Primary Care

Impact of NHS Direct on general practice consultations during the winter of 1999-2000: analysis of routinely collected data

BMJ 2002; 325 doi: (Published 14 December 2002) Cite this as: BMJ 2002;325:1397
  1. Rachel S Chapman, primary care scientista,
  2. Gillian E Smith (GESmith{at}, regional epidemiologistb,
  3. Fiona Warburton, statisticianc,
  4. Richard T Mayon-White, consultant in communicable disease controld,
  5. Douglas M Fleming, directora
  1. a Royal College of General Practitioners, Birmingham Research Unit, Harborne, Birmingham B17 9DB
  2. b Public Health Laboratory Service Communicable Disease Surveillance Centre (West Midlands), Lincoln House, Birmingham Heartlands Hospital, Bordesley Green, Birmingham B9 5SS
  3. c Public Health Laboratory Service Statistics Unit, Colindale, London NW9 5EQ
  4. d Oxford City Primary Care Trust, Headington, Oxford OX3 7LG
  1. Correspondence to: G Smith
  • Accepted 4 October 2002

The impact of NHS Direct on other primary care services in the United Kingdom has been the subject of recent debate.1 A hospital bed crisis occurred in the winter of 1999-2000, but according to routine primary care surveillance systems the incidence of influenza-like illness did not reach epidemic proportions (as conventionally described). 2 3 Considerable medical and media interest was given to influenza activity during this “millennium” winter.4 There was speculation that there was a genuine influenza epidemic but that people were telephoning NHS Direct and not seeking help from their general practitioner, resulting in an artificially low incidence of influenza-like illness. At the time there was partial coverage of England and Wales by NHS Direct; we therefore used this “natural experiment” to assess whether the introduction of NHS Direct had any impact on episodes of influenza-like illness and other cases of respiratory infections seen by general practitioners.

Methods and results

We used general practices' telephone area codes to categorise those practices that participate in the Royal College of General Practitioners' weekly returns service on the basis of the degree of cover provided by NHS Direct during the winter of 1999-2000.5 The three groups were practices covered by NHS Direct since April 1999 (“full cover”), practices covered since November 1999 (“part cover”), and practices not covered until April 2000 or later (“no cover”). We compared weekly data on new episodes—as diagnosed by general practitioners—of influenza-like illness, acute bronchitis, and an aggregation of respiratory diseases with probable infectious aetiology during the winter of 1999-2000 with the three preceding winters. We also examined overall numbers of consultations per population. For each cover group and for each disease we calculated the mean weekly incidence per 100 000 population for week 48 to week 8, when respiratory illness is at its peak, and for the remaining winter weeks 35 to 47 and 9 to 20. As a comparative baseline we calculated equivalent data on incidence for the winters 1996-7, 1997-8, and 1998-9, when NHS Direct was not operating.

The age distribution was similar in the three cover groups and similar to that of the national population. In every winter and for all diseases examined, the incidence of the diseases was highest in the part cover group, with the single exception of influenza-like illness in 1999-2000, when the full cover group had the highest incidence (figure). The incidence in the full cover group was similar to or slightly higher than the no cover group for all winters examined. Incidences in the winter of 1999-2000 did not differ from the other winters. Numbers of consultations per population were similar in the full cover and no cover groups.


Mean weekly incidence of respiratory infections for four consecutive winters in general practices fully covered (since April 1999), partly covered (since November 1999), or not covered by NHS Direct


This examination of consultation data does not support the suggestion that an influenza epidemic occurred in the winter of 1999-2000 but was under-reported as a result of people contacting NHS Direct instead of visiting their general practitioner. The introduction of NHS Direct had no impact on the number of general practice consultations for influenza-like illness and other respiratory infections. A small decrease in the incidence of influenza-like illness in the part cover group was seen in 1999-2000, compared with the other groups, but this is unlikely to be due to the introduction of NHS Direct, as the full cover group did not show a similar change and the incidence of aggregated respiratory diseases with probable infectious aetiology remained high in the part cover group.

NHS Direct was not introduced to decrease or increase the number of general practice consultations but to make consultations more appropriate; we have not explored this issue. Reporting levels in the weekly returns service are based on approximately 8000 contacts per 100 000 population per week, so it is unlikely that NHS Direct—with a peak contact frequency of 250 per 100 000 population per week—would have any substantial impact.


We thank Duncan Cooper (Communicable Disease Surveillance Centre, West Midlands) for his helpful discussions on NHS Direct and Ken Cross for his involvement with the analysis.

Contributors: GES and RM-W had the original idea for the study. RC carried out the analysis of data and produced the first draft. All authors commented on drafts of the paper. GES and DF are the guarantors.


  • Competing interests RC and DF are both actively involved in the organisation of the Royal College of General Practitioners' weekly returns service.


View Abstract