49. Learning and Developmental Disabilities

Interventions in Low- and Middle-Income Countries

Numerous interventions are effective in preventing LDD. Table 49.2 provides a summary of these interventions classified on two axes. The horizontal axis distinguishes whether the intervention would accomplish primary, secondary, or tertiary prevention of disability. The vertical axis distinguishes four levels of evidence for cost-effectiveness:

• evidence available for LMICs

• evidence available for high-income countries only

• evidence for cost-effectiveness not available, but cost-effectiveness can be estimated from existing data

• evidence not available, but potential for benefits exists.

The literature indicates that the economic outcomes of a given intervention may vary widely for two reasons:

• Variations exist across populations, even within the same country, in the prevalence of the disorder, the cost of health care, and the existing infrastructure available to implement the intervention.

• Differences between studies exist in analytical methods used, such as the willingness to pay versus the human capital approach to valuation, and in cost categories, such as whether to include parental time costs. Though these differences make cross-population comparisons difficult, the overall evidence of cost-effectiveness is demonstrated by repeated findings that the benefits of a particular intervention outweigh the costs in a number of different settings.

Current evidence suggests that three interventions are cost saving: folic acid fortification to prevent NTDs, prenatal screening and selective pregnancy termination to prevent DS, and neonatal screening and treatment for congenital hypothyroidism (CH).

Too little is known about the fourth type of intervention, community-based rehabilitation, to evaluate it. There is a paucity of knowledge and a history of failed interventions for the prevention of premature birth and the disabilities associated with premature birth.

Neural Tube Defects: Burden and Cost-Effectiveness of Folic Acid Fortification

NTDs, which are the most common malformation of the central nervous system, result from failure of the neural tube to close during the first month of pregnancy. Anencephaly typically results in pregnancy loss, stillbirth, or neonatal death. Spina bifida (open spine defect) is associated with a range of functional deficits (requiring multiple surgical and rehabilitative interventions), including paralysis of the lower extremities and often primary enuresis and cognitive disabilities. Large geographic variations in the prevalence of NTDs exist both within and between countries. The burden of disease is highest in South Asia and lowest in LMICs of Europe and Central Asia. Similarly, deaths from NTDs are high in South Asia but lowest in high-income countries. Estimates suggest that almost all NTD disease burden is concentrated in the age group zero to four years (Mathers 2006).

Folic Acid

Folate is a vitamin that occurs naturally in green leafy vegetables, legumes, citrus, and other foods. Folic acid (FA) is an easily absorbed synthetic form of folate that can be delivered as a dietary supplement or through FA fortification of flour or other common staple foods. NTDs can be reduced by 70 percent if women consume 400 Mug of FA daily around the time of conception and until closure of the neural tube. At a population level, either supplementation or fortification of the food supply is necessary to ensure that 400 Mug of FA is consumed at the critical period of fetal development, as this dose is higher than can reasonably be consumed by relying on naturally occurring folate in foods. Fortification is much more likely than supplementation to reach the population at risk because the benefit of enhanced FA intake occurs early, typically before the pregnancy is recognized. Fortification is of particular value to women who may not receive prenatal care until the third trimester.

This section considers only evidence of cost-effectiveness of FA fortification in LMICs with respect to the benefit of preventing NTDs. Additional health benefits can be expected with respect to stroke, heart disease, and cancer.

Cost-Effectiveness of Folic Acid Fortification

A cost-benefit analysis of grain fortification in the United States (Romano and others 1995) included costs related to the addition of FA to food, to annual testing and surveillance, to a one-time packaging change, and to potential (though not substantiated) adverse health effects associated with undiagnosed vitamin B₁₂ deficiency. Benefits included avoided costs of NTDs, such as mortality costs (particularly for anencephaly) and costs of caring for those with spina bifida. The benefits of fortification outweighed costs with cost-benefit ratios of 1 to 4.3 for low-level fortification and 1 to 6.1 for high-level fortification.

Cost-effectiveness relative to status quo of FA fortification depends on several factors:

• Costs of food fortification depend on the types and quantity of food that are fortified and the level of fortification.

• The proportion of the target population reached by the fortified food is important since, in most LMICs, many people consume food produced on their own farms or within their villages.

• Grains from large mills are relatively cheap to fortify; more resources are required to fortify grains milled in smaller neighborhood mills.

• The amount of folate consumed by different populations in the absence of fortification varies.

• Prevalence of NTDs varies across populations, and the cost-effectiveness increases with prevalence.

