Everything you do as a health professional has a cost: cost effectiveness analyses

Developing countries, more than any others, need to get their money's worth when it comes to investing in healthcare interventions. Martin Dawes explains why learning about cost effectiveness analyses is important—because every intervention has a cost

This month's paper is Adam T, Lim SS, Mehta S, Bhutta ZA, Fogstad H, Mathai M, et al. Achieving the millennium development goals for health: cost effectiveness analysis of strategies for maternal and neonatal health in developing countries. BMJ 2005;331:1107. You can read it by clicking on the link.

“Cost effectiveness analysis of strategies for maternal and neonatal health in developing countries.” Or “This will help developing countries choose effective interventions to reduce maternal and neonatal deaths.” Or “Everything you do as a health professional has a cost.”

The sentence that usually explains the reason for undertaking the study is found in the first sentence in the introduction. Often this will be a dramatic phrase to catch the eye of the reader, and this article is no exception. “Each year more than half a million women die during pregnancy or childbirth and more than 4 million babies die in the first 28 days of life.” They follow this with the disturbing fact that there is a 10-fold difference in rates of neonatal death between rich and poor countries. So what can be done to reduce this gap? This study is done to provide advice to governments who have limited funding and need to ensure that any health spending is cost effective. Until now, individual interventions have had cost effectiveness analysis but this is the first time the whole package of potential interventions has been compared using the same methods of analysis. As such this study is incredibly important to women and children in developing countries.

The authors have identified the cost effectiveness of 21 separate interventions and also examined combinations of these interventions. One intervention may reduce pre-eclampsia and subsequently reduce maternal mortality while another may reduce malnutrition and subsequent neonatal mortality. You could assess the number of life years gained by introducing these interventions, but it would be misleading because you cannot directly compare a neonatal death with a maternal death. The method used to compare multiple interventions with multiple outcomes is by converting outcomes to a common outcome such as a disability adjusted life year (DALY). This measure takes into account the effects of mortality and morbidity. Briggs and Gray have written a further description of how this sort of calculation is made.¹

The clinical effectiveness has already been assessed for all these outcomes and is in table A on bmj.com. Clearly this is a huge amount of material. For example, screening and treating for asymptomatic bacteruria will reduce neonatal pre-term deaths by 10% and also reduce maternal deaths from sepsis by 10%.

The authors then list all the other interventions in the same way, with references, so that you can trace exactly where their information came from. They then describe the actual intervention in detail (table E) so that for this example what they actually mean is the screening of urine of all pregnant women at antenatal visits and treatment of identified cases with amoxicillin (500 mg three times a day orally for three days). The researchers then identified the cost of this intervention (tables F and G) for the cost of the programme, the training of staff, the costs per patient, and the total costs.

PEP BONET/PANOS
Now, what's the most cost effective way to make you stop crying?

The authors have then used PopMod software (World Health Organization, Geneva) to calculate the DALYs from this data (http://www3.who.int/whosis/cea/Tutorial/4SPopMod_tutorial.htm). The main consideration is that their methodology for doing all this work is entirely transparent.

The last methodological issue is the use of incremental cost effectiveness ratios (ICERs) and average cost effectiveness ratios (ACERs). If I give one drug to reduce blood pressure I will prevent a certain number of strokes for a given population and cost. If I add a second drug I will reduce blood pressure further than by using just one drug, and I will prevent more strokes. The ACER is the ratio of the price of the two drugs combined against the total reduction in strokes.

The ICER is the difference in cost of the second drug in relation to the number of strokes avoided that the addition of this drug produces. So the first drug may prevent 1000 strokes at a cost of $10 000. The second drug may be equally effective on its own preventing 1000 strokes and cost the same. However when added to the first drug there is already a prevention of 1000 strokes by the first drug. The two drugs together may prevent 1500 strokes but at the total cost of $20 000. The average cost is now $13.3 (20000/1500) for each stroke prevented. The 500 additional stokes were prevented at a cost of $10 000 so at an incremental cost of $20. This is double the cost per stroke prevented than just using the one drug.

The costs and outcomes of the 21 interventions would occur over different periods, and may differ widely between programmes. For example, you might pay now to set up neonatal care but the benefits would be achieved over many years with that investment. To compare these interventions now, you need to identify the future costs on the basis of present values, by adjusting future costs with a discount rate. This is to take account of the fact that a health benefit today, for example, reducing pain, is worth more than the same health benefit in the future, and similarly money spent today is worth more than the same money spent in the future. Torgerson and Raftery have written a further description of what a discount rate is.²

The results present what each intervention or group of interventions would cost and what the likely benefit would be. The authors have identified that newborn care in both countries in South East Asia as well as countries in sub-Saharan Africa would be effective. Further down the list, however, there are some differences in syphilis screening and management of neonatal pneumonia. Screening and treatment of syphilis in countries in South East Asia with high adult and high child mortality is relatively less cost effective than in countries in sub-Saharan Africa with very high adult and high child mortality owing to its lower prevalence. Community based management of neonatal pneumonia is relatively more cost effective in sub-Saharan Africa than in South East Asia because of its higher prevalence. These are highly relevant to the health policy decision makers of those countries.

Yes. This is perhaps one of the most important studies in terms of potential impact on mortality and morbidity throughout the world. The authors are very open that this is only one way of viewing the healthcare system and that it should be considered alongside other healthcare system goals such as equity and acceptability to the people.

Is this relevant to you? Again, the answer is undoubtedly yes. It is a very clear lesson on the complexity of health care and also shows that everything we do as clinicians has an opportunity cost. That is, the money could have been used for other interventions that might have been more effective in your healthcare setting.

1. Briggs A, Gray A. Economics notes: using cost effectiveness information. BMJ 2000;320:246.
2. Torgerson DJ, Raftery J. Economic notes: discounting. BMJ 1999;319:914-5.