Geographic information

Using GIS to get foundation trust membership right from the start

The Health and Social Care (Community Health and Standards) Act 2003, now superseded by the National Health Service Act 2006 (consolidating previous legislation) relates both to NHS Foundation Trusts (FTs) and the foundation trust regulator known as Monitor.

Foundation trusts are part of the bedrock of NHS reforms, being free from central Government control and able to make local plans and take local action on service improvement, inter-linked with financial independence that includes retaining surpluses and borrowing money.

Effective governance is a pre-requisite and this is transferred from the Department of Health and SHAs to a local governing body when a foundation trust is authorised. This governing body is elected from a membership which is intended to be representative of and have strong connections with the local community.

Defining the local community

To achieve excellence in governance of FTs, the local community must be defined in two dimensions: first there is the geographic dimension that deals with the physical limits of a “catchment” and the “constituencies within; and second the demographic dimension which looks at the representation of societal groups within the catchment. Both of these dimensions rely heavily on spatial analysis. Let's take each in turn.

Peterborough & Stamford Hospitals NHSFT was part of the first wave of ten FTs authorised on April 1st 2004. Shortly after authorisation, the trust wanted to look more closely at the catchment which it served, having adopted a high-level definition for authorisation and initial membership recruitment purposes.

A system of patient-to-public representation was adopted to establish a geographic catchment that truly reflected the localities that saw the trust as their own. Using inpatient and outpatient data, related to total population over the age of 16 in local authority wards revealed a set of near coterminous areas that would become the new catchment. Figure 1 (below) shows the outcome of the inpatient study.

Figure 1. Outcome of the inpatient study.

Not surprisingly, the further away from the hospital a ward is, the less the representation of patients travelling to the trust becomes. Hence the high proportion in Peterborough (red) and the low proportion in the outlying areas (blue). In consultation, the trust decided upon a 1:10 patient to public representation cut off. The outlying ward to the north-west was joined to the main body by selecting the wards in between as it was considered essential to have a single catchment without “islands”.

Jane Pigg, Head of Corporate Affairs/Company Secretary of Peterborough and Stamford Hospital NHS Foundation Trust says, “Our trust had a strong desire to ensure that our FT catchment was both fair to all the patients that use our services, and manageable in terms of the formation of constituencies and membership engagement. The transparent evidence-base that this approach to catchment definition provided satisfied our needs at the time of our authorisation and helped us to further develop the quantity and diversity of our members.”

Once a catchment has been defined, the constituencies within need to be set out. The design of constituencies is crucial as they will form the basis for the election of governors from the public membership. Constituencies are more a mixture of demographic and geographic study. Constituencies need to contain broadly the same number of citizens, but also need to be balanced and well distributed to ensure an even representation of the catchment in the governing body.

Turning to the demographic dimension, we need to first look at the appropriate definition of diversity in the context of FT authorisation. Surprisingly, both Monitor and the Foundation Trust Network say very little on this subject. Broad age groupings are listed, as are high-level ethnicity clusters (not in the detail available in the 2001 census) and social economic groups.

But in reality, both Monitor and FTN are silent on this important matter. Each FT has to work out their own approach to the detail of diversity, which is leading to inconsistency in reporting and standards. There is a great opportunity for Monitor to provide detailed guidance on this and establish a basis for national reporting and study on citizen participation in FT membership.

Moreover, amongst the consultation on amendments to Monitor's Compliance Framework dated 31 January 2008 it was said that: “membership plans should include growth in numbers of members but also measure and develop the effectiveness of the engagement between NHS Foundation Trusts and their members.”

Figure 2 (below) comprises the baseline demographic diversity of the catchment of Homerton University Hospitals NHSFT, another of the first wave of 10 FTs. Here we can see that of the 4,214 public members, there is a bias of females to males (56% v 42%) whilst the balance of females to males in the catchment as a whole is closer (53% v 47%).

Figure 2. Recruitment 2001 census baseline/ membership comparison.

Accordingly, females are slightly over-represented (index 107, where 100 would be near perfect representation), whilst males are slightly under-represented at an index of 89. From this table we can see at a glance that Homerton has 2.4 times the number of 'other mixed' ethnic classification members, and only 10% of the 'white and asian' classified members that it should have.

