We found a considerable shift in the prescribing of antihypertensive drugs towards the use of thiazides in our trial. In our economic analysis of the trial data we found that reduced drug expenditures due to increased use of thiazides did not outweigh the costs of the intervention. However, assuming that the effect is sustained for a second year, the picture changes and the intervention can be expected to lead to savings. We predict modest savings within 2 y if the intervention is implemented in a national program.
The key issue in determining the cost-saving potential of the intervention is whether the intervention effect will be sustained beyond a year, which was the duration of the trial. There are two good reasons to believe that this will be the case. Firstly, it is reasonable to believe that most patients who start using thiazides will continue to use these drugs rather than the more expensive alternatives, although many patients will be started on additional drugs regardless of the first drug given. This may translate into savings in subsequent years. Secondly, it is likely that the intervention will continue to influence physicians' choice of antihypertensive medication for new patients beyond the first year after the outreach visit took place. In the trial, the rate of thiazide prescribing to new patients was stable throughout the study period in the intervention group, with the effect showing no sign of waning over time.
Whether it is reasonable to assume that the intervention effect will be sustained beyond 2 y is less certain. Firstly, the pharmaceutical market is dynamic, with prices constantly changing and the regular introduction of new products. Secondly, countermeasures from pharmaceutical companies may be anticipated if health authorities launch national outreach programs aimed to influence physicians' prescribing.
There are few other economic analyses with which we can compare our results. Whether a cost of US$454 for changing a prescribing decision is a good or bad result is unclear. The intervention also targeted other behaviors, such as physicians' use of risk assessment tools, and had little or no impact on these. Thus, for these behaviors the intervention would be more costly but no more effective, i.e., it would represent an inefficient use of resources.
In light of these results, policy-makers need to consider alternative strategies if they wish to influence prescribing more cost-effectively. Recently, authorities in Norway made thiazides mandatory as first-choice medication if patients were to have drug costs reimbursed through the national health insurance system. The effectiveness of this strategy is under evaluation. Another alternative is establishing a maximum level of reimbursement for a group of drugs assumed to be therapeutically equivalent (reference pricing). This strategy has been thoroughly evaluated in British Columbia, Canada, for one class of antihypertensive drugs (angiotensin-converting enzyme inhibitors), and was found to lead to substantial savings on health budgets [
]. However, applying one reference price for all antihypertensive drugs may be more difficult to implement and has not, to our knowledge, been evaluated rigorously.
There is some debate about whether the choice of antihypertensive drug has an impact on health outcomes, which again may have economic implications. For some drug classes, e.g., alpha-blocking agents [
], there is sound evidence of inferiority compared to other classes, while for other drug classes, e.g., angiotensin receptor blockers, it is unclear whether the therapeutic effects are as good as for other classes [
]. It could be argued that an increased use of thiazides would lead to improved health outcomes [
], but the superiority of thiazides over other drugs is small in any case, and is not likely to have had much impact on the present analysis.
There are weaknesses with our study, mainly related to the many assumptions we had to make. However, almost all the sensitivity analyses we conducted, including our suggested best estimate, point in the same direction: if the intervention effect is sustained for a second year, modest savings can be expected.
We have assumed that the intervention had no impact on the volume of prescribing of antihypertensive medication. The trial data on prescribing volumes may be interpreted differently since the absolute difference between prescribing volumes in the experimental and control groups differed between the baseline period (one year before the outreach visit) and the study period (one year after the outreach visit) (
). However, the ratios of prescribing volumes during the study period versus during the baseline period were similar for the two groups, and the observed difference may well be a chance finding. It is also highly unlikely that taking this finding into account would change the conclusions.
This study is one of very few comprehensive economic analyses of quality improvement strategies that include educational outreach visits. The lack of economic evaluations has been pointed out by others, and our findings illustrate that such analyses are important for well-informed decision-making.
The findings from our trial were similar to previous studies of outreach visits aimed at influencing prescribing patterns. Despite savings due to increased use of less expensive drugs, the intervention resulted in a net cost within the 1 y that the trial lasted. However, it is likely that the intervention would lead to modest savings over a 2-y period.
Publicly funded educational outreach programs should perhaps be targeted primarily at improving quality of care rather than costs of health care, given the cost of outreach visits, competition from drug companies, and the availability of alternative strategies for controlling costs (such as reference pricing, pre-authorization, and restrictions on reimbursement).