Systematic review of cost effectiveness studies of telemedicine interventions

Dear Editor,

The paper by Witten and colleagues [1] reviewing the cost
effectiveness of telemedicine interventions was interesting and somewhat
provocative.

I am concerned that the authors confuse the quality of an evaluation
with levels of evidence and evaluation design (see the rapid response by
Alan Haycox to Paul McCrone). Quality should not be judged on levels of
evidence or if consequences were measured, but on how well these were
undertaken and reported. From their review, economic evaluations of
telemedicine tend to concentrate on costs and cost-minimisation ignoring
benefits. This is the conservative approach, that if benefits were
measured and valued, the cost-effectiveness of telemedicine would only
appear better.

In addition, Witten and colleagues have rated studies based on the
'presence of a clear hypothesis'. Although this is essential for
assessing efficacy, it is substantially less important for economic
evaluations. Hypotheses require the null to be accepted or rejected
through statistical tests; economic evaluations typically use point
estimates for costs and, without variance, cannot be subjected to
statistical testing. Hence, there is normally a battery of sensitivity
analyses undertaken around point estimates. What is required is a clear
statement of the aims, and as noted by Witten et al., most economic
evaluations do clearly state the aims.

Cost-effectiveness is about the additional cost for an additional
unit of benefit; it is a value judgement about whether it is worthwhile to
spend the additional costs to obtain the additional benefit (i.e. the
incremental cost-effectiveness ratio or ICER). A systematic review
judging the cost-effectiveness of interventions is not meaningful unless
we know what the ICER or incremental cost threshold is. At best, cost-
neutrality, as used in cost analyses and cost-minimisation analyses,
suggests that the outcomes from telemedicine and conventional healthcare
services can provide be obtained from telemedicine without additional
costs. It is a pity that these authors did not report their cost-
effectiveness threshold before claiming "Telemedicine is not cost-
effective" [2].

There were some easy questions not addressed in the review by Witten
and colleagues. I am disappointed that the review did not report
circumstances where telemedicine was and wasn't cost-effective. For
example, 'live consultations' are resource and cost intensive compared
with asynchronous consultations, but in some cases, such as the Highlands
and Islands Teledentistry project [3], might be cost-effective. Likewise,
telemedicine might be cost-effective for some disciplines such as
dermatology and radiology, and not others.

In their critique of the generalisability of results, the issue of
equity through access to healthcare services is overlooked. There are
many remote communities in the United Kingdom where there are substantial
barriers to access of usual healthcare services. For example, the rate of
referrals for restorative dental consultations from the Shetland Islands
(this requires a trip to Aberdeen) is one-tenth of the Orkney Islands rate
(where a consultant visits for one to two days per annum) [3].
Teledentistry and other telemedicine interventions, even if there are
additional costs, will help address the inequities that exist in access to
healthcare services.

1. Whitten PS, Mair FS, Haycox A, May CR, Williams TL, Hellmich S.
Systematic review of cost effectiveness studies of telemedicine
interventions. BMJ 2002; 324(7351):1434-7.

2. Whitten PS, Mair FS, Haycox A, May CR, Williams TL, Hellmich S.
Telemedicine is not cost effective. In "This week in the BMJ" BMJ 2002,
June 15; 324(7351).

3. Scuffham P, Steed M. An economic evaluation of the Highlands and
Islands teledentistry project J Telemed Telecare 2002; 8(3):165-77.