Costs of food fortification may be lower in high-income countries, where most people consume cereals processed in a few large mills, equipment for fortification is likely to be in place, and quality assurance is facilitated. In contrast, mills in LMICs lack fortification equipment and capital, and running costs are higher in the short run.

Costs of Food Fortification

For optimal daily consumption, the actual level of food fortification (defined as Mug of FA per 100 grams of the food item) should be adjusted for storage and other losses so that a daily dose of 400 Mug is achieved. Food items that should be fortified depend on specific dietary habits. Staples such as rice and flour are obvious choices; salt, sugar, bread, milk, and edible oils are promising candidates. There are economies of scale in FA fortification. It can be and usually is carried out in conjunction with other forms of fortification, such as iron, iodine, and vitamin A fortification. Many food items are already fortified in high-income countries. Other factors to be considered in the choice of food for fortification are items that are centrally processed and allow for quality control. Soy sauce in China is an example: it is consumed on a daily basis by 70 percent of the population and is prepared in a few large factories.

The recommended fortification level is thus 240 Mug per 100 grams of the staple food. This fortification rate is assumed for all regional strata where the per capita staple consumption per day is less than 300 grams. Wheat, rice, maize, or a combination of these foods is the staple in most countries. The recommended level of FA fortification varies from 150 Mug to 240 Mug per 100 grams of cereal. So that women receive a daily dose of 400 Mug, the target cereals for fortification should be those for which daily per capita consumption is at least 200 grams. In Sub-Saharan Africa, daily per capita cereal consumption exceeds 200 grams only if wheat, rice, and maize are considered together.

Quality assurance is done through analytic testing of fortified products to confirm FA levels. Quality assurance costs in the United States are estimated at US$0.64 cents per ton of fortified grain in quality assurance costs.

The costs of FA fortification include the cost of FA, setup, and analytic testing. The analysis is done using two different cost estimates: US$0.15 and US$0.50 per ton of grain fortified. The cost of FA determines the cost of premix added to the flour. FA is almost never added alone; usually FA, iron, zinc, and niacin are added in combination. The material cost of FA alone is about US$0.10 to US$0.20 per metric ton of milled wheat. However, a more realistic cost for the premix (including other supplements) is about US$0.50 per metric ton of milled wheat. This higher estimate is conservative and does not account for the health benefits from the other supplements. Either way, the per capita costs are only a few cents in each region. The low per capita cost in high-income countries of US$0.009 assumes that 80 percent of the cereal supply is fortified. In South Asia, where NTDs have the highest burden, the per capita cost is estimated at US$0.067 (Bagriansky n.d.).

Benefits of Folic Acid Fortification

The cost-effectiveness of FA fortification in terms of its cost per DALY and per death averted assumes that the fortification strategy will reduce the incidence of NTDs by 50 percent. The costs are relatively high because of the high cost of FA. Even a few cents per capita becomes expensive if the per capita prevalence of NTD is very low.

Other Costs and Benefits

The benefits of FA fortification outweigh the costs. The benefits estimated here are conservative for three reasons:

• Strokes and coronary deaths are also prevented by FA fortification and occur more frequently than NTDs.

• The percentage of NTDs that can be prevented by FA fortification may be greater than 50 percent, because up to 70 percent of NTDs can be prevented by 400 Mug of periconceptional FA daily.

• These estimates do not take account of the costs of clinical care and management for complications when NTDs are not prevented.

Interventions to Prevent Disability Caused by Down Syndrome

Screening programs are critical public health interventions that use universal or targeted screening tests to identify potential causes or cases of LDD, including DS.

Prenatal and Neonatal Screening

Prenatal screening for genetic abnormalities allows parents to determine whether to continue with an affected pregnancy, whereas neonatal screening's fundamental purpose is to improve the infant's prognosis through early diagnosis and treatment.

A number of LDDs have been screened for in high-income countries since the 1960s, and researchers have conducted economic evaluations of these screening programs, including those for Tay-Sachs disease carriers, DS (Cusick and others 2003), sickle cell disease (Panepinto and others 2000), phenylketonuria (Lord and others 1999), and several other inborn errors of metabolism (Insinga, Laessig, and Hoffman 2002).

Estimates for Prenatal Screening, Diagnosis, and Selective Pregnancy Termination for Down Syndrome

DS is the most common genetic cause of mental retardation. Identifying a fetus with DS before birth and giving parents the option to terminate the pregnancy early can help decrease the burden of the disease on families and society. During counseling, parents may receive information about the consequences of DS, which will allow them to make an informed decision about the best care for the newborn or about termination of the pregnancy. Prenatal screening services provide an opportunity to profoundly reduce the impact of MR. The cost-effectiveness of prenatal screening for DS is based on two parameters: efficacy (by assessing the false positive rate of screening procedures and the number of fetal losses caused by screening) and financial costs (costs of screening per DS pregnancy averted). On the basis of the evidence, the best screening method is proposed, and sensitivity of the parameters of interest to the LMIC is tested. No formal comparisons are made between the costs of screening and care for a person with DS. The purpose of this analysis is to suggest the most cost-effective way of screening that provides families with information about the health of the child; it is not a cost-benefit analysis of whether a couple should terminate a pregnancy.

Burden

DS is caused by trisomy of chromosome 21—an extra chromosome rather than the usual diploid form—and is a major cause of severe MR (IQ less than 50 with substantial deficits in adaptive behavior). The incidence of DS is higher than the birth prevalence because many fetuses are spontaneously miscarried and, in some cases, selectively terminated. In the absence of prenatal screening and intervention, most DS conceptions (71 percent) result in spontaneous abortion; another 3 percent result in stillbirth, and 26 percent result in live birth with subsequent LDD (Kline, Stein, and Susser 1989). Because the incidence of DS cannot be determined without doing surveillance of all conceptions, the frequency of DS is typically measured in terms of prevalence per 1,000 live births rather than in terms of incidence. Thus, the population prevalence of DS varies depending on the maternal age structure (steep increase after age 35 years) as well as the availability and use of prenatal diagnosis followed by selective termination. Estimates from 10 LMICs range widely, from 0.1 out of every 1,000 live births in Indonesia to 4.4 out of every 1,000 live births in Pakistan (Institute of Medicine 2003). Most studies, in both high-income countries and LMICs, show DS birth prevalence in the range of 1.0 to 1.6 out of every 1,000 births. The birth prevalence of DS is likely higher in LMICs because of a higher proportion of births among women over age 35 (11 to 15 percent) relative to that in high-income countries (5 to 9 percent) (Kline, Stein, and Susser 1989) and possibly because of differential access to prenatal screening for chromosomal abnormalities.

Life Expectancy and Quality of Life

Life expectancy for children with DS is substantially lower than that of the general population. Congenital heart disease occurs in 40 to 60 percent of children with DS and accounts for 30 to 35 percent of deaths. Survival and life expectancy of children with DS have increased dramatically: In a 1940-60 birth cohort in England, only 50 percent of infants with DS survived beyond age two. By comparison, in 1981-85, 90 percent survived beyond age five (McGrother and Marshall 1990). Table 49.3 describes the estimated total deaths caused by DS by region, as well as the estimated total DALYs lost.

DS is always associated with cognitive impairment. Disability can range from mild to profound, and most children are affected moderately (IQ 40-55). Early intervention and therapy can improve functional outcomes. Of children with DS, 60 to 80 percent have hearing loss, and approximately 70 percent have ophthalmologic problems. As life expectancy of DS individuals has increased, many grow to adulthood and face an increased risk of early onset Alzheimer's disease, cataracts, hearing loss, hypothyroidism, and degenerative vascular disease.

Costs of Care

Based on 1988 data, the estimated incremental lifetime economic costs of DS are US$410,000 per case or US$647,709 in 2004 dollars (Waitzman, Romano, and Scheffler 1994). In another study, the estimate of per capita incremental costs of DS, converted to 2004 dollars, include net medical costs of US$168,567, developmental services costs of US$80,530, special education costs of US$171,593, and total costs of US$420,690 (Waitzman, Romano, and Scheffler 1994).

An estimate of lifetime costs per live born baby with DS—including education, health, and lost productivity costs—ranged from US$137,000 in 1990 to US$515,000 in 1993 (Gilbert and others 2001). Net savings using the annual program of screening, diagnosis, and selective termination was estimated to be US$885, with costs of US$446,000 per 10,000 pregnancies for a program that detects and prevents 9.7 DS births per year and a lower bound estimate of US$137,000 of potential lifetime costs per 9.7 births prevented.

The increased life span of individuals with DS and accompanying age-associated morbidity impose heavy demands on medical care and community services, as well as on sustained support from family members. It is also important to note that dollar costs of care for a DS child in LMICs would be much lower than such costs in high-income countries because of lower prices as well as lower treatment intensity. For example, in some countries, congenital heart disease, which affects 40 to 60 percent of DS children, cannot be treated effectively. This lack of treatment will lower costs of care as well as life expectancy, and cost estimates will vary for each individual area or region.

Cost-Effectiveness of Prenatal Screening, Diagnosis, and Selective Pregnancy Termination

Prenatal screening can be implemented to allow selective termination of DS pregnancies and prevention of disability related to DS in the population. This intervention raises ethical, social, and cultural concerns for some individuals and populations that may preclude its applicability.

A screening program incorporating maternal serum triple screening in all pregnant women, regardless of maternal age, yields an excellent DS detection rate and is associated with a low false-positive rate (Wald and others 2003). DS pregnancies yield lower levels of alpha-fetoprotein and unconjugated estriol but have elevated levels of human chorionic gonadotropin compared with other pregnancies. Ultrasound evaluation of the fetus neck thickness improves screening sensitivity. It is also useful when used in conjunction with serum screening (Wald and others 2003). A positive screening result is followed by diagnosis using amniocentesis or chorionic villus sampling (CVS).

Although both diagnostic procedures are guided by ultrasound to reduce risk, they are invasive, are more expensive than the screening procedure, and carry a small risk of miscarriage of an unaffected pregnancy. Thus, only a select group screening positive for possible trisomy 21 are offered the invasive diagnostic procedures. Amniocentesis, which involves the aspiration of amniotic fluid, is performed between the 14th and 16th weeks of pregnancy. CVS involves aspiration of villi and can be performed between the 10th and 12th weeks of pregnancy. Although CVS can be performed earlier in the pregnancy, amniocentesis is easier to perform and is more widely used in the second trimester. Following diagnostic confirmation of DS, parents are provided with genetic counseling and the option of terminating the pregnancy.

Although DS risk increases with maternal age, most births occur in younger women and, therefore, two-thirds of all DS births occur in younger mothers (Ross and Elias 1997). If prenatal diagnosis is available only for mothers 35 years or older, only 33 percent of DS births will be detected. Studies demonstrate that heavy reliance on maternal age to screen for DS may not be desirable in LMICs. Maternal age factor is not so useful in settings where early marriage and motherhood are the norm and most DS pregnancies involve mothers younger than 35 (Gupta and others 2001). Therefore, maternal serum screening of all pregnant women is important in preventing DS births and achieving cost-effectiveness (Wald and others 2003).

Procedure Costs

Genetic screening and counseling services are expensive. Even after initial high fixed costs to establish prenatal screening services, provision of high-quality services requires staff training, equipment, and laboratory maintenance. A recent report suggests establishing genetic screening services when other public health interventions have reduced the infant mortality rate to the range of 20 to 40 out of every 1,000 live births (Institute of Medicine 2003). Above this level, other public health interventions may have greater benefits.

The breakdown of tasks is as follows:

• screening costs, which consist of laboratory expenses (consumables and staff); informing women of results (by mail if negative, by phone if positive); service costs (processing results and monitoring the service); training in ultrasound measurement of neck skin translucency; and overhead expenses

• diagnostic costs, which comprise counseling before CVS or amniocentesis, equipment and staff for these procedures, laboratory expenses (consumables and staff), and overhead expenses

• costs of termination of selected pregnancies, which include surgical dilation, evacuation (11 to 13 weeks), or medical termination with mifepristone (after 13 weeks).

We assume infrastructure exists for prenatal screening, diagnosis, and intervention. We use the following costs: triple serum test, US$70; amniocentesis, US$1,200; genetic counseling, US$100; and termination of pregnancy, US$2,000. These cost estimates have been widely used in the literature (Cusick and others 2003). However, the medical costs can be significantly lower in LMICs and will also vary across and within countries.

Cost-Effectiveness and Efficacy

We assume that 100 percent of women attend a prenatal clinic between 10 and 14 weeks of gestation and are offered tests in the first trimester, or between 15 and 19 weeks for the tests in the second trimester. We discuss the effect of low uptake of prenatal care and its effect on cost-effectiveness of prenatal screening programs in our sensitivity analysis.

In terms of economic considerations, it is desirable to balance the probability of the birth of a DS child with the risk of procedure-related miscarriage. Sensitivity of prenatal screening and the false-positive rates vary widely, depending on the method used. The risk of procedure-related miscarriage can vary from 0.04 to 0.8 percent (Nyberg and others 1998). We use the conservative fetal loss rate of 0.9 percent (Gilbert and others 2001) for both procedures.

