Intended for healthcare professionals


Initiatives to improve childhood immunisation uptake: a randomised controlled trial

BMJ 1998; 316 doi: (Published 23 May 1998) Cite this as: BMJ 1998;316:1570
  1. Maria Z Morgan, research officer,
  2. Meirion R Evans, consultant in communicable disease control (mre{at}
  1. Public Health Directorate, Bro Taf Health Authority, Temple of Peace and Health, Cardiff CF1 3NW
  1. Correspondence to: Dr Evans
  • Accepted 6 May 1998

Levels of childhood immunisation are high in the United Kingdom but are proving difficult to maintain. Several initiatives to improve uptake have been described, including sending written information to parents,1 specialist immunisation clinics,2 and prompts to health visitors and general practitioners.3 However, none of these interventions has been the subject of a randomised controlled trial. We tested the effectiveness of two such interventions.

Subjects, methods, and results

The child health system, which maintains computerised data on immunisation status of all children, was used as the sampling frame. The study population comprised children resident in the former county of South Glamorgan who were (a) born between 1 April and 30 September 1995 and scheduled to complete the primary course of diphtheria, pertussis, tetanus, polio, and Haemophilus influenzae type b immunisation or (b) born between 1 April and 30 September 1994 and scheduled to receive measles, mumps, and rubella immunisation. Children were included in the trial if they had not completed their primary course by 9 months of age or their measles, mumps, and rubella immunisation by 21 months of age.

Each week between 1 January and 30 June 1996 we received a computer generated list of children eligible for inclusion in the study and randomised each child using computer generated random numbers to one of two interventions or a control group. Intervention A comprised a non-directive telephone call to the child's health visitor to confirm the child's personal details and immunisation status. The health visitor was not informed of the trial and, although follow up of the child was anticipated, it was not specifically requested. Intervention B comprised a single mailed reminder to the child's parents together with a questionnaire about details of immunisation status and reasons for non-immunisation, and a reply paid envelope. Parents were not informed of the trial.

Study end points were completion of (a) primary immunisation by the first birthday or (b) measles, mumps, and rubella immunisation by the second birthday. We performed statistical analysis on an intention to treat basis, using the χ2 test with Yates's correction for baseline comparisons, and calculated 95% confidence intervals for the difference in proportions.

Table 1

Baseline characteristics and immunisation uptake in intervention and control groups. Values are numbers (percentages) of subjects unless stated otherwise

View this table:

In total, 153 children (76 primary course and 77 measles, mumps, and rubella immunisation) were randomised to intervention A, 159 children (82 primary course and 77 measles, mumps, and rubella immunisation) to intervention B, and 139 children (74 primary course and 65 measles, mumps, and rubella immunisation) to the control group. The study had a power of 80% to show a 15% difference between each intervention and the control group at 5% two sided significance. Distribution of baseline characteristics in the three groups was similar. There was no significant difference between either intervention group and the control group in the proportion completing the primary course or measles, mumps, and rubella immunisation (see table). Nor was there a significant difference in study end point, when both interventions combined were compared with the control group. Subgroup analysis by maternal age and parity showed a substantial but non-significant effect of intervention in promoting completion of primary immunisation in firstborn children (56%, 10/18) compared with firstborn controls (25%, 3/12), and in children of young mothers aged ≤30 years (31%, 27/86) compared with controls (13%, 5/38). There was no effect on uptake of measles, mumps, and rubella immunisation.


Randomised controlled trials provide the best evidence for effectiveness of interventions. However, we found only one other trial of an intervention to promote childhood immunisation. This was carried out in preschool children in the United States and found that a computer generated telephone reminder resulted in a significant but modest improvement of 12% in immunisation uptake in the intervention group, after excluding the 20% of households with no telephone.4

Neither intervention we studied improved immunisation uptake. The results suggest that district-wide initiatives directed at individual families are unlikely to be worth while, although there may be some benefit from targeting young or primiparous mothers. There is evidence that initiatives by primary healthcare teams such as opportunistic immunisation of children attending the surgery and domiciliary immunisation by nurses can improve uptake,5 although these approaches would benefit from more formal evaluation. More use should be made of randomised controlled trials to evaluate interventions to promote uptake of preventive services in primary care.


We thank the child health support team at the Welsh Health Common Services Authority for providing immunisation data; Mrs Margaret Morgan, Cardiff Community Healthcare NHS Trust, for her invaluable assistance; and Dr Frank Dunstan, Department of Medical Statistics and Computing, University of Wales College of Medicine, for statistical advice.

Contributors: MZM helped design the protocol, collected and analysed the data, and helped write the article. MRE conceived, designed, and supervised the study; wrote the article; and is the guarantor.

Funding: None.

Conflict of interest: None.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
View Abstract