Systematic review of cost effectiveness in telemedicine

Editor—Witten et al reviewed the cost effectiveness of telemedicine.¹ In their response on bmj.com they confuse the evaluation of quality with levels of evidence and study design.² Quality should not be judged on levels (strength) of evidence or whether health consequences were measured but on how well these were undertaken and reported.³

Additionally, studies were rated on the “presence of a clear hypothesis.” Although essential for assessing efficacy, this is less important for economic evaluations. Hypotheses require statistical testing to acceptance or rejection of the null hypothesis; economic evaluations typically use point estimates and, without variance, cannot be subjected to statistical testing. Hence, sensitivity analyses are required for testing the robustness of results. A clear statement of aims is required, and as noted by Witten et al, most economic evaluations do clearly state the aims.

Economic evaluations of telemedicine often concentrate on costs and cost minimisation, assuming health outcomes are at least as good as those from conventional services. For example, in teleradiology a reasonable assumption is that patient outcomes will be equivalent to or better (such as in terms of quicker results and avoiding travel) than conventional reporting by a radiologist. This is the conservative approach; if benefits were measured and valued the cost effectiveness of telemedicine might only appear better. Cost neutrality, used in cost analyses and cost minimisation analyses, suggests that telemedicine provides the same health outcomes without additional costs. Cost effectiveness is concerned with additional costs for additional units of benefit; it requires a value judgment about whether additional spending to obtain additional benefits is worth while.

I was disappointed that the authors did not report circumstances where telemedicine was and was not cost effective. For example, live consultations are resource and cost intensive compared with asynchronous consultations, but occasionally—for example, the highlands and islands teledentistry project⁴—might be cost effective. Likewise, telemedicine might be cost effective for some disciplines such as dermatology and radiology but not for others.

In the critique of the generalisability of results, equity issues of access to healthcare services were overlooked. Many communities in the United Kingdom have barriers in accessing common healthcare services. For example, the referral rate for restorative dental consultations from the Shetland Islands (requiring travel to Aberdeen) is a 10th of that of the Orkney Islands, where a consultant visits for 1-2 days each year.⁴ Teledentistry and other telemedicine interventions, even with additional costs, help address inequities in access to healthcare services.