Household Deprivation and Its Linkages with Reproductive Health Utilisation

NATIONAL FAMILY HEALTH SURVEY-3Economic & Political Weekly EPW november 29, 200867Household Deprivation and Its Linkages with Reproductive Health UtilisationK Srinivasan, S K MohantyThe household deprivation scores, based on the availability of some basic amenities to a household and the presence of a literate adult member, have been applied to data sets of the three National Family Health Surveys to study the trends in deprivation levels over 1992-2006 and the correlates of selected reproductive and child heath parameters with household deprivation levels. It is found that the proportion of households classified as “deprived” on the basis of the HDS has recorded a secular declining trend over this period and that the quantum of declinein the proportions of the deprived is strongly associated with improvements in reproductive and child health parameters. Analysis of the data on malnourishment of children reveals that the availability of some basic amenities at the household level makes a significant contribution to children’s growth and prevention of malnutrition. K Srinivasan(ksrini_02@yahoo.com) and S K Mohanty (sanjay@iips.net)are with the International Institute for Population Sciences, Mumbai.In an earlier paper we constructed a simple index to measure the level of deprivation of basic amenities and services at the household level, termed the household deprivation score (HDS), and studied its association with the basic nutritional and health parameters of women in households (Srinivasan and Mohanty 2004). This index of deprivation, or the HDS, is based on six variables at the household level for rural and urban areas, four of them common to both. It measures the deprivation of households in three dimensions: (1) basic economic assets, (2) basic amenities, and (3) basic communications with the outside world. The variables used in each of these dimensions are on a binary scale. For the first dimension, they are (1) whether the household has a pucca house in an urban area or a pucca or semi-pucca house in a rural area, and (2) whether the household in an urban area has toilet facilities and the one in a rural area has some land.For the second dimension, (3) whether the household has electricity, and (4) whether the household has drinking water facilities in the residence; and for the third dimension, (5) whether there is at least one literate adult member in the household, and (6) whether the household has a radio/transistor, a TV or a bicycle. The total score is an addition of the six variable scores and ranges from 0 to 6. The lower the score, the higher the level of depriva-tion; 0 being the absolute level of deprivation and 6 indicating no deprivation on the selected variables. The details of the variables are given in Table 1 (p 68).1 Introduction We postulate that in a modern market-oriented economy the possession of basic economic, social and physical necessities of life could be considered the basis of dividing lines between dif-ferent levels of deprivation. The advantage of such a classifica-tory system is that it is based on actual physical or social posses-sions (for example, adult literacy) rather than income data, which is more difficult to measure, and that it can be used to measure the changes in deprivation levels over time. The HDS is not a direct measure of the economic condition of a household like per capita income or expenditure, or the Standard of Living Index, but a measure of the extent to which a household is deprived in the three dimensions. It is a prevalence data rather than an incidence data such as consumption expenditure. A respondent might for-get some of his or her expenditure over a period of 30 days, but he or she is not likely to make a mistake about stating what his or her basic possessions are. Using the data available from the second National Family Health Survey (NFHS-2) conducted in 1998-99, this score was

NATIONAL FAMILY HEALTH SURVEY-3 november 29, 2008 EPW Economic & Political Weekly68constructed for different segments of the population in various states. It was found to be significantly correlated with various health-related parameters such as the body-mass index (BMI) of women in a household, their nutritional intake of milk and other forms of proteins, and the extent of utilisation of public health services by household members. It was also validated in broadterms with National Sample Survey Organisation (NSSO) data on the monthly per capita consumption expenditure (MPCE) of households. This simple measure of deprivation at the house-hold level showed health conditions and income levels were highly related. 