Population healthcare: designing population-based systems

When health services get into difficulty, a common response is to change their structure, method of financing or both. However, these together comprise only one aspect of a service. A service consists of people and an organisation, and the organisation, in turn, has three different elements, as shown in Figure 1.

Figure 1.

The relationship between aspects of an organisation.

Complex challenges

The right structure is useful for simple tasks, but most health and healthcare issues are complex, where complexity is defined as ‘the dynamic state between order and chaos’. ¹ This means there is no linear relationship between input and output, but that their relationship is unpredictable; the carefully planned input may not lead to the desired output, and unintended consequences abound.

Some healthcare problems are managed in a relatively orderly way. The young man who fractures his anklebone playing football will usually reach the right service quickly, receive a standardised treatment and recover without complications. Some populations, for example older people with frailty, have conditions whose care lies closer to the opposite end of the spectrum from order to chaos. As mentioned, most healthcare problems are complex, but exactly how complex they are will depend on the given condition.

The solution to the problem of complexity is not more regulation, inspection or more detailed contracts, but rather the development of complex adaptive systems. These are defined by Sidhar Rihani as

Certain nonlinear systems … are commonly described as being Complex, because their behavior is defined to a large extent by local interactions between their components. When such systems are capable of evolution they are also known as Complex Adaptive Systems. ²

In healthcare, a system is a set of activities with a common set of objectives and an annual report. A population-based system focuses on a particular subgroup of the population with a common need, defined by:

• a symptom such as breathlessness, or

• a condition such as rheumatoid arthritis, or

• a common characteristic such as being frail and elderly or being a teenager.

These types of subgroups define the population in need. There are about one hundred common problems that health services have to tackle.

Focusing on common conditions

Rare conditions, as a category, are very common – but there are not many people with a particular rare condition in one population at any one time. In other words, a population served by health and social care will contain a large number of people with different rare conditions. Perhaps surprisingly, rare conditions are often much better managed than common conditions. Partly this is because they escape the attention of those who pay for or manage healthcare, which allows clinicians and patient representatives to work together to find solutions. For example, to cover the whole population of England, the clinicians interested in Type 2 neurofibromatosis have developed four networks which work well and which do not conform to any recognised boundary or structure in the National Health Services. These networks could be classified as comprising a ‘microsystem’, which is defined as an

Organizational unit built around the definition of repeatable core service competencies. Elements of a micro-system include (1) a core team of health care professionals, (2) a defined population of patients, (3) carefully designed work processes, and (4) an environment capable of linking information on all aspects of work and patient or population outcomes to support ongoing evaluation of performance. ³

The simple definition of a system given above can be further subdivided into open versus closed and hard versus soft systems.

Open systems and closed systems

It is customary to talk about open and closed systems. Closed systems are relatively unaffected by factors outside the system, whereas open systems are impacted by factors external to them. Healthcare is an open system liable to be influenced by many events outside the direct control of those who are trying to run it (Figure 2). Economic pressures, political interests and press stories all come to bear on people delivering health services.

Figure 2.

Healthcare as an open system.

Hard systems and soft systems

In a hard system, everyone is trying to achieve the same objective. In a soft system, however, all players are interested in the same problem, but some may be working in opposition to others. In a system to reduce teenage pregnancy, for example, one group may be keen to promote the pill and the morning-after pill, while another promotes abstinence and opposition to the pill and morning-after pill. In health and social care, it is common to find different stakeholders with different priorities and values, even while these are working towards a common goal of securing the best outcome for a population or subgroup. Healthcare meets the criteria for what is called a ‘wicked problem’.

Rittel and Webber (1973) originally developed this concept within the social planning literature to refer to problematic social situations where: (i) there is no obvious policy solution, (ii) many individuals and organizations are necessarily involved, (iii) there is disagreement between such stakeholders, (iv) where desired behaviour changes are part of the solution, and (v) there may be a call for co-production with citizens. ⁴

For this reason, they require what systems theorists call a ‘complex adaptive system’.

Complex adaptive systems

System scientists would classify healthcare as a complex adaptive system. The best example of a complex adaptive system is an ant colony. Ant colonies are marked by the fact that all the ants work for the good of the colony, not for the good of the individual or for their particular caste. This is analogous to the health service seeking to function to promote the good of a given population – i.e. its ‘colony’.

It is important to emphasise that the terminology in this area is by no means standardised. For example, there has been excellent work to create a population-based system of care for people with, or suspected of having, an acute stroke, in London. ⁵ There, they used the term ‘network’, a term we use simply to describe the set of organisations that delivers the system.

Further, the terms ‘managed network’ and ‘networking’ can be used to describe the actions of all the players in a system. Their applications are illustrated in the following quotes:

An activity by organisations which, and/or individuals who are themselves members of the network. The objective of such management is to exercise control, regulation, inducement, incentive or persuasion over some but not necessarily all other members of the network. ⁶

Networking is a broad concept referring to a form of organized transacting that offers an alternative to either markets or hierarchies. It refers to transactions across an organization’s boundaries that are recurrent and involve continuing relationships with a set of partners. The transactions are coordinated and controlled on a mutually agreed basis that is likely to require common protocols and systems, but do not necessarily require direct supervision by the organization’s own staff. ⁷

As always, it is essential to clarify the meaning of any term such as system or network before a workshop on how to set up a system is held.

The seven key stages in system design

There are seven key stages in system design, each of which will be covered in brief here:

setting up a design group;

clarifying the geographical population to be covered;

agreeing what is in scope and what is outside the scope;

defining the aim;

agreeing the objectives;

choosing valid and reliable performance criteria;

setting standards.

