Dr. Stephen Kinsella:

My submission is on a body of work Dr. Rachel Kiersey and I carried out for the Health Research Board. It is a review of five different workforce planning models and consisted of an evidence-based review and a series of interviews with the people who make the models in these countries. A workforce model is a series of forecasts based on a baseline data set of the number of doctors, nurses, physios, etc. in the system and what one thinks the demand for the services they provide will be over a given period. It might be the number of doctors or physiotherapists in Limerick University Hospital in 2025, and it might cover the integrated care pathway patients can expect to go through to get the highest level of care and the requirements for same.

It is also about how these services change over time. An example that kept coming up involved dialysis nurses across four of five countries. When technological change takes place, certain service specialties tend to be used to a lesser extent. Technology means there is less demand for nurses specialised in dialysis because people can dialyse themselves at home to an increasing extent. How can the system know that it should reduce the number of training places for dialysis nurses? How does one know whether one is producing enough or too many doctors? It is about the connection between the demands of the system in five, ten or 15 years, the supply and also the mitigating factors. For example, we know that migration is a major issue across all health professionals, including nurses and doctors. Most of these models involve very simple modelling. It is just forecasts being formed. What is really difficult is getting the data on a fine-grain level and obtaining a very qualitative understanding of what is going on.

I will take the committee through what we found. I ask members go to the results section of the paper. We found that the onset of health workforce planning emerged independently in each country we studied. We studied Australia, New Zealand, Scotland, Wales and the Netherlands. Each time it happened, it was in response to some pressure. Either it was demographic pressure, financial constraints or issues surrounding future supply. In several countries, for example, it was the lack of available nurses that caused people to ask how many nurses were being trained and how many were being bought in from abroad. There are very few integrated national care models. Many take modules and only look at doctors, nurses or allied health care professionals, respectively. Very often, they start with doctors, for various reasons. It is very interesting that once one has moved beyond doctors and their sub-specialties, one sees that there are is a great deal of thinking and planning to be done. Data are always the first problem. Typically, we do not know what the data are on a given day. If one stopped the health system today and asked the chief information officer of the HSE how many people were working that day in the health system, it is most likely he or she would not know the answer. If one asked him or her how many doctors or phlebotomists the service would require in Temple Street in 2020, he or she would not be able to provide an answer. That gives a sense that the planning system is short-term in nature. While there is a workforce planning model in place right now, everybody agrees that it could change.

Each workforce planning model that we studied was country specific. That is to say it was generated with respect to the institutional structures that underlay each individual country's needs. Australia is completely different from New Zealand institutionally and similarly Wales and Scotland, notwithstanding similarities in terms of the National Health Service, NHS, have quite different financial and legal structures. However, they all begin in the same way. One creates a baseline analysis of how many doctors, nurses, etc., one has in a particular year and then forecasts based on demographic supply. We know our population is ageing and that we have both a very young and a very old population and that will increase over time. We are aware that demands on the service will change. We are also aware that there are regional differences in the numbers that are accessible. Once there is sight of the regional data and the patient flow data, it is possible to analyse the evolution of our system, which is very important.

When I started lecturing years ago, people said, "Get the summary in first, in case people are not sufficiently caffeinated." Here is the conclusion. This is not just a quantitative process. That is, it is not just the case that some nerd in a room comes up with a forecast, a line goes up and that is fine. This is a qualitative process, which is to say that it is a way for the Department of Health and the Health Service Executive to initiate a dialogue across all the sub-specialties. As an aside, I note that I will be setting out some thoughts on governance structures in a while. This qualitative process starts with saying, "Here is our forecast. Is it right?" It is about asking the individual services and specialties whether it is correct. Every time, of course, they will come back and say it is not and that more of specialty X is required or less of specialty Y. However, what is really valuable is being able to have that dialogue. This was a real surprise to me as a technocratic, quantitative person. In all of the expert interviews we did, this qualitative process and the idea of generating fora surfaced. It means one could get the deans of medicine of all the major colleges and universities, a representative of the Department of Justice and Equality, who will talk about migration issues, and representatives of the Department of Health and the HSE around a table with the Departments of Finance and Public Expenditure and Reform. If the medium-term forecast to 2021 is that health spending will increase by 7%, we can drill down and ask the different specialties how much of each resource they need and where, if we had the money, resources should go. We can then ask the colleges if they are training the necessary personnel. Before the Fottrell report and the massive increase in the number of doctors we produced, we were not producing enough doctors. Now, it seems like we are producing too many. The system needs a check and a balance. There needs to be feedback into the system to slow things down.