2 Objectives of the Study In this study, we want to build further on the above arguments byconstructing the measure on a uniform basis over different periods between 1992 and 2006, based on the data sets of the NFHS conducted in 1992-93, 1998-99 and 2005-06. We analyse the trends in the level of deprivation in rural and urban areas in India and the major states from 1992 to 2006. We revalidate the utility of this simple measure in terms of its association and interaction with the utilisation of public health services and the nutritional levels of women and children. We hypothesise that improvements in the scale and density of public health services cannot in themselves improve the health conditions and repro-ductive health-seeking behaviour of the population beyond a level,and that this needs to be supported and strengthened by theprovision of certain basic minimum amenities at the house-hold level. We thus not only construct and validate a simple measureof deprivation at the household level, but also empiri-callydemonstrate its utility to assess various health conditions of the population at the macro level. Its simplicity is its strength and it is useful for measuring levels and changes in deprivation at the household level across states. The extent to which the provision of basic amenities contained in the index has influenced the extent of utilisation of health services and health outcomes is clearly borne out by this study. The specific objectives of the study are the following: (1) To test the validity and usefulness of the index of deprivation based on time series data at three periods. (2) To examine the levels, trends and change in deprivation since liberalisation of the economy in 1990. (3) To study the association of household deprivation with the utilisation of various health services and health out-comes. (4) To analyse the extent to which household deprivation levels influence health outcomes even when controlling for the availability of health services.3 Methodology and Data Sets UsedHere we first discuss the household deprivation score.Household Deprivation ScoreThe deprivation scores mentioned above werecomputed for 1992-93,1998-99and 2005-06 and were revalidated in terms of their usefulness in assessing the utilisa-tion of public health services and their impact on some basic health indicators such as the prevalence of antenatal carefor pregnant women, medical attention at the time of delivery, immunisation of chil-dren, use of contraceptives, registration of births, child malnourishment, nutri-tional intake of women, and tuberculosis in households. The data for these variables were computed from the household, women, and child questionnaires of the NFHS-1,NFHS-2 andNFHS-3. The NFHS-3,the most recent in the series, provides information on population, health, and nutrition in India and each ofthe29 states. It collected information from a nationally representative sample of 109,041 households, 124,385 women aged 15 to 49, and 74,369 men aged 15 to 54. The NFHS-2had a sample of 91,196 households and interviewed 89,199 ever-marriedwomenaged 15 to 49 while the NFHS-1 had a sample of 88,652 households and interviewed 89,777 ever-married women aged 15 to 49. We refer to the periods of the surveys as 1992, 1998 and 2005 in the discussion.A description of the variables used in the construction of the HDS, healthcare utilisation and health outcomes, the data files used, the unit of analysis, and the reference periods are given in Table 2 (p 69). The data were compiled from the original data sets of the three surveys. It may be mentioned that the estimates published in the official reports are not comparable over the years owing to varying reference periods and definitional differences in some cases. Efforts were taken to re-estimate the parameters to make them comparable over the years. For example, the antenatal care in the NFHS-3 refers to antenatal care of women who gave birth in the five years preceding the survey, whereas in theNFHS-2 it refers to the last two births in the three years Table 1: Variables Used in Computing Household Deprivation Score Variable Variable Used Description Categorisation of Households on Deprivation Based on Total ScoreRural 1 Adult literacy 0= No adult literate in the household 0: “Abject Deprivation” (AD) 1= Presence of a literate adult in household 2 Type of house 0= Kucchha house 1-2: “Moderate Deprivation” (MD) 1= Semi pucca/pucca house 3 Electricity 0 = House is not electrified 3-4: “Just Above Deprivation” ( JAD) 1= House is electrified 4 Drinking water facility 0= No arrangement in the residence 5-6: ”Well Above Deprivation” (WAD) 1= Own arrangement in the residence 5 Radio/transistor, 0 = No radio/transistor, bicycle or TV bicycle or TV 1= At least one of these 6 Landholding 0= No land 1= Have some land Urban 1 Adult literacy 0= No adult literate in the household 0: “Abject Deprivation” (AD) 1= Presence of a literate adult in household 2 Type of house 0= Kucchha/semi pucca house 1-2: “Moderate Deprivation” (MD) 1= Pucca house 3 Electricity 0 = House is not electrified 3-4: “Just Above Deprivation” (JAD) 1= House is electrified 4 Drinking water 0= No arrangement in the residence 5-6: “Well Above Deprivation” (WAD) 1= Any arrangement in the residence 5 Radio/transistor, 0 = No radio/transistor, bicycle or TV bicycle or TV 1= At least one of these 6 Toilet facility 0= No toilet facility/others 1= Own/shared flush toilet/own pit toilet

