide – can social and biomarker data help policymakers work out where to target limited resources?
Health care costs have risen faster than economic growth in OECD (Organisation for Economic Co-operation and Development) countries, and this is projected to continue as a result of new medical technology, rising expectations and the increasing needs of the ageing population. In this policy setting, it is important for policymakers to be able to identify the sections of the population where costs are high and rising, to establish priorities for resource planning and preventative policy.
In Britain, the NHS spends about 10% of the UK’s GDP on health care, which is broadly in line with other European nations. However, if we compare the UK with other OECD countries, the NHS has less than the average number of doctors, nurses and hospital beds per head. The UK’s performance on some health measures is also below average – survival rates for breast and cervical cancer, for example, and preventable deaths.
So, with limited resources, it is important to be able to identify the sections of the population where costs are high and rising, so that policy makers can work out where to target resources and preventative efforts.
Bringing different data together
For our working paper on this issue, we used Understanding Society data, looking at over 2,300 adults who were apparently healthy in 2010/11, and had not been diagnosed with any long-lasting health condition. We also had a set of nurse-collected and blood-based biomarker data for them, collected at that point. Five years later, we looked again at how many times they had seen their GP, and been a hospital inpatient or outpatient each year over that period. Having biomarker data at the beginning of the five-year period allows us to explore the power of this information for predicting health service use and the resulting cost to the public.
By bringing detailed social, economic, and health care use data together, we can also estimate the impact of differences in people’s socio-economic characteristics on future health service use and the corresponding public cost. We do this using administrative NHS cost data and individuals’ responses on Understanding Society questions on GP and outpatient consultations as well as inpatient days.
What we found
We found that a biomarkers summary measure, which captures several dimensions of a person’s physical health, is a powerful predictor of what their health care eventually costs. Specifically, we found that people with elevated biomarker levels will add about 18% to NHS costs in five years’ time, compared to those with normal biomarker levels.
In addition to the expected strong effect of ageing on costs, we also found a large gender difference, with women experiencing at least 20% higher health care costs than comparable men – because women visit the GP and use outpatient services more than men do. There is also a strong education gradient in health care costs. For example, the five-year-ahead costs are about 16% higher for those with no educational qualifications compared to those with intermediate qualifications.
What does this mean for policy?
These results have potential implications for health policy in the UK and beyond. They can be used to indicate priority areas for interventions, such as screening programmes and health education initiatives, which could control future treatment costs for those who have not yet reached the stage of being diagnosed, but who are at risk of generating higher future health care costs.
The results could be used to work out priorities for preventative care such as screening programmes and health education campaigns to control future costs before people need treatment. For example, the NHS Health Check programme is currently available to everyone aged between 40 and 74 and, thus, targeted only at certain age groups. If these checks were targeted more precisely to make sure people at risk take up the opportunity, the programme could identify better the parts of the population which have the highest potential future healthcare needs and costs.
This research also has implications for the GP reimbursement system in the UK (and several other countries). At the moment, GPs’ budgets are allocated using capitation payments – giving practices a set amount for each person enrolled with them, per period of time, irrespective of an individual’s actual use of the health service. If the capitation formula was more tailored to patients’ morbidity data and other characteristics, this could mean better allocation of resources and better health outcomes.
Apostolos is Senior Economist at the Office of Health Economics, researching and publishing in the area of applied health econometrics, and focusing on evaluating public health and social care policymaking
Stephen is Professor of Econometrics at the University of Sheffield, working mainly on health and disability issues, and former Director of Research at the Institute for Social and Economic Research