GPs and purchasing in the NHS: the internal market and beyond By Bernard Dowling Ashgate 276 pages £39.95

GP fundholding was the most controversial component of the 1991 NHS reforms. To its supporters, it offered a flexible and patient-sensitive mechanism for purchasing secondary care and providing primary care. To its opponents, it imposed heavy transaction costs (through multiple contracts) and, most importantly, generated inequity between the patients of fundholding and nonfundholding GPs. Shorter waiting times for elective surgery experienced by fundholding patients were one of the most widely cited inequities.

Unfortunately, most of the evidence on this subject was anecdotal and partial. As one prominent researcher put it: most evidence came through 'poncing' (ie partial and non-comparative evaluation).

This criticism cannot be levelled at Bernard Dowling's meticulously conducted study.

Originally the subject of a PhD thesis, GPs and Purchasing in the NHS investigates variations in waiting times for elective surgery - and the reasons for them - between GP fundholders and health authority purchasers in West Sussex over the period 1992-93 to 1995-96.

Dowling found that patients of GP fundholders did indeed have shorter waiting times. Possible reasons for these discrepancies - such as more generous budgets for fundholders and differences in case-mix - are exhaustively explored and rejected.

Ultimately, the explanation he favours is that fundholders were able to get better deals for their patients by exerting more purchasing leverage over providers than HAs. This was possible because fundholders had an 'exit' option: that is, they were able to direct their patients to alternative providers if existing ones were unable to offer satisfactory terms. In fact, the threat of such a switch was often sufficient.

It was rarely used, and then only at the margins.

The criticism of excessive transaction costs imposed by GP fundholders is also investigated.

Certainly, multiple cost-per-case contracts did impose heavy costs on providers. But Dowling maintains that these could have been reduced substantially and did not necessitate the abolition of fundholding.

Overall, the study offers a strong evidence base in an area where ideology tended to dominate debate. It suggests that GP fundholders were, in many senses, better purchasers than HAs. But fundholding is now history. What lessons does the research hold for the future development of primary care goups and primary care trusts?

To the extent that GPs are still involved in purchasing, Dowling is able to offer 'cautious optimism'.

However, his research emphasises the advantages of small-scale flexible purchasers.

Large PCGs will find it difficult to maintain this relative advantage.

One way in which the benefits of small-scale purchasing could be sustained is through budgetary devolution within PCGs.

More fundamentally, though, there is bound to be a tension between an approach to purchasing that emphasises local innovation and flexibility (such as fundholding) - which inevitably benefits the patients with go-ahead GPs - and a policy which places high priority on national standards and the avoidance of inequalities, even if these are short-term and transitional.

The findings of this study clearly favour local flexibility, but it is not apparent that ministers currently share this view.

Ray Robinson Professor of health policy, London School of Economics.