CHDGP - The National Project

This Project now continues as PRIMIS from the new web site

This site will remain online but will no longer be updated

CHDGP - Plans for the Pilot Stage

THE NATIONAL PROJECT

Aims of the Project

In response to the needs outlined above, the aims of the national project as a whole are:

	To facilitate and support the establishment of local schemes within the NHS to collect, analyse and make practical use of data from General Practice systems, with suitable safeguards for patient and practice confidentiality. The project will aim to facilitate sharing of experience and lessons learned, and look for ways to avoid duplication of effort and cost in setting up new data collection schemes.
	To ensure that data collected from participating practices is accurate and comparable, such that information derived from it is credible and useable. This involves providing guidelines for data recording, validation, extraction and analysis, and training and assistance with the application of these guidelines. Participating practices will not be expected to record anything they do not consider useful to support patient care. Project documentation will be freely available on request within the NHS.
	In a pilot stage to work with at least 10 local schemes, organised on a health authority area basis, with each scheme having at least 10 participating practices. Participation is voluntary. Working with these pilots, the project will seek to demonstrate a variety of practical uses for the information derived, and explore the cost-effectiveness of the process.
	To set up and operate a comparative analysis service for the duration of the pilot stage, taking data from and producing reports for all practices able to provide data of acceptable quality.
	To organise an evaluation of these pilot activities after one year of operation to assess the feasibility and desirability of wider adoption of the methods employed, and to assess the benefits that have or could accrue to participants and the wider NHS.
	As budgets and resources permit, to establish links and to work with organised data collection schemes, and also groups of practices and individual practices, in addition to the official pilots. In particular, to work with new types of primary care organisations (eg Locality commissioning groups, Total Purchasers, PCAPs) to explore the value of the process and the information derived.

The project is funded by the NHS Executive, to the end of the Pilot Stage.

Expected benefits

The potential benefits of the project may be summarised as:

	better informed decision-making - by GP practices and groups of practices, and health authorities - from use of the information that will be derived from the data collected.
	avoiding some of the costs involved in setting up stand-alone data collection schemes; and
	improved effectiveness and efficiency from more comprehensive and consistent use of computer systems by individual practices, which will be encouraged and assisted by participation in the project, or use of its guidance and training materials.

It may also be expected that a national initiative will enable GP data collection schemes to be established at an earlier date and with less effort than would otherwise be the case. This will be particularly welcome given the range of other information and IT related initiatives that staff will need to deal with over the next few years.

It may also be argued that schemes are also likely to be more effective and to produce better quality results if they are undertaken within a national framework, developed from best practice and expert advice.

Work Completed

Early Stages

A study conducted in 1993 examined the feasibility and desirability of some form of national initiative to increase the use of health data from GP systems. It has set the agenda for subsequent work, but its recommendations were not taken forward immediately because of the need at the time to assess the potential of the General Practice Research Database (GPRD), which was unexpectedly acquired by the Department of Health in November 1993. Following a review of evaluation studies of the GPRD conducted in 1994, in early 1995 it was decided to resurrect the national project work in the form of a Planning Stage, which was completed in May 1996, at which time it was decided to proceed with various development activities (termed the Development Stage), leading to a national pilot.

The Development Stage

The Development Stage commenced in September 1996 and will be completed in December 1997.

The main activities during this stage have been:

	To prepare a set of guidelines and policies for the pilot stage covering data recording within practices, data quality, data validation, methods of data extraction, confidentiality, and data analysis and interpretation. This work has built upon the outputs and activities of earlier projects.
	To prepare training materials for the above for use by project staff, scheme facilitators and practice staff.
	To recruit local pilot schemes to work within the framework defined by the national project.
	To set up a small training and support team to provide practical assistance to pilot schemes, and others within the NHS who express an interest.
	To work with GP system suppliers to make suitable data entry aids and query and data extraction tools available for their systems. In particular, to encourage and support the wider availability of MIQUEST, a flexible query and extraction methodology recommended by the project.
	To specify and establish a comparative analysis service, which during the Pilot Stage will provide pilot schemes with tabulations and reports that compare data collected by local schemes from their participating practices.

CHDGP Guidelines

The Guidelines document is the main reference publication for pilot participants, and others interested in taking advantage of the work of the project.

The latest version of the Guidelines is a refinement of the original prepared by consultants with advice from an invited working group. More significant revisions and improvements are expected in Version 2, which will draw upon experience gained to date in pilot schemes, and is due to be ready in July 1998.

The Guidelines document is available from the Training and Support Team (see Contacts section). Its constituent sections are summarised below.

Data Recording

These sections includes suggested good practice for data recording, techniques for checking data consistency and accuracy, and avoidance of common faults. The guidance includes direct entry at consultation, indirect entry after the event from notes, data capture by other PHCT members, the use of templates and protocols, and an overview of Read codes.