In many of the cases we studied, clear and legislated connections to policy levers were very important.

For example, if an individual service did not do its forecasting plan, it does not get funded. It is legislated in Wales that the medium-term strategy for workforce planning is allied with the financial plan, so if one does not do these forecasts one does not have one's service level agreement signed off. If one does not have the forecast carried out, that is it. Similar actions are being taken in New Zealand. In Australia, where it is most developed, people can tell one down to the hour what each pharmacist, doctor, nurse and every other sub-specialty is doing. They have activity data by the hour so they are able to compute productivity levels not just by hospital, but also by individual service. I am not suggesting that we go that route just yet. Perhaps we could find out where all the doctors were, although I am being slightly facetious because we have quite good data on that.

However, if we had a robust data gathering exercise, the modelling is not very difficult. What is difficult is getting these fora established and having them repeat. In Scotland, for example, when it was rolled out there were 150 individual contact sessions between the workforce planning unit and the relevant sections. We are not just talking about the workforce planning unit talking to the HR manager of a hospital but to all of the people who run the services, who might say they need three administrative personnel rather than seven, that they do not need ten nurses but two nurses and six more doctors and so forth. One might think that system could be gamed. I am a university lecturer and if somebody asked me how many more university lecturers are required in economics, my answer would be "all of them". One would imagine that people who hold budgets would say that. In fact, it is not true. Most of the time if one is asked for a fairly credible forecast, typically one gives it. People are not that incredible, especially if they are held to it - "Last year you said your expectation was that you would need eight physiotherapists, but you did not need eight because you have coped perfectly well with four. I know this because I have latent demand data and I know that there was not much of a waiting list and it was cleared quite quickly." On the other hand, the service might not be sufficiently resourced with support staff. This is a pretty big issue in certain areas. The support staff is not there, so one has highly qualified medics, nurses and other allied health care professionals running around finding labels. Much of this is about balancing and understanding the team composition.

This qualitative aspect is really important. One of the matters I wish to impress upon the committee today is the need to think about what the structure of those engagements would be. What would that structure be like if one had to have the deans of all the colleges of medicine, all of the various representative bodies in the Higher Education Authority, HEA, justice and so forth there, as well as a sense that the funding will follow whatever strategic priority is set? The strategic priority is typically on a five to seven year basis, not one to two years, which means the funding must be set roughly in those parameters. As I will outline later, the Irish context does not put the lie to that medium-term planning approach, but historically it has been quite difficult to do it.

We found that the engagement with the whole workforce planning process across all the countries we studied, and particularly a discourse around the modelling itself, is really influential. It changes the wider health workforce policy, because it simply promotes a conversation. This allows stakeholders concerned with health workforces to become actively involved, even me. I am talking about doctors and nurses. Very few people ask how many phlebotomists we need. It turns out that this type of work gives what one might call the medical scientists a voice at the table, that they did not previously have because everybody is concerned that the public discourse is around doctors and nurses. In fact, one needs a large number of allied health care professionals to make the system run and creating these engagements gives them a large voice at the table which they would not have had previously. They get input into policies which they previously did not have and that is very good. These types of conversations give one a sense of being able to sense and identify trends in a system, and these trends in a system turn out to be very important. One will not get this data sitting in a room with a spreadsheet. One only gets the data by talking to the people who run the service; that is the only place to get this data.

