Periodic Reporting for period 4 - HealthcareLabour (Empirical evidence on the impact of the labour market on the production of healthcare and health)
Okres sprawozdawczy: 2023-04-01 do 2024-09-30
The focus on the research is labour markets in healthcare and the effect of the behaviour of key employees in these markets on the quality, quantity and distribution of health. The growing body of economic theory on motivation and incentives (for example, Besley and Ghatak 2005, Benabou and Tirole 2006, Delfgaauw and Dur 2010, Francois 2000) has proved important in helping us to understand the factors that affect the behaviour of public service suppliers. However, empirical tests of this literature remain relatively rare. The aim is to fill this gap.
The focus of my research is on the micro-foundations: the role of key agents in contributing to the persistent variability in productivity, the take-up of innovation, and the variability in quality within and between healthcare suppliers and the implications for users of healthcare. The issues addressed are important in any labour market in which human capital is central to production and where there are market failures which mean government intervention is important and society cares about critical outputs. The focus to date has been on the behaviour of two groups - senior clinicians and management - both of whom are central in the production of healthcare. Similar agents are found in all areas of public (and many private) services production, so our findings will have implications more generally for public service production.
The size of the healthcare sector in the economy; the importance of the output to social wellbeing, the ability to exploit policy changes to create ‘natural’ experiments and the availability of large administrative datasets makes healthcare an ideal test-bed. Healthcare accounts for at least 10% of GDP in almost all OECD countries and this proportion is steadily rising. Thus understanding how healthcare production can be improved is key to social welfare and, because healthcare is often paid for by the public purse, to the fiscal position of the government.
The research has shown the importance of non-financial incentives in driving the behaviour of healthcare suppliers and users. It has shown that in the case of adoption of new technology, and in established practices, providers respond to the behaviour of their peers, their work networks and to new information that is available to all. For consumers of healthcare, it has shown that access to information via the internet changes behaviour, that common information through the internet has different impacts on different individuals, and that responses to such information is not always beneficial for the healthcare system or for individual health. It has also shown that public policy changes that mimic private sector corporate governance do not translate well into public hospitals in the UK.
The focus of my research is on the micro-foundations: the role of key agents in contributing to the persistent variability in productivity, the take-up of innovation, and the variability in quality within and between healthcare suppliers and the implications for users of healthcare. The issues addressed are important in any labour market in which human capital is central to production and where there are market failures which mean government intervention is important and society cares about critical outputs. The focus to date has been on the behaviour of two groups - senior clinicians and management - both of whom are central in the production of healthcare. Similar agents are found in all areas of public (and many private) services production, so our findings will have implications more generally for public service production.
The size of the healthcare sector in the economy; the importance of the output to social wellbeing, the ability to exploit policy changes to create ‘natural’ experiments and the availability of large administrative datasets makes healthcare an ideal test-bed. Healthcare accounts for at least 10% of GDP in almost all OECD countries and this proportion is steadily rising. Thus understanding how healthcare production can be improved is key to social welfare and, because healthcare is often paid for by the public purse, to the fiscal position of the government.
The research has shown the importance of non-financial incentives in driving the behaviour of healthcare suppliers and users. It has shown that in the case of adoption of new technology, and in established practices, providers respond to the behaviour of their peers, their work networks and to new information that is available to all. For consumers of healthcare, it has shown that access to information via the internet changes behaviour, that common information through the internet has different impacts on different individuals, and that responses to such information is not always beneficial for the healthcare system or for individual health. It has also shown that public policy changes that mimic private sector corporate governance do not translate well into public hospitals in the UK.
During the grant period I have advanced research that contributes to understanding of the impact of non-financial incentives on healthcare production. The focus has been on identifying causal effects of changes to information and incentives on the behaviour of healthcare suppliers and users.
Strand 1 is a set of projects that initially focused on the impact of networks and information in adoption of medical innovations and then broadened to consider the role of information on medical production more generally.
The initial focus on adoption of innovation was driven by the fact that medical care sector is an important source and user of technology and its uptake has important effects on costs and outcomes and networks are important in shaping individual and firm behaviour, including technology adoption. But despite this, there is little empirical research on the effect of medical networks in explaining variation in adoption of innovation, partly because is it empirically non-trivial to determine the causal effect on networks on behaviours. Our research has shown that a physician's network size and the adoption behaviour of their peers affects the adoption of a new and cost-effective technology in surgery. We also provide evidence of whom should be targeted in order to increase uptake of innovative surgical procedures. We also show that fast adoption is not always optimal - sometimes it pays to be more cautious with a new technology. In complementary papers I examined the impact of news conveyed through the internet on the behaviour of users of services. We have examined the effect of false medical news about the safety of the MMR vaccine on childhood immunisations in the UK, looked at the impact of internet access on the spread of COVID cases in Indonesia. In both cases we find heterogeneity in behaviour, with greater responses by those who were more educated. We have complemented this focus on physicians and patients and examined hospitals and the publication of information about quality. We find that publicly disseminated higher hospital quality ratings are associated with lower deaths during COVID suggesting that quality ratings, which are often contested, can convey meaningful information. We also show systematic differences between for-profit and non-profit hospitals in quality, with the largest gap arising when for-profits hospitals are organised into chains and operate in uncompetitive markets.
Strand 2 focuses on the role of labour in hospital productivity and more widely on the medical labour market. To do this we exploited data sets that are newly accessible to researchers. With these we investigated the impact of a pension reform on the labour supply of senior doctors to the UK NHS and to examine the causal impact of staffing, team and hospital familiarity of nurses on a range of patient outcomes. Both papers show the importance of skilled labour and that short term agency staff or less qualified staff are not a substitute for degree qualified staff. I also examined whether reforms which implemented private sector governance models and strengthened market incentives for CEOs and top directors of public sector hospitals resulted in greater alignment of production with key government targets. My findings indicated that this model may not be the best model for operating public sector hospitals.
Strand 1 is a set of projects that initially focused on the impact of networks and information in adoption of medical innovations and then broadened to consider the role of information on medical production more generally.
The initial focus on adoption of innovation was driven by the fact that medical care sector is an important source and user of technology and its uptake has important effects on costs and outcomes and networks are important in shaping individual and firm behaviour, including technology adoption. But despite this, there is little empirical research on the effect of medical networks in explaining variation in adoption of innovation, partly because is it empirically non-trivial to determine the causal effect on networks on behaviours. Our research has shown that a physician's network size and the adoption behaviour of their peers affects the adoption of a new and cost-effective technology in surgery. We also provide evidence of whom should be targeted in order to increase uptake of innovative surgical procedures. We also show that fast adoption is not always optimal - sometimes it pays to be more cautious with a new technology. In complementary papers I examined the impact of news conveyed through the internet on the behaviour of users of services. We have examined the effect of false medical news about the safety of the MMR vaccine on childhood immunisations in the UK, looked at the impact of internet access on the spread of COVID cases in Indonesia. In both cases we find heterogeneity in behaviour, with greater responses by those who were more educated. We have complemented this focus on physicians and patients and examined hospitals and the publication of information about quality. We find that publicly disseminated higher hospital quality ratings are associated with lower deaths during COVID suggesting that quality ratings, which are often contested, can convey meaningful information. We also show systematic differences between for-profit and non-profit hospitals in quality, with the largest gap arising when for-profits hospitals are organised into chains and operate in uncompetitive markets.
Strand 2 focuses on the role of labour in hospital productivity and more widely on the medical labour market. To do this we exploited data sets that are newly accessible to researchers. With these we investigated the impact of a pension reform on the labour supply of senior doctors to the UK NHS and to examine the causal impact of staffing, team and hospital familiarity of nurses on a range of patient outcomes. Both papers show the importance of skilled labour and that short term agency staff or less qualified staff are not a substitute for degree qualified staff. I also examined whether reforms which implemented private sector governance models and strengthened market incentives for CEOs and top directors of public sector hospitals resulted in greater alignment of production with key government targets. My findings indicated that this model may not be the best model for operating public sector hospitals.
The research has applied ideas more broadly used in the labour economics literature to the healthcare sector. It has been possible due to extensive use of matched administrative data sets and ideas from outside the field of health economics. It has extended understanding our understanding of the role played by networks for physicians, with a focus on adoption of innovation. It has also raised important issues as to the impact of management of complex organisations in public sector.