Pauline Brown, Director of Corporate Development at Homerton says “Membership diversity will only be achieved if the demographic segmentation of our membership parallels that of our catchment. This is also most certainly impossible to achieve and the Trust must strive to get as close as it can through targeted recruitment and engagement of our communities.”

Alternative demographic approaches

The comparison of a 2008 membership with 2001 census of population contains an obvious inaccuracy.

The 2001 Census of Population achieved a reliable return rate of 98% of resident citizens. It is this near completeness of response that has, to-date, underpinned the decennial cycle of census. But the period since 2001 has seen one of the most dramatic changes in societal diversity fuelled by immigration patterns, the rate of new home construction and changing work/life patterns.

When studying the 2001 census at a low level of geography (such as the output areas in which the data was reported which contain on average 140 households), now in 2008 the accuracy has to be drawn into question. These inaccuracies decrease with the higher levels of geography such as wards and local authorities.

An alternative to census data is lifestyle data such as Experian's Mosaic data and CACI's Acorn data. These differ from census data in that they are updated on an annual basis so household estimates are more reliable in these inter-censal years, but the application of a lifestyle type to a household and postcode is largely modelled from relatively small samples. So, on the one hand, census is a near complete survey but only updated every 10 years, whilst lifestyle data is largely modelled from a small primary survey but updated every year.

Monitor has alluded to the potential for lifestyle data to be used as well as, or instead of census data in reporting on diversity. This first appeared in the compliance framework document published in April 2007 but remains as a suggestion. Interestingly, the compliance framework also talks about “profiling techniques” that are intended to exploit the attributes about lifestyle data, so more of a shift towards lifestyles in the lead up to the next census is likely.

With the next census planned for April 2011 and a realistic expectation of 2012/13 before low level data is available, there is prospectively a 4-5 year window in which the 2001 census data will continue to erode in accuracy, and lifestyle data will become more the norm.

Having said that, lifestyle data is the only basis on which real targeting of membership recruitment, before and after authorisation, can be achieved.

At Essex Rivers Healthcare NHST, which is expecting its authorisation in May of 2008, considerable work has been done to achieve the best level of diversity in its membership in the lead-up to authorisation. This work is partly due to the Trust's commitment to excellence in governance predicated on excellence in membership, and partly due to Monitor's more stringent expectations of public membership quality as well as quantity as the waves of FTs progress.

With random mailing campaigns generating 1-2.5% return rates, the 5.1% recruitment rates achieved by Essex Rivers from a mailing of 10,000 letters specifically tailored to targets in known, previously under-represented lifestyles, was highly cost effective. Not only did this campaign add 510 new public members, it also overcame some of the most significant imbalances in the diversity of the membership.

Figure 3 shows the clustering and spatial distribution of the full postcodes containing the target 10,000 households by under-represented lifestyles within the catchment of Essex Rivers. Each colour represents an individual lifestyle and the grouping in defined areas should be noted.

Figure 3. Distribution of under-represented lifestyles in the trust catchment area

Paul Searle, Head of Communications at Essex Rivers Healthcare says, “Only by using the intelligence behind the lifestyle data and the ease with which target households can be identified could we achieve such a recruitment response, and redress an evident inequity in our membership diversity. A census-based action plan could not have achieved the same outcome.”

Figure 4 shows the before and after representation of one of the four target lifestyles. In the top map, the red areas within ERHT's catchment show where people who live a “small time business” lifestyle are under-represented in the overall catchment. In contrast, the bottom map shows how much of the red has turned to green, indicated a swing in these areas from under to over-representation, giving a more significant balance overall in the membership to the representation of this lifestyle.

Figure 4. Membership recruitment map for 'small time business' category.


Grey shows no representation
Red shows under-representation
Orange shows average representation
Green shows over-representation

Conclusions

Geographic and demographic information has a significant role to play in the formation of Foundation Trusts and their ongoing governance. Systems that utilise spatial databases for the management and development of FT membership provide a “real world” picture of the way people live their lives and the opportunities for their involvement in the running of FTs. GIS helps bring to life the issues of FT membership development covering the whole member life-cycle from recruitment, through engagement and retention.

Paul Smith, Managing Director of Active Solutions Europe Ltd.
Email: