Economics of antenatal prophylaxis
This paper briefly surveys the literature on the economics of antenatal anti-D prophylaxis. For the purposes of this paper in order to qualify for inclusion studies must explicitly report and combine information on the costs and the effectiveness of antenatal anti-D prophylaxis. Studies that comment on the cost effectiveness of antenatal prophylaxis without presenting this minimum level of information have been excluded.
There are two broad questions of economic relevance raised by antenatal anti-D prophylaxis. The first is whether or not such prophylaxis is worthwhile, that is, whether or not overall the advantages outweigh the disadvantages. The second concerns what form prophylaxis should take. Here, whether or not prophylaxis should take place is not considered, the question is how should it be organised if it is to be provided. These are closely related issues because answers to the former question may depend on the particular programme to be evaluated. The distinction between the two is not always clear. A comparison of a prophylaxis programme for different groups, such as primi-gravidae and other women, might be viewed as either being concerned with the cost effectiveness of different forms of prophylaxis or with whether it is worth-while providing prophylaxis to particular groups.
To answer the question of whether or not prophylaxis is worthwhile requires a comparison of the costs of the programme with the benefits it generates. The relevant costs are the opportunity costs or the benefits that must be foregone elsewhere if resources are to be devoted to antenatal prophylaxis. Identifying these foregone benefits is often difficult and consequently, as an operational alternative, it is standard practice to compare the cost effectiveness of different uses of resources in terms of their cost per unit of outcome, for example, cost per year of life gained or cost per quality-adjusted life-year (QALY) gained. The limitations of this alternative approach in the context of antenatal anti-D prophylaxis are discussed below.
Cost effectiveness information is generally most useful when comparing different means of achieving a given end (such as a reduction in the rate of alloimmunisation). In such situations it may be adequate to simply compare programmes in terms of the cost per alloimmunisation prevented. However, this type of information does not generally provide a clear answer to whether or not a programme is worthwhile. It can, however, offer some guidance on how the programme compares to other uses of health care resources, and assist decision making with respect to the design of programme (for example, choice of dose regimen).
The efficiency with which resources are used is not necessarily the only consideration when allocating resources to antenatal prophylaxis. For example, considerations of fairness might also influence decision making (for example, treating all women in a similar fashion rather than treating primigravidae differently). Studies of the efficiency of resource use are still relevant, however, in that they can identify the additional costs of treating all women in a similar fashion.
Ten studies which provide some economic evaluation of antenatal anti-D prophylaxis are summarised in Table 1. They range from studies where the economic element is minimal9 to others where it is substantial2,10. The question of whether or not the benefits of antenatal prophylaxis exceed the costs has not been directly addressed but several studies have compared the cost effectiveness of antenatal prophylaxis with that of other healthcare interventions and have drawn conclusions regarding the desirability of introducing antenatal prophylaxis. Of the ten studies, four evaluate antenatal prophylaxis for all Rh-negative women3,4,6,7 four consider primigravidae only1,2,5,9 and two studies compare programmes for primigravidae with programmes for all Rh-negative women8,10
Table 1. Economic evaluations of antenatal anti-D prophylaxis.
|1||1979-81 US $||Compared predicted outcomes with and without intervention||Antenatal for primigravidae||Cost per HDN case prevented|
|2||1985 UK £||Compared actual with predicted and without intervention||Antenatal for primigravidae||Cost per life saved. cost per life-year saved and cost per QALY|
|3||1986 Can $||Compared actual with predicted outcomes||Antenatal and postnatal for all women||Cost per immunisation prevcnted|
|4||1986(?) Can $||Treatment group compared with historical controls||Antenatal for all women||Cost per immunisation prcvented, cost per life saved and cdst per QALY|
|5||1985 Fr F||Randomised controlled trial||Antenatal for primigravidae||Cost per immunisation prevented|
|6||1981 US $||Compared actual with predicted outcomes||Antenatal and postnatal for all women||Anti-D IgG per immunisation prevented. cost per immunization prevented|
|7||1988-91 DM||Three studies from the literature that used historical controls||Antenatal for all women||Cost per case of HDN prevented|
|8||1983 Can $||Compared actual with predicted outcomes||Antenatal for primigravidae and for all women||Cost per irnmunisation prevented, cost per HDN case prevented. cost per life saved. cost pcr life-year saved and cost per QALY|
|9||1980 UK £||Treatment group compared with historical controls||Antenatal for primigravidae||Cost per immunisation prevented|
|10||1995 UK £||Two studies from the literature that used historical controls||Antenatal for primigravidae and for all women||Cost per immunisation prevcnted and cost per Rh loss prevcnted|
Cost of antenatal prophylaxis
All of the studies have considered the cost of anti-D IgG. However, only four of the ten studies included the future healthcare costs which could be avoided as a result of successful prophylaxis1,3,8,10 and just two of these studies discounted these costs to allow for their futurity8,10. The importance of such information depends on the purpose of the analysis. If the cost effectiveness of these programmes is to be compared with that of spending in other areas of healthcare, or if an assessment is to be made of whether or not antenatal prophylaxis is worthwhile, such information must be included. If the purpose of the analysis is simply to compare different antenatal prophylaxis programmes it will generally not be necessary to include this information.
