PJ Online | MM Features (Morbidity mapping; a technique to help link needs and services in primary care)

Home

logo

Medicines Management
Issue no 2, pp13-14
March/April 2002

Features

Morbidity mapping; a technique to help link needs and services in primary care

Ensuring medicines are correctly targeted in the NHS is increasingly important, both to tackle health inequalities and as a means to maximise scarce resources. Steve Chapman and Heath Heatlie describe a technique that they have developed to help, specifically, primary care organisations in England

Many people and organisations, including the Government, recognise that historically there have been differences in the levels of health care treatment across England. The NHS Plan stated that "health inequalities were compounded by a failure to match the provision of services with health needs". The plan went on: "The issue is not whether the NHS has to set priorities and make choices, the issue is how these choices are made.''

It is implicit that such choices should be made by the newly formed primary care organisations (PCOs) in an equitable and transparent manner within a framework of national guidance, and adhering to the principles of clinical governance.

Central guidance comes in the form of National Service Frameworks (NSFs) which aim to set national standards for prescribers to aim for and thus overcome the "postcode lottery of prescribing and care". In addition the National Institute for Clinical Excellence provides guidance on specific drugs, or groups of drugs, in particular new chemical entities.

National guidance is limited by the number of selected therapeutic areas it is possible to cover in the short to medium term. This leaves primary care managers and prescribing advisers with the responsibility for deciding which local health needs are a priority.

Key issues to be considered are what diseases represent a particular burden of morbidity in the locality, what interventions will best cure, alleviate, or slow the progress of those diseases, and how current practice compares with best practice.

In the case of prescribing, pragmatically, it also make sense to compare PACT data with those of neighbouring PCOs to check whether your PCO is an outlier (either prescribing much more or much less than your neighbours) in any particular therapeutic group.

This is of limited value by itself, because the volume or cost of prescribing has to be considered within the context of the morbidity patterns for your own and those neighbouring PCTs. These data come from disparate data-bases, and combining them for planning purposes can be laborious.

Also, presenting such data at planning meetings for health improvement programmes and service and financial frameworks requires a series of graphs and tables which, particularly in the context of a public meeting, may not engage stakeholders with a limited knowledge of therapeutics.

We have created a geographic health information system to link these data and primary care healthcare need to service provision, for all PCOs across England.

Objectives

The study aim was to use a geographic information system (GIS) to map the relationship between patient population health care need and primary healthc are provision.

Methods

We used a spatial framework to link primary care prescription records to measures of morbidity. We selected asthma for the pilot study because it has a significant burden of morbidity and prescribing expenditure and is not yet covered by a NSF. A GIS was used to derive a map for all PCOs in England. These maps were used as the framework to display and analyse the morbidity and prescribing data.

The data

Retail sales area prescribing data from International Medical Systems for prescribing of prophylactic treatment for asthma (inhaled corticosteroids) were aggregated into English PCO areas. PCO prescribing data were then standardised using the age, sex and temporary resident originated prescribing units (ASTRO-PU).

ASTRO-PUs are designed to weight individual practice populations for age, sex and temporary residents and are used by the NHS Executive in prescribing allocation methodology.

The best approximations routinely available for morbidity data are the hospital episode statistics data which are available for specific diagnostic areas. In this study we used asthma diagnostic codes (ICD-10 J45-J46). The data are indirectly age and sex standardised.

Analysis

Both the prescribing and morbidity data were categorised into quintiles, that is five categories with an equal number of cases in each category. These were then displayed as a map for the whole of England and for regions across England, thus providing an easily understandable comparator for all potential stakeholders at PCO meetings (below).

In order to show the relationship between the two databases, the PCO data were also plotted using a modified Boston matrix (below).

This shows each PCO as a point on a scatter graph, with the x-axis representing variation from mean prescribing and the y-axis representing variation from mean morbidit. This, therefore, highlights PCOs with relatively high patient need (morbidity) and relatively low service provision (prescribing).

It is important to note that although the maps and the matrix show apparent disparities in prescribing, they do not show cause and effect because the data are not linked at the individual patient level.

An interactive computerised model has now been developed so primary care managers and prescribing advisers can use this model and input their own PACT data to generate maps and analyses to target resources to meet unmet need.

Acknowledgements

The morbidity maps were developed with support of an unrestricted research grant from 3M Health Care. Copies of the interactive programme based on this work are available through the company.

Stephen R Chapman BSc PhD MRPhamS,Professor of Prescribing Studies, Dept Medicines Management, Keele.
Heath F Heatlie BSc MSc PhD Research Fellow, Dept of Medicines Management, Keele