8 1 2 3 Estimating health impacts of interventions with a focus on small scale drinking water interventions 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 Helen Risebro, Paul R Hunter INTRODUCTION Having established a framework for identifying and costing interventions, we now move to the benefits side of economic assessment. In this chapter the focus will be on estimating the health impacts of small scale interventions giving improved access to safe drinking-water for a target group of people. The method could also be applied to other environmental health interventions such as new sanitation facilities or the introduction of educational or behavioural change programmes. Interventions can be conducted at the individual, household or community levels. They can include, for example, educating children about the importance of handwashing, digging pit latrines in people's homes or building small-scale water treatment plants collectively. Most of the discussion will be around how to estimate the likely change in disease prevalence and incidence that would follow a particular intervention. However, any environmental intervention is likely to have broader impacts than just those that impact on 1 Estimating health impacts of interventions with a focus on small scale drinking water interventions In Press, World Health Organization, Geneva http://www.who.int/water_sanitation_health/economic/chapter8.pdf20 21 22 23 24 25 26 27 28 29 disease occurrence. These broader impacts will largely be dealt with in Chapter 9. This chapter will also consider, however, those livelihood impacts that are primarily mediated by the expected reduction in disease. In order to estimate the impact of an intervention on disease burden, one needs to know the how common the disease in question is currently (see Chapter 4) and the likely impact of the proposed intervention on disease occurrence (usually: what proportion of illness will be prevented). The Expected disease reduction is then given by impactExpectedfrequencyCurrentreductionExpected ___×=30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 The expected reduction in disease burden can then be derived from the expected reduction in disease frequency multiplied by the average disease burden per case. As suggested in Chapter 5, probably the best estimates come from studies conducted in the area of the proposed intervention. However, few analyses will be fortunate enough to have access to such primary data and they will rather have to be rely on previous studies or global estimations. In this context the options available to the analyst are to conduct an intervention study in the area under investigation, to rely on previous studies of interventions in similar situations (or, preferably, on the conclusions of systematic reviews of such studies), to use risk assessment or rely on global estimates of disease. This chapter will briefly describe these approaches, giving primary data where available. Finally, the chapter will consider the situation of phased interventions, and the special situations where chronic disease is a major impact of lack of access to safe water or where a substantial proportion of the population is particularly vulnerable, as, for example, in the case where the prevalence of HIV infection is high. INTERVENTION STUDIES When attempting to estimate the impact of an intervention within a primary study there are several factors that the analyst should consider: • First: is it necessary to conduct a new study or is there sufficient information already available to enable him/her to estimate the likely health impact? • What resources are available in order to conduct any study that may need to be done? Certain study designs can be very expensive. • How urgently are the results likely to be needed? • What is the nature of the proposed intervention? Is it likely to have impact at the individual, household or community level? 260 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 The impact of interventions can be introduced in a variety of study designs including cross-sectional, cohort and randomised controlled trial (RCT) interventions. Intervention studies can be conducted under natural conditions (i.e. accidental trials, such as, outbreaks), under uncontrolled conditions (i.e. public measures, such as the introduction of a new water treatment plant), or under controlled conditions (i.e. clinical trials or field studies) (Payment & Hunter, 2003). For small-scale interventions, such as the installation of a new water treatment plant for a village, a cohort or cross-sectional survey may be most appropriate and would certainly be less costly than an RCT. As discussed in chapter 4, in the Cohort approach the investigator follows a group of people over time to determine whether they develop illness or how frequently they become ill. The cross-sectional survey will interview people within the target population at some single point to determine the prevalence of disease. In either cohort or cross-sectional study design, it is important to survey the target group that has received the intervention and one or more comparable control groups yet to receive an intervention. The incidence or prevalence of illness can then be compared between the groups and, assuming other things being equal, any difference can be attributed to the intervention. The double difference evaluation approach combines both longitudinal and cross-sectional data, but is very resource demanding. Without such controls, however, the big problem with these study designs is attribution and the assumption that other things are equal. It is often the case that the wealthiest village with most political muscle gets the new water treatment plant first; this village would be expected to be healthier anyway. Some of these biases can be overcome if the communities under study are surveyed for a period before the intervention as well as after. The RCT is one of the most robust epidemiological study designs and permits simultaneous comparison of outcomes in a group of individuals. Study participants are randomly assigned to one or more intervention groups expected to influence disease status, or the control group which receives either the status quo or a placebo (sham) intervention. RCTs overcome many of the problems with bias that affect other study designs, though
View Full Document
Unlocking...