That is the reason the governance structure I recommend is so important. If one locates a workforce planning unit completely within the Department of Health, the links it has with service users may be limited. If one locates it entirely within the HSE, it will not have the strategic element which the Department of Health provides. My recommendation is to have the workforce planning unit, division or department span both and have links into other areas, simply because the health system is so big.

As part of the context for our review, we looked at the evolution of the Irish health system to date. I have a panel of figures for the members of the committee which is included in the series of graphs I have provided. I am delighted it is done in colour because mine printed out in black and white and is completely illegible, so I was a little worried we would be looking at the wrong things. These figures are indexed figures from the Department of Health, the Department of Public Expenditure and Reform and from the HSE's annual reports. I am seeking to give the committee a sense of what has happened to the evolution of the Irish health system since the mid-1990s. The members are looking at these numbers indexed; that is, one can imagine them as a type of horse race. We are not looking at the nominal numbers. We are only looking at them relative to a given index year, and for all but one 2008 is the index year. Everything is relative to the height of the crisis, which I believe is a valid way of seeing it.

If one looks at the first figure, it shows all the Vote if one adds in the voted expenditure for health and the HSE. That includes capital, non-pay, pay and pensions. If we look at the rest of the public sector, we can compare the evolution of both directly. One can see that from the early 1990s up to 2008 for much of the time health grew at a lower rate than the rest of the system. It caught up around 2003 and 2004 in terms of its funding levels, which it then matched exactly until the crisis hit in 2008. The emergency budget in 2009 pulled the wheels off the bus somewhat. One can see the jagged drop in health and HSE funding, which then jumped up again. One can see that the evolution of the health system relative to the rest of the public sector has been quite different. The spending on health, despite all of the negative press it has received, has been relatively high if one compares it with, for example, education which is rather different as a subject. Over the time there are two contrasting stories. One is of relative underfunding followed by a levelling out of funding, especially relative to the rest of the public service.

Look at health worker numbers in the second panel. Of course, they both fall relative to 2008 but health recovers faster but later. The rest of the public service begins recruiting. The public sector moratorium is removed in 2013 but it does not actually begin until 2014 for the health sector. In terms of the deficit of people, there is an approximate 10% deficit of people by the end of the crisis in the rest of the public service and it is a 12% or 13% deficit in the health service. If one looks at voted expenditure on pensions, we spent approximately €2 billion in 2016 on pensions.

An interesting point is that the health service accounts for approximately one third of all public sector staff - approximately 103,000 people are employed in the health service - but just a quarter of all public sector pension expenditure. It is clear from these statistics that health was relatively insulated from the decrease in capital expenditure over the period of the crisis. Again, that is to be expected. The next statistic to which I would like to refer explains the large increase. The voted expenditure on non-pay items is 25% higher in 2016 than it was in 2008. Non-pay expenditure in the health service is higher today than it was at the height of the crisis. We can get into the various reasons for that in response to members' questions. The level of voted expenditure on pay is relatively similar, as one might expect given that they are all public sector workers.

When we break down the staff levels, based on the HSE's annual report, we can see that 25% of those who work in the health service work in management, in administration or as support staff; 14% of health service staff work in health and social care; 34% of staff work in nursing; 9% of staff work in medical and dental roles and 18% of staff work in patient support. I would not attach massive amounts of explanatory power to these numbers, for several reasons. Support staff in this sense can be categorised in various ways. ICT personnel and pharmacists could be classified as support staff. It is not the case that all support staff are people who write letters. If one looks at what many of those who are counted as nursing staff do on a day-to-day basis, one will find that they are managing the system. There are other aspects of this in the patient support and medical and dental categories. This is a very imperfect measure, in terms of the breakdown of the levels, but it gives a sense of how the HSE sees itself and perhaps of how it has evolved.

We know for a fact that we will need to spend more on health in the coming years. Demographic pressures alone mean that current expenditure in the health service will have to be increased. This will have to happen independent of policy change or anything else. We know that demographic pressures will push this up. I would like to refer to an estimate of the dependency ratio that assumes there will be middling amounts of migration and fertility. It uses a pretty standard migration and fertility measure known as the M2-F2 assumption. It is estimated that the dependency ratio will increase by between 12% and 15%, which means that more older people will have to be cared for through their taxes by fewer younger people.

