How do we understand individual needs within groups?

Section 5

The groupings presented previously represent the primary organising logic. Within each of these groupings sit a set of cross-cutting themes or lenses to help us prioritise needs within the groups. These include age-related frailty, levels of economic well-being, behaviour, social connectedness, utilisation risk, presence of a carer, and a person’s own caring responsibilities. In addition to people’s clinical and social care needs, these lenses can have a significant impact on a person’s capacity and willingness to manage their condition as well as their reliance on statutory services. These lenses should therefore also be taken into account to help target individual services to best meet those needs (see exhibit 4.8).

Why not use utilisation risk or risk stratification, as the primary organising logic?

Utilisation risk, in the form of either the Combined Predictive Model or the PARR model, is widely used by providers to understand how needs are distributed across the population. This raises the question – Why not use risk specific stratification as the primary organising logic for our grouping?

There are three major reasons why not that are explained below:

  • Similar risk scores do not necessarily mean similar needs – two service users may have very high-risk scores, but vastly different needs. For example, someone who is very frail and elderly may have a risk score similar to someone who is middle aged and has diabetes and COPD, but the former may need fall prevention programmes and regular social-care visits, while the latter may need more intensive primary care and a biweekly visit to a specialist for managing his or her conditions.
  • Utilisation risk only measures risk of a non-elective hospital admission – while this is a useful proxy for needs, many needs are lost by only focusing on one aspect of care.
  • Risk stratification has a low ability to predict non-elective hospital admissions - most of the models have only very limited ability to positively predict the people who will be admitted to hospital. By understanding the population by their characteristics, fewer people will slip through the cracks.

The lenses mentioned above can be used directly within groups to improve care. For instance, risk stratification tools can be used within each group to understand where the magnitude of need is greatest, and therefore where within each group we should focus. Using the grouping around condition type and age combined with the risk stratification approach will give a more detailed and nuanced view of the population. The former tells us where in the population the types of needs are similar, while the latter tells us where the magnitude of needs are concentrated.

Exhibit 4.8 

 

A more detailed description of each of these lenses, and how they can affect the design of new models of care, can be found in Chapter 6: How do we innovate a new model of care working with users and carers? For example, by understanding which people within a given population are carers, providers could target carer respite programmes or carer peer support groups more effectively. It is important to understand all of these lenses when designing new models of care.


Case study/example

Essex County Council Isolation Index

Essex has a higher older population than the national average. Older people with care needs are projected to increase by 63% over the next 20 years to a potential 131,200 older people in need of social care and support. Essex County Council views the impact of loneliness as a complex and pressing issue for the authority.

In order to determine which communities are most affected by loneliness, Essex County Council has developed a unique ‘Isolation Index’, using commercial demographic data.

Using Mosaic UK (a consumer classification based on in-depth demographic data), 11 common factors – identified by research as drivers of isolation – were selected as variables, defined at household level and then combined to create an index. These included:

  • Single pensioners
  • Widowed
  • Retired
  • Unlikely to meet friends family regularly
  • Unlikely to interact with neighbours
  • Poor health
  • Suffering from depression
  • Suffering from poor mobility
  • Visually impaired
  • Hard of hearing
  • Struggling financially

Different scenarios were explored by weighting the relative importance of the common factors, for example one scenario focused on older people that are widowed and in poor health. Another scenario placed emphasis purely on contact with friends, relatives and neighbours. The weighting varied from 0.5 (ie. Half as likely) and 3 times as significant, but the result did not vary substantially from the uniform weighting applied to all the common factors.

The variables were then mapped at a Lower Super Output Area (LSOA) level to identify clusters of households that are potentially vulnerable to loneliness and isolation.

The initial exercise showed 55,000 households were statistically likely to be lonely, or at high risk of isolation. However, if the 37% of older people single occupancies in Essex is taken into account, over 80,000 people are statistically likely to be lonely.

The index has been designed as an adjustable tool and it can allow the council to allocate priority by resource availability as well as risk. For instance moving the statistical risk to 1.75 times more likely (than an standard household) to be lonely brings us closer to the national average of 10-14% of older people who self-identify as lonely all or most of the time. Essex County Council is now working with the voluntary sector more closely to design a behaviour change approach to encourage individuals and communities to build social networks and reduce their risk of loneliness.