NATIONAL FAMILY HEALTH SURVEY-3Economic & Political Weekly EPW november 29, 200869preceding the survey. It is different in the NFHS-1 as well. We madealltheestimatesuniform over the period of time to make them comparable. In the first phase, we analysed the trends in the proportion of households with differentHDS in 1992, 1998 and 2005 and the differentials across the states and in rural and urban areas. We then compared, by simple two-way tabulations, the extent of dif-ferences on the utilisation of various reproductive and child health services and health outcomes observed between the households categorised as deprived and those not deprived. This led to the conclusion that households deprived of basic economic and social amenities differ significantly, and adversely, in the uti-lisation of various reproductive health services and health out-comes compared to those not deprived. We empirically establish through a decomposition and multivariate analysis that even af-ter controlling for various demographic and health service avail-ability factors, the deprivation of basic amenities at the house-hold level contributes significantly to poor health conditions among its members.4 FindingsThe following are the main findings of the analysis.4.1 HouseholdDeprivationTable 3 (p 70) provides a summary of the distribution of propor-tion of households with anHDS of 0, 1, 2, 3, 4, 5 and 6 in the country as a whole as well as rural and urban areas in 1992, 1998 and 2005. Households with 0 are termed “abjectly deprived” (AD) and practi-cally have no possessions; those with scores 1 and 2 are “moderately deprived” (MD); those with scores 3 and 4 as “just above deprivation” (JAD); and those with scores 5 and 6 are “well above deprivation” (WAD). From the table it is interesting to see that during 1992 to 2006, the period of economic liberalisation and increasing glo-balisation, the proportion of households under AD in rural areas declined from 6.7% in 1992 to 3.7% in 1998 and to 2.1% in 2006. So the proportion under AD in 2006 was one-third the level in 1992. In urban areas, this declined from 2.0% in 1991 to 1.1% in 1998 and to 0.7% in 2005, again one-third the level in 1992. Thus the proportion of all households under AD in 2005 was one-third the level in 1992. This calls for a more disaggregated analysis of the income and expenditure data collected by the NSSO of house-holds below the poverty line. The meanHDS in rural areas was 2.9 in 1992, increased to 3.4 by 1998 and to 3.8 by 2005. In urban areas, it went up from 4.3 to 4.8 to 5.1. It is to be noted that rural households have con-sistently recorded a higher level of household deprivation (lower meanHDS) than urban ones. More interesting is that the coefficient of variation (CV) in rural areas was 55.2% in 1992 and this declined to 39.5% by 2005, and in urban areas from 39.5% to 24.5%, indicating that not only deprivation at the household level has declined but also variability across the mean values, implying uniform improvements in the conditions of the deprived. Further, we combinedAD andMD as a “deprived group” andJAD andWAD as a “not deprived” one to focus on the association of deprivation with healthcare utilisation and health outcomes.Table 4 (p 70)presents data on the percentage of households in the deprived category at the state level during 1992, 1998 and 2005. In 2005, 10 states had relatively low levels of deprivation, that is, below 10% – Gujarat (6.1%), Haryana (7.2%), Himachal Pradesh (1.6%), Jammu and Kashmir (4.4%), Karnataka (9.9%), Kerala (3.6%), Maharashtra (9.0%), Punjab (2.7%), New Delhi (1.4%) Table 2: Data Files, Variables Used and Reference Periods of HDS, Healthcare and Outcome Variables in Three Rounds of NFHSDomain Variable Defined Unit of Analysis and NFHS-1 NFHS-2 NFHS-3 Variable Used Data File Used (1992-93) (1998-99) (2005-06) Household deprivation Composite score on ownership Household Not available Not available Not available Computed uniformly forall periods score of household durables, amenities, land, house quality, and a literate adult member Healthcare variables Antenatal care visits Three or more visits to or by health Woman Threelivebirths Twolivebirthsin Last live birth in Last live birth in last three years professionals in last four years last three years last five years (since Jan 1988) (since Jan 1995) (since 2001) Medical assistance Deliveries assisted by a trained Children Three live births Twolivebirthsin All live births in Two live births in last three years at delivery health professional either at in last four years last three years last five years hospital or at home (since Jan 1988) (2001) Childhoodimmunisation BCG+3 DPT+ 3 Polio + Measles Children Threelivebirths Twolivebirthsin All live births in Two live births in last three years (For children 12-23 months old) in last four years last four years last five years (since Jan 1988) (2001) Contraceptive use Current use of any method of family Woman Use of any Use of any Use of any Useofanymethodamongcurrently planning, including traditional method of method of method of married women or their husbands methods contraceptioncontraception contraception Health outcome variables Birth registration Registration of births to the child Member Not available Not available Available only Used based on NFHS-3 born in last five years in NFHS-3 Malnutrition of children Children under three years who Children Children under Children under Children under Children under three years are underweight for age three years three years five years Nutritional intake Consumption of milk/curds and fruits Children Not available Available Available Used only for NFHS-3Prevalenceoftuberculosis Any member in the household Household Available Available Available Used only for NFHS-3NFHS-1: Antenatal, natal and immunisation were recorded for last three births in last four years (since January 1988).NFHS-2: Antenatal, natal and immunisation were recorded for last two births in last four years (since January 1995).NFHS-3: Antenatal care was recorded for last live birth in last five years (since January 2001). Natal and immunisation were recorded for all live births in last five years (since January 2001).

