The NHS plan envisages autonomy for primary care trusts and multiagency providers. But is the data available adequate for monitoring whether these agencies are operating according to the plan's principles for providing a public service?

Under the new arrangements, the NHS will collaborate more closely with local authority social services and the private sector. The Department of Health will take progressively less central control with increasing power and responsibility being devolved to service providers, PCTs and NHS and local authority commissioning bodies.

The key question is how this delegation of power and authority to multi-agency commissioners and providers, including PCTs, social services and the private sector, squares with commitment to deliver the core principles. The plan suggests that this will be achieved through the new arrangements for monitoring and reporting performance. But are data collection systems adequate for this purpose?

The radical statistics health group has identified significant gaps in the data required.

1It has looked at how data from official censuses and surveys, registration of births, deaths and cancer and notification of communicable diseases are used to assess the health of the population.

In the past, the private sector, local authorities and the NHS operated their services in isolation from each other. As a result, systems for data collection evolved separately, resulting in wide differences in the scope and range of data available. In general, less data is collected in the private sector and local authorities than in the NHS. There are also wide differences within the NHS, with much more data available for hospitals than for community services and primary care.

Data collected about NHS hospitals include aggregated data about bed availability and occupancy, staffing, waiting lists and times, while the hospital episode statistics include data about people receiving each episode of inpatient and day case care and the care they receive.

2Although the government had legal powers to collect any information it deemed necessary from the private sector, little was collected.

Data collection from private hospitals, clinics and nursing homes now lags far behind the NHS. No person-based data is collected about the characteristics of patients and the care they receive, and hospital episode statistics data are not submitted - even when patients are funded by the NHS.

With three exceptions - legal abortions, birth notifications and detentions under the Mental Health Act 1983 - no data is collected routinely about episodes of care provided. Nor is aggregated data collected routinely about the activities, such as operations, undertaken in private facilities. Data collection is limited to the capacity of the sector in terms of numbers of beds and their occupancy, whether patients are publicly funded, and staffing levels.

3This means that unless data collection improves, no information will be available about the characteristics of, and the type and quality of care provided to, the estimated 270,000 people expected to receive rehabilitative or intermediate care in nursing homes each year under the new plan.

4In local authorities data collection has also lagged behind the NHS. Until recently there were few attempts to get local authorities to use common definitions. Each local authority devised its own definition of who was eligible to become 'client'and how clients were categorised.

This included deciding who should be regarded as having a 'learning disability'being categorised as 'elderly mentally ill'.

4In April 2000, the 'referrals, assessments and packages of care'returns were introduced. These use standard definitions to collate information about assessments of individual adults. A different approach has been used for children.

Starting in 2000, data will be collected annually in a sample week about all children each local authority works with, including their age, gender, ethnic group, legal status and reason for receiving help. The DoH has introduced a series of outcome indicators for children in need.

These are crude, as no allowance has been made for other relevant factors such as the social or economic circumstances of the families in which children live.

In the Health Act 1999 and the NHS plan, primary care groups and trusts were given a central role as gatekeepers to all other forms of care. This means their systems should, in theory, hold information about all care funded by the NHS which each individual receives, irrespective of the type of provider.

Although funds have been allocated to upgrade GPs' computer systems to enhance their ability to provide care for individuals, the ability to derive statistical data from GP systems lags far behind.

Data about community health services is equally sparse and focuses on staff activities - for example, numbers of face-to-face contacts with community nursing and midwifery staff, rather than on their clients and the care they receive.

The need for functional client-based approach to data collection and the construction of indicators has been acknowledged for many years.

Several attempts have been made to define a community 'minimum dataset' in England, but this was dropped in the NHS information strategy.

In the 1980s and 1990s, care for long-stay patients was moved from the NHS to private nursing and residential homes. Although the majority of this care remained publicly funded, initially from social security and more recently from social services, little data is available to monitor the way these public funds were spent or the contributions individuals were required to make to their care.

5The implementation of the NHS plan will mean that increasing amounts of care of all types will be provided by local authorities and by the private sector. This means that care formerly provided free at the point of delivery, such as the personal care elements of rehabilitative and intermediate care provided in private nursing homes, is likely to be means-tested.

Social services departments and the NHS are now co-operating to develop common definitions, and record linkages and to share data, but there are many problems to overcome.

The new Care Standards Act 2000 gave the government powers to collect whatever information it deems necessary from the private sector, but there are many problems to be resolved before the information obtained becomes equivalent to that used to monitor NHS care. Private providers may resist providing fuller information to the government and the public if there is a risk that the data may show the outcomes and quality of their care in a poor light, giving the regulators a more powerful lever over them. The government also may be less than enthusiastic about increasing data collection, if the cost is passed on to the NHS through higher charges.

A positive step is that from April 2001, all general practices must have a system to monitor referrals, but the plan does not specify the range of providers to be included. It is essential that these systems should not be restricted to acute care and should include referrals to social services and the private sector.

The cornerstones of the NHS information strategy for England are electronic patient records, which will bring together information about all the care a person receives from a given provider, and the electronic health record, which will bring together information about all the health and social care a person receives.

The plan envisages that 75 per cent of hospitals and 50 per cent of primary and community trusts will have implemented electronic patient record systems by 2004 and that smart cards will be developed to make the information available to patients. These are encouraging developments, but the plan does not mention how electronic health records will be developed or how statistics will be derived from systems whose prime purpose is to support individual patient care.

This is a worrying omission, given the reliance the government plans to put on the relatively small numbers of 'high-level'indicators in its performance framework. The limited nature of the data available for these is acknowledged in the plan. A further problem is that PCGs or trusts are based on people registered with general practices and only loosely on geographical areas.

This means that problems arise in using primary care data in conjunction with census, birth and death statistics and other populationbased data.

The NHS plan places great emphasis on innovation and the devolution of control to the periphery - in effect to PCTs. We have a right to expect the government to make the necessary resources available to extend and strengthen data collection in the NHS and social services, while using the Care Standards Act to extend it to the private sector as a condition of registration.

Data is needed for sufficient transparency and leverage over services provided by other sectors to ensure that public money is spent effectively.

Patients should not be expected to pay for services that are of poorer quality than those currently provided free at the point of delivery by the NHS.