Efficacy of prenatal screening is defined as the number of unaffected fetuses lost due to prenatal testing per each DS birth averted (Institute of Medicine 2003). The goal is to minimize this ratio. The efficacy of prenatal screening varies with prevalence, and the primary determinant of variations in prevalence of DS is the age structure of women giving birth. The prevalence of DS and the efficacy of prenatal screening increase with the percentage of births to mothers over the age of 35. In this analysis, a 90 percent rate of selective termination is used (Waitzman, Romano, and Scheffler 1994). On this basis, the number of fetal losses per DS birth avoided varies from 7.13 (for 1 in 10,000 prevalence) to 0.16 (for 44 in 10,000 prevalence). Therefore, in countries with low prevalence of DS, such as Indonesia, more unaffected fetuses are lost than DS births averted because of screening. In areas where the ratio of unaffected fetal losses to DS births avoided is above 1, the efficacy of screening for DS is questionable.

Because of higher loss rates for CVS, we use a 1.5 percent fetal loss rate in our sensitivity analysis (Lippman and others 1992). Other costs not considered in this study are the psychological effects of a positive test on the parents, anxiety that may persist from a false-positive test, and potential complications resulting from pregnancy termination. Complications from termination may vary (Stray-Pedersen and others 1991) and may not be the same in LMICs, which should be taken into account. The sensitivity rate for the triple serum test followed by the amniocentesis is 62.3 percent in the clinical trials (Vintzileos and others 2000), and the uptake of amniocentesis is 90 percent for affected mothers and 80 percent for unaffected mothers (Waitzman, Romano, and Scheffler 1994). We assume the false-positive rate of 5 percent. The false-positive rate affects the probability of losing an unaffected fetus as a result of invasive testing that follows serum screening.

Financial cost-effectiveness is defined as the screening costs per DS birth averted. It is presented in table 49.4. Cost-effectiveness is the highest in countries with high birth prevalence of DS, given that women have access and receive prenatal care. Costs of prenatal screening and termination per DS birth averted vary from US$1,497,390 in Indonesia (for 1 in 10,000 prevalence) to US$37,185 in Pakistan (44 in 10,000 prevalence). A similar relationship is seen between prevalence and cost per DALY. In our analysis, we use costs data that are based on estimates from developed countries. Because costs of care will vary widely across and within countries, cost estimates should be done for individual regions. Lower costs of care will reduce cost-effectiveness of prenatal screening for DS. However, even after the cost adjustment, it is unlikely that the benefits will completely go away, because of the large difference between a relatively cheap screening program and high burden of disease of DS.

Sensitivity Analysis

The results of the analysis above depend on assumptions that may not hold in some LMICs. For example, if many women accept screening but few decide to have an amniocentesis, cost-effectiveness will be adversely affected. The public health benefits of screening for DS in socioeconomically deprived areas are small because of low uptake of amniocentesis (Ford and others 1998). With lower uptake rates of amniocentesis, both efficacy and financial cost-effectiveness are adversely affected as a result of low detection rates, and the number of unaffected fetal losses decreases. It is also important to note that, in some countries, many women may not have access to prenatal care or may not seek prenatal care and prenatal testing. In such areas, programs that try to reduce DS prevalence will have limited success, especially if a population at greater risk of DS is not tested.

Cost-effectiveness is often measured per DS birth averted since reduction in DS prevalence is the ultimate goal of prenatal testing. In many cultures, an abortion is not an acceptable option. Acceptance of elective termination of pregnancy may also vary across ethnicities and other subgroups within a given country. A study in California found the uptake of termination following the DS diagnosis varied from 47.5 percent for Hispanics to 65.8 percent for whites and 70.8 percent for Asians (Cunningham and Tompkinison 1999). If few families decide to terminate pregnancy to avoid having a child with severe disability, cost-effectiveness per DS birth averted will be adversely affected, and the screening program may fail to reduce the birth prevalence of DS. If a large percentage of families are opposed to induced abortion of fetal DS, the uptake of amniocentesis also will be low.

Because fetal losses following CVS are often higher than those for amniocentesis, efficacy analysis should be conducted assuming a 1.5 percent fetal loss risk attributable to invasive testing in the first trimester. With higher fetal losses, the efficacy of the prenatal screening is adversely affected, although the cost-effectiveness will not change.

In addition, assuming a higher false-positive rate of 8.3 percent increases the number of invasive tests on unaffected mothers and the number of unaffected fetal losses, thus adversely affecting the efficacy of the prenatal testing (Vintzileos and others 2000).

The analysis presented above is limited to an evaluation of the cost-effectiveness of prenatal screening for DS only. Some serum markers (for example, alpha-fetoprotein) will identify other abnormalities, the benefits of which are not included in this analysis.

Equity and Access

The desirable policy is that women of similar risk for DS have equal access to diagnostic tests. With limited access to prenatal care, the introduction of the screening programs can have small public health effects. Although the approach used in cost-effectiveness analysis is optimization of societal net benefit, the policies to be recommended for the prevention of disability must also consider individuals' freedom in decision making at each step of the prenatal diagnosis. Successful policies need to be based on cost-effectiveness estimates that take into account the needs, sensitivities, and values of individuals and cultures (Institute of Medicine 2003).

Interventions to Prevent Disability Caused by Congenital Hypothyroidism

For CH, like DS, screening programs are critical public health interventions.

Neonatal Screening in Low- and Middle-Income Countries

When considering the costs and benefits associated with a CH screening program, one must first have an estimate of how prevalent CH is in the population so that the avoided costs associated with disability can be calculated. It is important to note that in high-income countries and in several middle-income countries screening is usually done for a series of conditions rather than for a single disorder. This fact is likely to affect the cost. In several of these conditions, the treatment includes dietary modification as well as costly prepared foods and formulas. Policies in countries where this type of screening occurs include labeling of food to alert potentially vulnerable consumers.

Several studies have examined the prevalence of CH in specific populations, with substantially varying results. A review of 13 studies reporting findings on CH prevalence identified through individual screening programs found the lowest rate to be 1 case of CH per over 6,000 screened in Thailand (Wasant, Liammongkolkul, and Srisawat 1999). Contrasting this is the highest rate reported: 1 case in 1,000 screened in Pakistan (Lakhani and others 1989). Prevalence can vary not only from one country to the next, but also within countries, depending on different analyses or subpopulations within one country. These variations demonstrate the need for identifying the appropriate population in order to conduct economic evaluations of screening interventions.

According to three cost-benefit analyses of CH screening (Layde, Von Allmen, and Oakley 1979; Barden and Kessel 1984), the benefits included savings from institutionalization, special education, medical care, lost parent and child productivity, and slightly decreased life expectancy. The costs included those of the screening program as well as the cost of treating detected cases. Overall, CH screening programs are substantially cost saving, with a cost-benefit ratio as high as 1 to 8.9 in high-income countries (Dhondt and others 1991). Such savings have not yet been evaluated in LMICs. Because the treatment is inexpensive and highly effective, it is anticipated that CH screening would also be substantially cost saving in LMICs.

Burden

Congenital hypothyroidism is a common cause of MR that can be prevented by newborn screening and treatment. By the end of the 1970s, neonatal screening programs had been established in many regions of Canada, Europe, Japan, and the United States. Thyroid hormone is required for normal brain development, and little or no thyroid hormone in the neonatal period results in damage to the nervous system. Various causes of anatomical maldevelopment of the thyroid gland are responsible for CH, and several genes have been implicated. With biochemical newborn screening (best conducted in centralized regional laboratories) using dried blood spots and diagnosis in the first few weeks of life, MR is avoidable. Without appropriate treatment, two-thirds of patients with CH have low IQ, and 30 percent experience severe or profound cognitive disability (Beaulieu 1994). Even with appropriate treatment, some subtle intellectual impairment and behavior deficits may still occur—the mean IQ may be approximately 10 points lower than that of the general population (Tillotson and others 1994). In the United States, infants are screened as newborns and again at two to six weeks of age to detect missed cases. For optimal outcomes, lifelong treatment with thyroid hormone is required, with subsequent monitoring and adjustments recommended every 3 to 12 months until growth and puberty are complete. Many females born with CH are now reaching childbearing age and require increased dosages of thyroid hormone during pregnancy for optimal neuropsychological outcome in their offspring.

Costs of Care

Estimated lifetime costs of care for the child with CH include the following (Barden and Kessel 1984):

• Institutional care. At the time of the study, 15 percent of congenital hypothyroid individuals were institutionalized from age 5 to 70.

• Foster care. About 25 percent of congenital hypothyroid cases received foster care from age 5 to 20.

• Residential care. Such care was provided for 40 percent of affected cases.

• Special education expenditures. Such expenses varied with the level of MR (15 percent severe, 25 percent moderate, and 40 percent mild).

In 2004, estimated lifetime costs of CH care is US$191,000, with a 6 percent discount rate. This estimate of the financial costs of care for an affected person is fairly conservative; it does not take into account lost productivity of the person with CH, a potential loss of income attributable to the time inputs of the family members who are taking care of the affected person, or effects on quality of life.