Setting up a design group

A system should be designed by a group of people made up of representatives of all the key agencies, including patient and carer organisations. The remit of a design group is:

To ensure that all the key organisations and stakeholders are represented in the systems design group.

To identify similar projects taking place in other parts of the country.

To complete the system design template, basing the work on the assumptions that there are no additional resources to be invested.

To identify related healthcare issues, for example those specified as not being in scope, and record the need for action on these other issues.

To manage the transition to become a network co-ordinating group when the design has been accepted.

All the key organisations and stakeholders in the specified care pathway, including patients groups, need to be involved in the systems design group. It is important to clarify with members of the systems design group that they are representing their constituencies but that they are not necessarily expected to commit their constituency or employing organisation. For example, someone who is representing the local hospital may not have the authority to commit that hospital’s board to a particular action and she, like all other members of the systems design group, needs to be reassured that significant decisions will be put back through the formal channels to the relevant organisations.

The design group needs to have one person to act as the co-ordinator of the group, and it may not matter too much which organisation that person comes from.

Clarifying the geographical coverage of the system

Within National Health Services care has been, is and always will be delivered locally, but the definition of ‘local’ requires clarification. ‘Local’ does not necessarily mean that care conforms to the geographic boundaries of any existing bureaucracy. In general, the optimum population size relates to two factors:

the incidence and prevalence of the condition, namely, how common it is, and

the need for expensive capital investment or rare super specialist skills best developed in teams.

The geographical shape and population size of services designed on the principle of population healthcare may bear little or no relationship to the jurisdictions that ensure their funding. Furthermore, these jurisdictions are ephemera because they change from time to time, whereas needs are relatively constant (with the occasional appearance of new needs, such as the emergence of HIV infection).

Agreeing what is in the scope and what is out of scope

Having picked a topic like headache or breathlessness, it may seem that the scope of the system has been chosen. However, it may take some time to reach agreement on what is inside and outside the scope. For example, in the development by Public Health England of a system for people with atrial fibrillation, many of whom have preventable strokes because of inadequate anticoagulation, the initial confused debate was clarified by agreeing that it was necessary to separate the more effective anticoagulation of people with known atrial fibrillation, which was deemed to be within the scope, from the task of finding people who were as yet undiagnosed, which was deemed outside of the scope. Two examples of defining the scope are shown in Box 1.

Box 1.

Examples of scope definitions.

Scope of a system for women with pelvic pain

The management of people with chronic pelvic pain (CPP) (defined as pain in the pelvic and lower abdominal region) for whom good quality generalist care has been ineffective, treatable causes have been excluded, and that persists beyond three months.

Scope of a system for teenagers with diabetes

The system is designed to meet the needs of children and young people with diabetes including the years of transition.

Articulating the aim

The aim is a single broad mission statement, something that could be put on a T-shirt and which is therefore not too long. How easy it is to measure whether the aim is met is not very important at this stage, as measurement will relate primarily to the objectives attached to the aim. Furthermore, the aim may take years to achieve, as is the case for the aim of the NHS Breast Screening Programme to reduce the risk of dying from breast cancer. Even then it might be attributed to interventions and developments other than screening.

Agreeing the objectives

The formulation of objectives, criteria and standards is interlinked, as demonstrated in Figure 3.

Figure 3.

Standards incorporated into an integrated system of care for cervical screening.

Clinicians and patient representatives are good at setting objectives, but they usually focus on clinical objectives and simply apply them directly to the population. During a RightCare workshop, for example, the following objectives for people with asthma were produced:

to prevent asthma;

to diagnose asthma quickly and accurately;

to slow the process of the disease by effective and safe treatment;

to help the individual afflicted adapt to the challenges;

to involve patients, both individually and collectively, in their care.

These are admirable, important objectives, but there are five other objectives which need to be included if the population is to be well served:

to make the best use of resources;

to mitigate inequity;

to promote and support research;

to support the development of staff;

to report annually to the population served.

The objectives given in this example form a good basis for objective-setting for any long-term condition. They could be given to a systems design group as a starting point, but it is often good to get the group to work together in pairs to set objectives in the first instance.

Choosing valid and reliable criteria

Criteria are those measures that are used as proxies to assess whether the objectives of a system are being met or not.

It is increasingly important to identify outcome criteria and to ensure that these are relevant to patients as well as to clinicians. In developing a system for people with arthritis of the knee, there are objective outcome criteria that can now be used, for example the Oxford Knee Score, which is an example of a PROM, namely, a Patient Reported Outcome Measure. However, from the patient’s point of view, there may be a different outcome relating to the problem that brought them to have their knee assessed, and therefore replaced, in the first place. To take an example, consider a person who came to the knee pain clinic because she was having difficulty bending to the extent that it interfered with her hobby of gardening. If she were to find out that she would not be able to kneel on that particular knee after a replacement, and therefore that her ability to garden would be even more affected than it had been before the operation, she would experience that as a bad outcome, even though the Oxford Knee Score showed an improved range of movement and less pain. There is therefore an increasing trend not only to measure objective outcomes but also to ascertain whether the patient’s presenting problem (i.e. the problem that first brought them to seek help) is:

better

no different

worse

Some would argue for a seven-point Likert scale from ‘much better’ to ‘much worse’, but three points are sufficient for most purposes.

Setting standards

A standard is a pre-set level of goodness, against which the measured criteria are compared. It is important that quality standards are set to allow performance to be compared to a reference. Often it is wise to set more than one level of a standard. For example:

a minimum acceptable standard below which no service line should fall, and

an achievable standard which is reasonable for all services to aspire to.

One way in which achievable standards can be set is simply by choosing a cut-off point between the top quartile and the rest of the services. If no data are available, then professional judgement can be used to set minimal acceptable and achievable standards.