We have studied five different national models in the system. We have spoken to experts on these systems in several countries. I will break down the main findings from these models. There is no one single model that I would recommend for the Irish system, simply because we are so institutionally different from some of the other models. The simplest thing to do would be to take the Scottish model, find a logo and slap it onto the Irish system. I think that would be a mistake. The NHS system is very different from the Irish system, especially at the service delivery level. It is far less diffuse than our system, which has a very different public-private mix. While I would be reticent to recommend an individual model, I will go through the strengths and limitations of the various models.

The authorities in Scotland have spent a substantial amount of time engaging with the people they are going to be working with. They have been doing this for approximately ten years. One of the limitations of this approach is that Scotland organises its medical and other nursing training differently, relative to other areas. The Scottish authorities feel there is an over-reliance on quantitative data. I was completely shocked when this came to the surface in the expert interviews. I had a notion that this was a very technocratic exercise. Those who went out to talk to everybody said it was actually a matter of structured engagement where they got to look people in the eye and ask whether everything was going okay. As part of this engagement, the interviewees are asked whether they have enough people and whether they need more. This helps trust to build up over time. I think there is great value in thinking about that and about what kind of structure enables that.

As I said earlier, Wales has a legislative foundation for workforce planning. It is based around five-year planning cycles. There have been some strong quantitative studies in Australia. They are less focused on the qualitative elements. A far more collaborative approach is taken in New Zealand, where data quality is a huge challenge. A great deal of the data in New Zealand comes from surveys. Anyone who has ever decided not to fill out a survey on one's phone will appreciate that a representative sample is lost when surveys are used. This turns out to be a big issue.

The medical manpower model used in the Netherlands is objectively the simplest one. It is based on GP planning. If we want to adopt the simplest possible system, we should choose the Dutch system. The authorities in that country are able to pursue such an approach because they use the universal health insurance model. That may be a bad phrase to use around here - I am not sure. While this is very useful, it is important to note that an important aspect was lacking in the Netherlands model because there was no way to mix the different teams. It produced a headline number of doctors, nurses, phlebotomists and so forth. Under that model, one could not tell whether two doctors were needed for every three nurses and so forth. There was no skill mix.

Three things are needed if workforce planning is to be done properly. First, a well-resourced unit dedicated to information collection and analysis needs to be established. Second, this unit needs to be split between the Department of Health and the HSE - the strategic and operational sides of the health service - and needs to have links to the Departments of Justice and Equality, Education and Skills, Public Expenditure and Reform and Finance. Obviously, much of this is about financial control as much as anything else. Third, there needs to be a commitment to the generation of a minimum data set, rather than a maximum data set, to make this thing work. There is an international standard for what that is. There are data quality indicators that enable us to say that our estimates of nursing provision in Tipperary are as good as our estimates of nursing provision in Dublin, for example.

The qualitative side of this workforce planning process cannot be neglected. The establishment of forums and the dissemination of information about workforce planning should be a crucial task for the workforce planning unit that is set up. This should be done before any formal model is put in place. I will set out a rough governance structure to give the committee a sense of what this would look like. Of course we have a large private sector in our system. This includes GPs. Tusla, the voluntary bodies, the various section 30 groups and the various hospital groups are also involved. It is quite a diffuse system. It is quite difficult to get one's head around just how complex the Irish health system is. The workforce planning unit will take data from all the various bodies, work between the Department of Health and the HSE and contact external stakeholders like the ESRI and the OECD. I think something like this can deliver a workforce planning model in two to three years, which would be a pretty good return on the taxpayers' investment. I suggest we would get a significant level of engagement from the various health care professionals along the way. They would start telling us in a structured way what it is like to run these services and what they think needs to change. I do not think we would end up merely with people asking for more and more. I think most of the people who work in the health service are committed to making it better. Obviously, they are personally incentivised to make it better. I think we should create a system like this, to the extent that we can do so, in order to help our growing and ageing population.