The ten studies identified above, on the whole, provide little detailed information on the costs that they report. Most studies display a widespread reliance on fee schedules or healthcare prices rather than using sufficiently disaggregated information to enable analysis of individual resource use. This has two consequences: the costs are presented in terms of average values so that confidence intervals for the costs cannot be constructed; and it is difficult to estimate the economic as opposed to financial consequences of antenatal prophylaxis.
The estimation of the cost of anti-D IgG is not straightforward. This is because blood products exhibit an interesting characteristic which can be referred to as supply jointness. The cost of extending the supply of one particular blood product depends on which component of whole blood currently drives the system. For example, the cost of IgG dilutant will depend on the extent to which IgG is produced as a consequence of meeting a Factor VIII target. Moreover, the component that drives the system at any one time can be changed as a result of changes in demand, the availability of alternatives, etc. Another consequence of joint production is that there is no uniquely correct way of allocating the fixed costs of production across the different products that are jointly produced. As a result the cost of producing additional quantities of anti-D IgG is poorly specified and it is thus not surprising that many authors have simply used the current price of anti-D IgG as an indication of its costs. Two UK studies2,10 present more detail on the costs of producing additional anti-D IgG
Effectiveness of prophylaxis
The studies can be classified in terms of the source of evidence for estimates of the effectiveness of antental prophylaxis. Three of the studies4,5,9. qualified for inclusion in the companion clinical review to this paper11, where they were graded IIa, Ib and IIa, respectively. Two other studies7,10. base their assumptions regarding effectiveness on studies graded IIa in the clinical review. Ib represents evidence obtained from at least one randomised controlled trial and IIa represents evidence from at least one other type of well-designed controlled study without randomisation. The rest either assume a particular level of effectiveness or make reference to published estimates rather than formal studies.
A range of outcome measures have been reported including: cost per immunisation prevented; cost per haemolytie disease of the newborn (HDN) case prevented; cost per life saved or loss averted; cost per life-year saved; and cost per QALY saved. The first two are useful when comparing different antenatal prophylaxis programmes. If one antenatal prophylaxis programme is superior to another in terms of cost per immunisation prevented, it is likely to be superior in terms of all of the other measures. The other measures are more appropriate where antenatal prophylaxis is to be compared with other health care programmes. However, care must be taken with these measures because it is unclear whether or not a fetal loss averted should be regarded as being directly comparable to a child or to an adult life saved, in part because of the greater scope for replacing fetal losses. It could be argued that the costs per Rh loss averted as a result of antenatal prophylaxis should not be compared to alternative interventions measured in terms of cost per life year saved but only to other interventions preventing fetal loss.
Torrance and Zipursky8 calculated a gain in QALYs, assuming that all survivors of Rh HDN lead a normal life and that the gain in QALYs is achieved entirely through the elimination of mortality. If this is the case, the cost per QALY is the same as the cost per life year saved. A similar approach is taken by the other two studies which make estimates of the cost per QALY2,4 However, Torrance and Zipursky note that a number of papers have reported instances of handicap. In Scotland between 1987 and 1991, two infants were handicapped due to premature birth related to Rh HDN, and it might be assumed that this handicap would have been prevented had the mother not been alloimmunised10. However, another recent Scottish study12 reports no developmental differences between infants with severe Rh HDN and the population as a whole. Given the relatively small number of cases and the difficulties involved in determining an appropriate quality adjustment, it is not surprising that these studies have effectively assumed that the cost per life year saved and cost per QALY gained are identical.
