Calculating the Effects of Malnutrition on Economic Productivity, Health and Survival in China using PROFILES
Jay Ross
The Academy for Educational Development
Washington, DC
Chen Chunming, Fu Zhenying, He Wu, Fu Gang, Wang Yuying, Chen Mingxia
The Chinese Academy of Preventive Medicine
Beijing
December, 2001
Acknowledgements
PROFILES was originally developed in 1993 by the Office of Nutrition, United States Agency for International Development through its Nutrition Communication Project, with technical assistance from the Academy for Educational Development and The Futures Group. Main contributors to the development of the model include Bart Burkhalter, Ed Able, Margaret Parlato, Jay Ross and Victor Aguayo. This application of PROFILES in China has benefited from earlier applications and further development of the model in over 20 countries in Asia, Africa and Latin America. These would not have been possible without assistance from USAID, UNICEF, the Asian Development Bank, the World Bank, Helen Keller International and The Micronutrient Initiative.
The current use of PROFILES in China was funded by The Western Pacific Regional Office of the World Health Organization with direction from Tommaso Cavalli-Sforza. Project leader in China was He Wu of the Chinese Academy for Preventive Medicine.
This report has benefited from the useful comments of Helen Stiefel.
Contents
Acknowledgements ii
Summary of Key Results iv
1. Background 1
2. Child Deaths Related to Malnutrition 2
2.1 General Approach 2
2.2 Low Weight-for-age and Mortality in Children Aged 6-59 Months 2
2.3 Vitamin A Deficiency and Mortality in Children Aged 6-59 Months 4
3. Productivity Losses due to Malnutrition. 5
3.1 General Approach and Model Assumptions 5
3.1.1. Benefits 5
3.1.2. Costs 6
3.2 Productivity Gains Due to Improved Iodine Nutrition 6
3.3 Productivity Gains Due to Reductions in Child Stunting 7
3.4 Productivity Gains Due to Reduction of Iron Deficiency Anemia 9
3.4.1 Current Productivity of Adults 9
3.4.2 Future Productivity of Children 10
3.5 Summary of Productivity Consequences 11
3.6 Comparing Benefits and Costs 12
4. Non-Communicable Diseases 12
5. Benefits of Breastfeeding 14
5.1 Monetary value of Human Milk 14
5.2 Mortality due to Sub-Optimal Breastfeeding in Infants 14
5.3 Morbidity due to Sub-Optimal Breastfeeding in Infants 15
5.4 Breastfeeding and Fertility Reduction 17
5.5 Summary of the Benefits of Breastfeeding 18
6. Conclusions 19
7. References 21
Summary of Key Results
Mortality
Protein Energy Malnutrition (Underweight)
Since 1992, the reduction in the prevalence of underweight (from 15.7% to 10.1%) has resulted in the prevention of 176,000 child deaths, representing 10% of the total reduction in under-5 mortality from all causes over this period.
In the next ten years (2001-2010), if there is no further reduction in underweight, it will contribute to 22% of deaths among children 6-59 months old, a total of 612,000 deaths.
Almost half (46%) of these deaths will occur in the 12 most disadvantaged western provinces, although these account for only 29% of the under-5 population.
The vast majority (82%) of child deaths due to underweight are caused by mild underweight, suggesting that prevention is the only effective remedy.
Reducing the prevalence of underweight (from 10.1% to 8%) over the next ten years would save an additional 62,000 child lives.
If this reduction were achieved by reducing the prevalence in the west from 16.8% to the current national prevalence of 10.1%, 56,000 lives would be saved (90% of the 62,000 deaths averted overall).
Vitamin A
At the current prevalence (10.8%), vitamin A deficiency in children under 5 contributes to 7.5% of deaths (6-59 months).
If this prevalence is reduced by half over the next ten years, 49,000 child lives could be saved, including 24,000 in the western provinces.
Productivity
Iodine Deficiency Disorders
If the total goiter rate remains at the current level (8.3%) the cognitive damage to the fetus caused by iodine deficiency during pregnancy will result in future productivity losses over the next ten years with a net present value of ¥199 billion.
Reduction of the total goiter rate to 5% over the next ten years would result in future productivity gains with a net present value of ¥40 billion.
Of these gains, ¥24 billion would be in the western provinces alone, assuming that the TGR there is also reduced to 5% (from the current 15.1%).
Protein Energy Malnutrition (Stunting)
The reduction in the prevalence of child stunting (from 32.7% to 14.4%) since 1992 has resulted in future economic productivity gains with a net present (2001) value of ¥101 billion.
If current stunting levels remain unchanged over the next ten years, the net present value of future productivity lost will be ¥159 billion.
Reducing stunting further over the next 10 years (from 14.4% to 11%), would gain ¥20 billion.
A 40% reduction of stunting in the west during the same period (from 22.7 %) would result in future productivity gains with a net present value of ¥10 billion.
Iron Deficiency Anemia (Adults)
If adult anemia remains at current levels (women: 35.6%; men: 13.7%) the value of lost productivity over the next ten years will be ¥702 billion.
The majority (69%) of these losses will be among women.
Reducing anemia by 30% over the next ten years would result in productivity gains worth ¥107 billion, including ¥30 billion in the western provinces.
Iron Deficiency Anemia (Children)
If the rate of anemia among children remains at its current level (21.7%) the value of lost productivity over the next ten years will be ¥2.4 trillion.
If childhood anemia is reduced by 30% over the next ten years (to 15.2%) the net present value of future productivity gained would be ¥348 billion.
Summary of Productivity Losses
The net present value of productivity losses due to stunting, iodine deficiency and anemia in adults and children in 2001 alone is ¥362 billion or 4.0 % of GDP.
All of these nutritional problems also reduce intelligence and learning ability. Although we are not able to quantify these effects, malnutrition will also reduce educational achievement, waste educational resources and lead to further reductions in economic productivity.
Non-communicable Diseases
Stunting
It is estimated that the current prevalence of stunting among children under 5 (14.4%) will contribute in the future to 22.3% of overweight, 13.9% of coronary heart disease, 2.1% of diabetes, 14.6% of hypertension, and 2.1% of stroke.
Low Birth Weight
Although low birth weight is known to be associated with non-communicable disease risk, at current levels of low birth weight (5.9%) its contribution is small relative to that of stunting.
Benefits of Breastfeeding
Economic Value
About 4.2 billion liters of human milk are produced each year in China, with a net value of ¥42 billion.
Sub-optimal breastfeeding in China results in a loss of breastmilk with a net value of ¥30 billion.
Survival
Pending verification of input data, no estimates are reported.
Health
Sub-optimal breastfeeding causes an estimated 15.6% of diarrhea cases and 13.2% of acute respiratory infections among infants. The cost to families of treating these illnesses is about ¥380 million.
Fertility
Breastfeeding in China results in an average delay in the return of fertility after birth by 13.4 months and offers a total of 21 million “couple-years of protection” against conception annually.
ii
1. Background
The analysis reported here was undertaken in collaboration with the Chinese Academy of Preventive Medicine (CAPM) with the financial support of the Western Pacific Regional Office of the World Health Organization. The purpose was to provide technical arguments for investments in nutrition. It involved the participation of a group of nutritionists and other technical experts at the CAPM, first in the definition of nutrition problems and priorities, then in the analysis itself, and finally in the formulation and articulation of the arguments. Input from a wider group of technical experts representing other fields and agencies was obtained at a day-long workshop where the methods and preliminary results were presented, feedback was obtained and an informal advisory network was set up.
PROFILES is both a process and a set of software tools designed to estimate the functional consequences of malnutrition in terms that are important to policy makers (morbidity, mortality, economic productivity, etc.) and to communicate these consequences to inform decision making. The software tools consist of spreadsheet workbooks with a series of user entry sheets that accept nutrition prevalences, demographic data, model coefficients, and economic information. Another series of sheets performs the calculations based upon relationships described in the nutrition literature, estimating functional consequences of malnutrition over a period from the present to some future year. Alternative scenarios can be described, representing different degrees of improvement in nutritional status, providing estimates of both the consequences of malnutrition if there is no improvement and the benefits of improving nutrition in line with specified targets. This paper describes the model assumptions, methods used for this purpose and the results for selected nutrition problems and consequences.
The period for the projections reported here is from 2001 through 2010. Additional analysis of the period from 1992 to 2001 provided estimates of the benefits of recent reductions in underweight and stunting on child survival and future productivity. Despite rapid improvements in overall health, nutrition and economic conditions in China, large disparities remain between urban and rural environments and between provinces. The national government has identified 12 provinces as disadvantaged and eligible for special development assistance programs: Inner Mongolia, Guangxi, Chongqing, Sichuan, Guizhou, Yunnan, Tibet, Shaanxi, Gansu, Qinghai, Ningxia, and Xinjiang, together comprising about 28% of the Chinese population. Additional separate estimates were made for this subset of provinces to illustrate the potential benefits of geographically targeted actions to improve nutrition there. This region is collectively referred to in this report as “the west” or “the western provinces.”
The calculations rely on demographic projections based on current United Nations estimates of the population in 1995 by age and sex, and projected life expectancies and fertility rates (UN, 1998). The total population obtained for the year 2000 using these methods was 1.293 billion whereas the current State Statistical Bureau population estimate is 1.266 billion. All estimates were therefore proportionally adjusted by a factor of 1.266/1.293 or 0.979. Similar procedures were used to derive demographic estimates for the 1992 model, using 1990 as the base year for projections. With the benefit of hindsight, the 1992 model was modified to reflect the observed trends in population, infant and child mortality, wages, and employment rates to bring its “projections” into line with the 2001 model.
Demographic data for the West were derived in the same manner as for China as a whole but the base year (1995) population pyramid was taken from the 1995 Chinese 1% sample survey, which provides population estimates by age and sex for each province. The fertility estimates were assumed to be the same as for China as a whole and the life expectancy as for China as a whole 5 years earlier.
A fourth model estimates the benefits of breastfeeding (BOB) in terms of reductions in morbidity, mortality and fertility and the economic value of breastmilk itself. This model provides estimates for a single year (2001).
2. Child Deaths Related to Malnutrition
2.1 General Approach
Nutritional conditions known to increase the risk of mortality in children aged 1-59 months include:
· Protein-energy malnutrition (PEM), as indicated by low weight-for-age
· Vitamin A deficiency
· Inappropriate infant feeding practices
In general, the risk of death increases with the number of nutritional deficiencies suffered. PROFILES spreadsheet models estimate the risk of death due to each condition independently. In theory, a child may be at risk of death due to the interaction of concurrent conditions so that the elimination of any one can save the child=s life. Also in theory, any of the conditions alone can be fatal. Because we do not have information on the prevalence of concurrent conditions or on the way these risk factors interact, we have estimated the effects separately. The number of deaths predicted as a consequence of each condition should not be added together because this would overestimate the combined effect of malnutrition and the overall benefit of reductions in more than one condition. The models presented here calculate the number of deaths in children 6-59 months due to low weight-for-age and vitamin A deficiency and infant deaths (0-12 months) due to suboptimal breastfeeding. To the extent that mortality is due to other nutritional conditions or due to these conditions but outside the specified age ranges, these models underestimate the full effects of malnutrition on mortality.
2.2 Low Weight-for-age and Mortality in Children Aged 6-59 Months
In children under five years of age, mortality increases exponentially (i.e., by a multiplicative factor) as PEM (as assessed by low weight-for-age) becomes more severe (Pelletier et al., 1993). This conclusion is based on a meta-analysis of the findings of eight observational studies from five countries (Bangladesh, India, Malawi, Tanzania, and Papua New Guinea). A key finding is that this exponential relationship is consistent across countries with different malnutrition prevalences, morbidity patterns, and mortality rates. Although these studies are observational and cannot strictly exclude alternative explanations, the consistency of a dose-response relationship, statistical control for potential confounders and biological plausibility argue strongly in favor of a causal explanation. Based on these findings, the model assumes that risk of death among children is 2.5, 4.6 and 8.4 times greater among children who are mildly, moderately and severely underweight, respectively. Calculations only apply to children 6 months of age and older because the studies used in the meta-analysis that generated these coefficients generally excluded mortality during the first 6 months of infancy.
To calculate mortality due to underweight, the prevalence for each level of severity is combined with the respective relative risk of mortality to obtain the proportion of deaths related to underweight. The relative risk estimates were derived from studies that used the %-of-median system and therefore require prevalence expressed in the same way. Prevalences in terms of z-scores (Table 1) were therefore converted to prevalences in %-of-median terms using the conversion formulas derived by Pelletier et al. (1994). The %-of-median prevalences, the relative risk of mortality for under-fives, and the percentage contribution to mortality by degree of PEM for China are shown in Table 2. Thus, 22% of all child mortality is attributable to PEM and its interaction with infectious diseases. The remaining 78% is due solely to other causes.