Intended for healthcare professionals

General Practice

Effectiveness of influenza vaccination policy at targeting patients at high risk of complications during winter 1994-5: cross sectional survey

BMJ 1997; 315 doi: (Published 25 October 1997) Cite this as: BMJ 1997;315:1069
  1. John Watkins, director of primary health care and consultant in public health medicine (101341.271{at}
  1. a Gwent Health Authority, Mamhilad, Pontypool, Gwent NP4 OYP, Wales
  • Accepted 25 April 1997


Each year the chief medical officer writes to general practitioners and other health professionals reminding them of the need to identify and vaccinate patients at risk of the complications of influenza—that is people who have chronic heart, chest, or kidney disease; people who have diabetes; people who are immunocompromised owing to treatment or disease; and people living in residential accommodation. Routine immunisation of elderly people is not recommended. Current data on the efficacy of influenza vaccine indicates that up to 70% of clinical cases could be prevented,1 2 an important finding as in 1989, 26 000 people, mostly elderly, or those recommended for vaccination, died in the United Kingdom from influenza and its complications.3 That year there was a good antigenic match between the epidemic strain and the one used in the vaccine, yet only one third to one half of all patients who would have benefited from vaccination received it.4 I investigated the implementation of current vaccine policy.

Subjects, methods, and results

In September 1994, 64 general practices in the county of Gwent, with a registered population of 291 908, took part in a study that entailed data collection from patients at the time of vaccination. Patients were asked their age, whether they suffered with any of the conditions for which influenza vaccination is recommended, and the method by which they came to receive vaccination. A numerical coding system was used to separate out each chronic disease and the method used to contact patients. Only practices that were computer linked to the health authority patient register were used, and this provided patient denominator data. Practices for which the authority held denominator data on chronic diseases were used to calculate uptake rates of vaccine in at risk groups. Statistical analysis was carried out with SPSS 6.0 for Windows.

For the 28 433 doses of vaccine given in the 64 practices, information was submitted on 21 001 patients (74%). Overall, the vaccine uptake rate was 97.4 doses/1000 patients (1), though individual practices showed wide variation (range 25/1000 to 275/1000). Uptake rates in specific at risk groups were calculated for the practices that had recorded all of their immunisations. Analysis showed that under half of those patients identified as high risk and recommended for vaccination received it: only 63% of patients with heart disease, 39% with diabetes, 41% with asthma, and only one in three of those over 75. One quarter of all doses were given to patients at low risk. The 1 shows that advice from general practitioners accounted for 40% of all those being vaccinated, most of the remainder resulting from self referral by patients on an annual basis or on advice from the practice nurse. Other health professionals, particularly hospital consultants, played an insignificant part in vaccine promotion. Under 4% of patients were recruited by proactive methods such as telephone, letter, or a message on repeat prescriptions; 80% were recruited opportunistically. Poster campaigns had little influence in targeting those who would most benefit. There was no significant difference in uptake rates between practices according to whether they were training practices or fundholders, had more than two partners, or occupied cost-rent premises. There was also no relation with list size, though those practices with the highest vaccination rates had the highest uptake in those who would most benefit.

Doses of influenza vaccine given to patients at high and low risk, showing relation with methods by which patients were contacted. Numbers in parentheses are percentages of total doses arising from that method of contact, unless stated otherwise

View this table:


The methods used in this study tend to overestimate the uptake of influenza vaccine in patients with heart and respiratory disease because of denominator deficiencies—for example, in calculating the uptake rate in patients with respiratory disease, the denominator population was calculated using the number of patients with known asthma and did not include other chest complaints, which would lead towards an overestimation. This study showed that personal advice from a general practitioner or practice nurse during the vaccination period was the greatest stimulus to vaccine uptake. There was little evidence of practices using vaccination registers to plan their vaccination programmes, and other health workers, though targeting risk groups correctly, did so too infrequently to make an impact.

Influenza is an important disease of major public health concern, with an effective vaccine. The United Kingdom currently spends over £30m on influenza vaccination (derived from the drug tariff cost per dose for six million doses of vaccine given in the winter of the study), yet this merely covers the cost of vaccine and fails to deal with organisational issues. I have shown that the present system, which relies on the idiosyncratic behaviour of individuals with minimal central guidance, no mechanisms to ensure effective targeting of vulnerable groups, and no link between renumeration and performance, results in less than half of those who require vaccination receiving it, while half is given to people at low risk. This approach falls short of delivering an evidence based public health policy aimed at reducing the impact of one of the world's major killer diseases, as shown during the winter of 1989-90.


I thank Amer Jamil and Kerry Ross Jones for help with data entry, and Professor Peter Ellwood for his help and advice during the study.

Funding: The study was made possible by a research grant from the Association for Influenza Monitoring and Surveillance.

Conflict of interest: None.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
View Abstract