Cost-Effectiveness of Neonatal Screening

Table 49.5 presents cost-effectiveness analysis of neonatal screening for representative countries in the World Bank regions. Screening costs include blood sample collection, laboratory costs, discounted lifetime treatment cost, and costs of care for those missed by the screening. Specimen collection and laboratory costs (Barden and Kessel 1984) constitute (in 2004 dollars) US$989,000 and US$969,000, respectively, per 100,000 children tested. Lifetime discounted (at 6 percent) treatment costs are US$6,292.64 in 2004 dollars (Barden and Kessel 1984). Analysis of costs and benefits in table 49.5 shows that, although screening for the population as a whole requires considerable investment and infrastructure, the burden from the disorder is high and treatment is cheap. Screening all newborns is beneficial compared with the high costs of lifelong care for the affected individuals. Cost savings are positive for all representative countries despite high variance in the prevalence of CH. Even for a low birth prevalence estimate of 4 out of every 100,000 in Thailand (Wasant, Liammongkolkul, and Srisawat 1999), the cost savings would be US$106,326.

Effectiveness of the newborn screening in identifying the affected infants depends on the ability of the screening program to collect blood samples from all infants in the first week and to perform tests in time to initiate treatment. This effort may be difficult in some settings, where infants are born at home or released on the first day after birth and do not have contact with the health care system in the first month of their lives (Sack, Feldman, and Kaiserman 1998). The wider the coverage of the screening program, the higher will be the cost savings of screening. Also, follow-up screening for those infants who test as false negative will increase sensitivity to 100 percent and improve cost-effectiveness of the program.

In our cost-effectiveness analysis, we assumed the lifetime care and treatment costs to be similar to those estimated for the United States. However, medical costs may vary significantly among and within countries. Such variation is unlikely to alter the cost-effectiveness analysis, because the difference between program and treatment expenditures and lifetime costs will remain even after we scale the medical costs.

The analysis presented above is limited to an evaluation of the cost-effectiveness of neonatal screening for CH. For minimal extra cost, collected blood samples for CH can also be used to identify other inherited disorders, including phenylketonuria, maple syrup urine disease, and other inborn errors of metabolism. Without the benefits of early detection and treatment for these conditions, the result is severe MR.

Community-based Rehabilitation

Community-based rehabilitation is a set of low-cost approaches to providing rehabilitation services such as physical therapy, occupational therapy, prosthetics, and assistive devices for people with disabilities in developing countries. Such rehabilitation also aims to minimize stigmatization of people with disabilities and to support inclusive education and integration of people with disabilities into society. WHO and other organizations have actively promoted community-based models for providing rehabilitation services—including services for children with developmental disabilities—in resource-poor settings (Institute of Medicine Committee on Nervous System Disorders in Developing Countries 2001).

Although 80 percent or more of the world's people with disabilities live in developing countries, only 2 percent have access to rehabilitation services (WHO Community-Based Rehabilitation 1982). If families of people with disabilities are taken into account, the number of people experiencing the effects of disability is estimated to be up to 25 percent of the world population. Community-based rehabilitation is designed to expand access to rehabilitation services in poor and rural areas by providing training in rehabilitation techniques to individuals with disabilities, their family members, and others in their communities. It also attempts to change negative attitudes toward disability and to remove barriers in the physical environment that prevent people with disabilities from fully participating in society.

Costs

In the community-based rehabilitation model, community interventions are shifted from institutions and centers to the homes and communities of people with disabilities and are carried out largely by family members and volunteers. By using volunteer workers and the existing infrastructure in the communities, this form of rehabilitation minimizes costs of delivering services and is assumed to be cost-effective relative to the alternative institution-based rehabilitation (Institute of Medicine Committee on Nervous System Disorders in Developing Countries 2001; WHO Community-Based Rehabilitation 1982; Lagerkvist 1992). Institutional care has higher costs because it relies on paid staff, medical equipment, building maintenance, and medical costs. Some advocate for provision of institutional, center-based, medical, and community-based approaches in a complementary fashion. In a Zimbabwean community-based rehabilitation project two-thirds of the patients were referred to hospitals or clinics (Rottier and others 1993). Annual costs for training workshops and salaries of rehabilitation workers amounted to US$60,000 to treat 1,614 individuals with disabilities.