All economic evaluations are based on numerous assumptions, partly because of the uncertainty concerning some of the parameters and because for findings to be generalisable they need to hold for a range of assumptions since circumstances vary in different settings. Consequently it is important that economic evaluations explore the sensitivity of their results to the assumptions being made. The cost effectiveness of antenatal anti-D prophylaxis is dependent on many factors including:
- • the cost of anti-D IgG;
- • the costs of caring for alloimmunised women;
- • the effectiveness of prophylaxis.
The data on these are of variable quality and local circumstances will often differ. It is thus an advantage if evaluations are repeated under a range of alternative assumptions so as to provide some indication of the robustness of any findings. However, only three of the ten studies present such a sensitivity analysis in addition to their base case findings1,2,10
What do these studies tell us?
There are Canadian3 and German7 studies that estimate that the cost of an antenatal prophylaxis programme is less than the additional cost of treating the cases of HDN which such a programme would prevent. In both of these studies the programme evaluated was for all Rh-negative women. A recent UK study also suggests that antenatal prophylaxis could be cost saving but only if it is restricted to primigravidae (and particular dose regimens)10. The general conclusion, however, is that antenatal prophylaxis will increase costs. In four of the studies2,4,6,8 the additional cost per unit of outcome (such as, immunisation prevented or life saved) is viewed as acceptable relative to the benefits such spending might generate in other areas of healthcare. One study1 concludes that ‘an antepartum Rh immune globulin program may be a relatively expensive way to reduce the incidence of Rh hemolytic disease of the newborn’ (p. 638). Of the remaining studies, two refrain from making any recommendations regarding the desirabilit of introducing antenatal anti-D prophylaxis9,10, and one suggests waiting because of concerns over the supply of adequate volumes of anti-D IgG5.
The two studies that compare programmes for primigravidae with programmes for all Rh-negative women both found that programmes restricted to primigravidae are most cost-effective.8,10 This is a consequence of primigravidae having a larger number of future pregnancies, and thus greater savings in future healthcare costs can be anticipated per alloimmunisation prevented.
Only one study has compared the cost effectiveness of different dose regimens10. It found that a I×1250 IU programme was more effective and cost less than a 2 × 500IU programe. Although the 2 ×1250 IU programme was more effective than the 1 ×12501U programme, its incremental cost effectiveness (or the cost per additional alloimmunisation prevented) was considerably poorer.
Two studies have considered policies of selective administration of anti-D IgG depending on partner's blood type1,8 If the Rh-type of the partner is not known, the additional costs involved outweigh any benefits from selective administration.
The literature on the economics of antenatal anti-D IgG prophylaxis is (with one exception) at least ten years old. This should not be taken as a sign that the economic issues are resolved. This brief review has highlighted a number of points:
- • that direct comparison of the studies is difficult;
- • the quality of the economic evaluations is variable;
- • the worth of antenatal prophylaxis has yet to be explicitly assessed.
First, direct comparison of the cost-effectiveness estimates reported in the literature is difficult. The studies use a range of currencies and of base years. They vary with respect to the interventions evaluated (all women versus primigravidae, different dose regimens). They also vary with respect to: the treatment of costs; the outcome measures used; and the sources for the estimated effectiveness of prophylaxis.
Second, the studies vary with respect to the quality of the economic evaluations. This is unsurprising in that some are clearly planned and presented as economic evaluations of antenatal prophylaxis while the economic content of others is fairly minimal. As a result it is important not to give the economic conclusions of the various studies equal weight. Moreover, it is generally the case that, no matter how much care is taken over the economic analysis, a study can only be as good as the clinical information upon which it is based. The clinical data in this area are not strong by current standards.
Third, although some authors have attempted to draw conclusions about the worth of antenatal anti-D prophylaxis, no study provides an entirely satisfactory basis for such conclusions. Such judgements should be based on an explicit comparison of the costs and benefits of antenatal prophylaxis. One potentially important benefit, that accruing to women and their partners, has yet to be assessed formally and integrated with information on changes in resource use. Furthermore, the cost of the substantially increased volumes of anti-D IgG required as a result of routine antenatal prophylaxis has been insuffciently investigated.
The Health Economics Research Unit is funded by the Chief Scientist Office of the Scottish Office Department of Health (SODoH). The opinions expressed in this paper are those of the author, and not necessarily the SODoH.