When encountering public services, it can be tempting to make the analogy with a supply chain, where products make a journey from factory, to warehouse, to distributor, to retailer, to customer and as shown in earlier Chapters this can be helpful in health and social care. However, sometimes in human systems things are less straightforward and the health and social care system is arguably much more complex than any industrial process.
Unlike production, where materials mostly flow in one direction,
in health and social care people flow in multiple directions, miss out some stages and perhaps repeat others. The same service capacity might be used for different purposes; for example in Fig.
, an ‘intermediate care’ service which has a single capacity, might be used BOTH ‘upstream’, as an alternative to hospital admission, AND ‘downstream’ to facilitate hospital discharge of people who have recovered somewhat, no longer need to be in hospital, but are not quite ready to go home. In the same system, some people might be waiting in hospital to be discharged to an intermediate care setting at the same time as other people in an intermediate care setting now need hospital admission, perhaps because their condition has deteriorated. Alternatively, people might be admitted to hospital having been users of a home care package, losing that package on or shortly after admission, then, several weeks later, cannot be discharged from hospital until another, identical or more intensive, home care package can be procured. Such eventualities are not only possible, but common.
(same as Fig.
) High level view of integrated care model 8.1
Therefore, at any point in time, people are waiting to be transferred between services, when capacity allows, in various combinations:
(waiting in A and needing B)
(waiting in A and needing C)
(waiting in A and needing D)
(waiting in B and needing A)
(waiting in B and needing C)
Etc. (but some pairings are not possible or likely).
This degree of disaggregation creates a challenge to the transparency of a stock/flow diagram. Depending on the software used, the model can be constructed with each service being represented as a single stock. These stocks are further disaggregated elsewhere on the model screen, using a co-flow structure and, in the case of Stella Architect software, using its modular capabilities.
The disaggregated representation of each service will look something like the diagram in Fig.
Disaggregated view of one service—note that the stocks drawn with dotted lines are copies of stocks that appear elsewhere in the model diagram
Each service might have people currently waiting for a place, of whom some are not currently using a service, and others are using other services. To make this diagram clearer, we have drawn ‘sector frames’ round the people who are ‘waiting for A’ and those who are using ‘Service A’. Note that, in the diagram, those waiting but currently using other services are shown as ‘ghosted’ stocks (dotted lines), copied from elsewhere on the model. The stock called ‘using service A’ represents the number of people currently using service A and are appropriately placed. Service A is also being used by people who need, variously, services B, C, and E. And, in this example, nobody ever goes from A to D. The total number of people using service A is therefore the sum of the four stocks.
Note that another way of representing this degree of disaggregation more simply, would be to ‘array’ (or subscript) the ‘waiting in’ stocks.
Allocation of People Waiting to Capacity Restricted Services
The key question that arises, for each service, is ‘what is the policy for allocating spare places across the various waiting stocks?’ Various possibilities apply, all of which can be modelled. For example:
Spare places might simply be allocated in proportion to the number of people in each waiting stock
Places might be allocated based on a very strict order of priority: we allocate spare places first to people waiting here, then to people waiting there, then there, and so on; one problem with this approach is that if the whole system is under pressure, some people ‘waiting here, needing a place there’ might never be allocated a place, in which case, the modeller must also bear in mind ‘what happens to people who wait for a very long time?’ because their state of health/or need for care is likely to change
Places could be allocated according to some kind of weighting system (we admit relatively more people from here but always some from there), which gets round the problem of some people never moving
A variation might be to base that weighting on some kind of ‘floating goal’ derived from some other dynamic in the model, such as ‘the length of time for which people have been waiting to move from x to A’; for example, people waiting in B should wait no more than twice as long as people in C.
Obviously, the modeller should create an allocation policy that most closely matches the actual policies that are deployed (Sterman
, Chap. 15), assuming that these are known and can be codified. It is perfectly possible that what actually happens is the result of various informal decision-making heuristics that are brought to light by the modelling process itself and are poorly understood by key people across the whole system. 2000
A further important consideration with this kind of model is the question of ‘what happens to people who wait for a long time in any one state?’ Depending on where they are in the system, their condition might improve or deteriorate.
Where people are improving/recovering, it is common for people in hospital on reaching the status of ‘no longer needing to be in hospital but needing something else’ to ‘step down’ to some kind of intermediate care setting, the purpose of which might either be to provide care whilst a permanent home care package is organised, and/or to provide additional rehabilitation. In these cases, if intermediate care is NOT provided, they will remain in hospital for the duration of their recovery, then be discharged to a home care package when it is ready. Going back to Fig.
, people waiting to transfer from A to B might actually transfer from A to C if B is not available in time (detail not shown in the diagram). 8.15
The opposite effect might happen in the case where the existing level of support for someone waiting is less than is needed (or completely absent). People waiting at home for a home care package to be arranged might be admitted to hospital. People on an elective waiting list might be admitted to hospital as an emergency.
To summarise, it is quite common in these structures to have at least two outflows from a waiting stock, the outflow to the service that is desired, and the outflow to the destination that will result if the desired service is not provided in time.