Little information is available about the full costs of community-based rehabilitation and how they vary across disabilities, age groups, and societies. The cost-effectiveness of such rehabilitation or whether its costs are lower than alternative rehabilitation models is unknown. It is usually implemented in settings where no other rehabilitation models exist. The costs to consumers in terms of their efforts, time, and money may be substantial (Thomas and Thomas 1998). No formal estimates are available of time costs and opportunity costs to family members involved in community-based rehabilitation. Meeting the needs of a family member with a disability may prohibit or disrupt labor force participation of the caregiver and reduce family income. This need for caregiving may especially affect women (Giacaman 2001). The effectiveness of community-based rehabilitation in improving functional outcomes for children with cerebral palsy in Bangladesh showed no improvement, but researchers unexpectedly found a significant increase in reported stress and symptoms of depression in the mothers of children in the community-based rehabilitation intervention group (McConachie and others 2000).

Efficacy

Other attempts to evaluate community-based rehabilitation in different settings and for a range of outcomes include one for preschool children with disabilities in rural Guyana. The children showed significant improvement after six months in the program, and noticeable improvements in the attitudes of parents and others toward children with disabilities were seen (O'Toole 1988). Community-based rehabilitation programs in the Philippines and Zimbabwe found gains in activities of daily living and communication as well as higher rates of starting school and employment after six months in the program (Lagerkvist 1992). Similarly, people with disabilities participating in a community-based rehabilitation program in Botswana showed high levels of independence in activities of daily living; 20 percent of adults were working, and most school-age children were attending class (Lundgren-Lindquist and Nordholm 1996).

Some have questioned the efficacy of community-based rehabilitation and its ability to expand access for people with disabilities. Many people with disabilities do not patronize such programs, and many who do try leave dissatisfied (Kassah 1998). After initial diagnosis, only half of the identified individuals with disabilities continued (Rottier and others 1993). Community-based rehabilitation programs also face the difficulty of working in diverse communities with unique cultural, religious, economic, and social conditions, making it difficult for a single model to meet the needs for rehabilitation services in developing countries (Crishna 1999).

Prevention of Premature Birth

Premature birth—after 20 weeks but before 37 weeks—is a powerful predictor of death and disability. Infants born before 28 weeks of gestation have a 50-fold increased risk of cerebral palsy (Drummond and Colver 2002) and heightened risk of sensory, cognitive, and educational impairment (Taylor and others 2004). The rate of preterm births in the United State has increased steadily since 1990 (MacDorman and others 2002). Survival of infants born before 34 weeks requires intensive and expensive medical care (Gilbert, Nesbitt, and Danielsen 2003), and the global survival rate differs depending on neonatal care availability (Lorenz and others 2001). Infection or inflammation of the placenta is common in preterm pregnancies (Goldenberg and Culhane 2003), and many deaths attributed to asphyxia may be caused by maternal, placental, or neonatal infection. The cerebral palsy rate is significantly higher in premature infants whose births are monitored electronically (Shy and others 1990). With the exception of magnesium sulfate administered to women in preterm labor (Crowther and others 2003), no specific interventions result in decreased cerebral palsy among premature infants (Crowther and Henderson-Smart 2004).

Dietary supplements might decrease the frequency of low-birthweight births and perhaps the frequency of marked prematurity. In Bangladesh, where the rate of preterm labor was high, women who went into labor before term were older, shorter, thinner, less educated, and more disadvantaged economically, with closer spacing of births (Begum, Buckshe, and Pande 2003). Deaths attributable to prematurity in LMICs are seldom due to poor management and are largely related to poor health facilities (Pattinson 2004).

Electronic Fetal Monitoring in Labor

For decades most cerebral palsy and a major share of MR, epilepsy, and learning and behavioral disorders of childhood were considered to be due to deprivation of oxygen supply to the fetus during birth. Recent research confirms that only a minority of cerebral palsy cases, as well as associated MR and seizures, are related to markers of birth asphyxia (Nelson and Ellenberg 1986; Torfs and others 1990). Low Apgar scores, the need for respiratory support, and neonatal seizures are more commonly due to etiologies other than asphyxia, most notably intrauterine exposure to infection (Wu and others 2003). So that medical workers could recognize the onset of asphyxia and "rescue" the fetus, continuous electronic fetal monitoring (EFM) of the fetal heart rate during labor was introduced in the 1970s. This intervention was disseminated before being tested in randomized trials to compare continuous electronic monitoring with intermittent observation by stethoscope (auscultation). Since the introduction of EFM, fetal death in labor has decreased, the cesarean section rate has quadrupled (Natale and Dodman 2003), and the rate of cerebral palsy has remained steady. Accordingly, EFM cannot be recommended for use in LMICs. Intermittent auscultation with a stethoscope appears to be the appropriate way to monitor fetal status during labor.