More specific guidance is offered on recording items within the Core Data Set, for which lists of recommended codes are provided.

Core Data Set

The purpose of the Core Data Set is to define a minimum set of conditions and types of data, sufficient to produce a useful set of comparative analyses, and to be able to assess data quality within participating practices. The Core Data Set provides a focus for baseline data quality audits and follow-up adjustments to practice procedures, and for initial data collection and feedback. The attainment of a consistently high standard of recording of the Core Data Set is seen as a first stage target for practices.

The list of conditions and data items in the Core Data Set is likely to evolve over time. It is not intended to limit more comprehensive data recording within practices, or the collection and analysis of data on other conditions.

The current Core Data set includes

Morbidities (prevalence) and presence of complications or sequelae, for

heart disease

cerebrovascular disease

diabetes

hypertension

asthma

severe mental illness

Risk factors relating to these and other diseases:

blood pressure

body mass index

smoking

alcohol consumption

Deaths, including immediate and significant contributory cause

and

Medication (prescriptions).

Data Extraction and Analysis

The objective is to obtain consistent, comparable data from all practices involved, and to ensure that confidentiality safeguards are maintained.

MIQUEST is a means for remote enquirers to express queries and extract data from different types of GP systems using a common query language. It incorporates safeguards against unauthorised and inappropriate access to patient data. It can also be used by GPs and practice staff to enquire on data in their own systems. Currently MIQUEST is available for EMIS and Meditel System 5; development of the necessary software is underway for several other systems. The use of MIQUEST is supported by the CHDGP Training and Support team, and detailed training materials are available for new users.

Other query and extraction tools and methods may be acceptable if MIQUEST is not available for a particular system, and can produce similar results. The Training and Support team can advise on this.

Analysis of data received from local groups of practices is carried out using locally developed systems, usually based on PC tools. To assist local schemes in specifying and developing the necessary routines, the Training and Support team will provide samples of commonly used analyses, and the reporting software used by the Comparative Analysis Service will also be available on request.

The Guidelines include advice on calculating incidence and prevalence, on grouping Read codes, avoiding double counting, processing linked health-event-datasets, and interpreting results.

Confidentiality and Scheme Management

Control over what data may be extracted, who has access to it, how it may be analysed and reported, and the uses to which the analyses can be put, are all of key importance. The guidelines for CHDGP on privacy and appropriate uses of GP data build upon national policies and guidance by DoH and professional bodies. The Guidelines also specify key aspects of management arrangements for local schemes and the Comparative Analysis Service, issues to consider when setting up a scheme, the types of data that may be collected, the uses to which it may be put, and the confidentiality safeguards to be employed.

Project guidelines require that all patient data extracted for use outside the practice must be anonymised. Practices will not be identified in any published reports, and if necessary results will be aggregated to prevent identification. GPs will have control over what data is collected, what results are produced from it, and the uses to which results are put.

It may be expected that local schemes will collect and store for analysis two types of data:

on a regular and routine basis, aggregates derived from calculations performed by software run against databases of patient records held in GP systems. These will normally be analysed by age and sex of patients, but no data on individual patients will leave the practice;

on an ad hoc basis, the selection and creation of health event-datasets capable of being linked via a patient number. (The patient numbers are allocated by and decodeable only by the practice.) This data can be analysed in various ways to answer audit questions or to support reviews of specific services. Only data needed for analysis will be collected, and this will exclude ‘strong identifiers’ ie key fields that might readily be used to identify individual patients.

For clarity, the following table shows what data items may not be collected, and what is considered acceptable and will normally be collected.

	Not permitted	Permitted
Any part of patient name	6
Any part of text address	6
Full postcode	6
Postcode sector *(eg GU21)		4
Full date of birth	6
Year of birth		4
NHS Number	6
Patient identifier generated by practice system		4

* However, where Health Authorities are working with groups of practices on reviews of health needs and service provided at a locality or small area level, by specific agreement with the practices concerned, the full post code might be collected, since otherwise the types of analyses required could not be performed. In these circumstances, special care would need to be taken over aggregation and reporting of results, to take account of the increased risk of patient identification.

It must be emphasised that, although no data collected will in itself be identifiable to individual patients, it will still be necessary to avoid ill-considered aggregation of data, which may threaten the confidentiality of the individual. This is especially important with rare conditions involving small numbers of patients, or conditions which are associated with recognisable characteristics. Organisers of and participants in local schemes must be alert to these potential problems and seek to prevent them.

With the management arrangements already in place within existing schemes, the data collected, the types of analyses and outputs that are permitted, and the uses made of them, are under the control of local management boards of participating GPs and public health doctors in the Health Authority concerned. Because all data extractions must be scrutinised and explicitly approved, individual GPs are in a position to monitor requests for data, and query or refuse to process those which may threaten patient confidentiality, or where they consider the data are liable to be used inappropriately or to their disadvantage.

[on to next section]