Improving effectiveness studies of immunization interventions for public health policy and practice

• Main Author: Hutchins, Sonja Suzzete
• Format: Dissertation
• Language: English
• Subjects: health sciences, public health
• ISBN: 0493611045; 9780493611044

Background. Evidence-based recommendations for public health practice should lead to more effective policies and programs. However, effective practice depends on the quality of the available evidence. Often, public health practitioners may not conduct effectiveness research because of limited resources. Selection of smaller than usual samples for effectiveness studies may improve the available scientific data, be affordable, and not sacrifice too much confidence in results.
Methods. To examine the feasibility of selecting smaller samples, an effectiveness study was examined of an immunization intervention with small (30 and 60) and large (100 and 200) control groups. Computerized simulations of 2000 hypothetical studies from a real study population were performed that compared small or large control groups with a large (200 participants) intervention group. Estimated population parameters included the intervention effect and the degree of certainty in the estimate. In addition, a cost-effectiveness study was conducted to identify the most efficient study design, and key public health informants were queried about information needed on smaller studies to increase their use.
Results. Average intervention effect (14%) from simulated studies was similar regardless of size of control group. Between 85%-99% of simulated studies found an intervention effect, varying by size of control group. Regardless of the size of the control group, the average effect was precise (0.6%-1.2%), average power (99%) was high, and the mean square error (0.7%-1.7%) was low. Moreover, a small control group of 30 was the most cost-effective design ($1618 per 1% likelihood of detecting an intervention effect). Six key public health informants identified cost to be a barrier to conducting effectiveness studies. If a decision is needed and no effectiveness data exist, informants seek expert opinion and, sometimes conduct small, pilot studies. Informants believe additional guidance is needed.
Conclusions. An example from an effectiveness study illustrates the situation-specific usefulness of a small control group. While findings may not be generalizable to other studies, results may be relevant to similarly designed studies. Because public health practitioners are interested in results of small studies for practice, additional guidance is needed to justify the selection of non-standard sample sizes.