States/ India Percentage Deprived 1992-93 1998-99 2005-06

Assam 49.85 38.44 21.04 Bihar 53.13 45.68 32.69 Gujarat 25.09 17.22 6.10 Haryana 13.55 9.45 7.16 Himachal Pradesh 5.72 5.03 1.56 Jammu and Kashmir 10.81 5.57 4.37 Karnataka 31.15 18.07 9.97 Kerala 20.98 14.61 3.59 Madhya Pradesh 39.10 21.65 22.55 Maharashtra 25.93 22.20 8.98 Orissa 48.67 42.64 30.41 Punjab 7.48 3.88 2.70 Rajasthan 35.93 23.29 21.64 Tamil Nadu 35.89 25.44 15.12 West Bengal 41.73 31.75 20.88 Uttar Pradesh 30.30 24.66 14.36 New Delhi 7.53 3.04 1.39 Chhattisgarh na 16.53 15.25 Jharkhand na 49.85 20.90 Uttaranchal na 14.04 8.65 North-eastern states 36.62 26.59 15.63 India 34.60 23.40 15.7

NATIONAL FAMILY HEALTH SURVEY-3Economic & Political Weekly EPW november 29, 200871theNFHS-1 to 24.4% intheNFHS-2, but declined to 23% in the NFHS-3. For the not deprived category in rural areas it increased from 37.6% in theNFHS-1 to 39.2% in the NFHS-2, and further to 42.6% in theNFHS-3. Among the deprived in urban areas, it increased from 28.8% in theNFHS-1 to 34.6% in the NFHS-2, but declined to 29.9% in the NFHS-3. Among the not deprived in urban areas, it increased from 56.8% in the NFHS-1 to 57.4 in theNFHS-2, and further to 60.1% in the NFHS-3. Only in the use of modern methods of contraception by married women in reproductive ages is there a secular increase in use over the time periods while the differentials between rural and urban households are significantly low. But the differentials between the deprived and the not deprived categories persist over the time periods. Table 6 (p 72) provides the percentages of services availed of by mothers and children in the four reproductive and child health (RCH) service dimensions – contraceptive use, three antenatal visits, safe delivery and full immunisation of children – in differ-ent states at the three time points. The information is provided separately for deprived and not deprived households to bring out the extent of differentials between these two groups in different states over the three time periods. It was pointed out in an earlier article (Srinivasan, Chander Sekher and Arokiasamy 2007) that the pace of improvements in these four dimensions is lower between 1998 and 2006 compared with between 1992 and 1998. A major finding from Table 6 is that in all the four dimensions, in all states except Kerala, women and children in deprived households recorded a lower utilisation of RCH services than those in not deprived households. Further, the gap between the two groups didnot narrow between 1992 and 2006 in spite of various health-care programmes targeted at the poor. Kerala is an exception where the utilisation of services with regard to contraceptive use, safe deliveries and three antenatal visits are higher among the deprived than the not deprived.Contraceptive use in the deprived category is the lowest in Bihar at 21.3% and the highest in Kerala at 81.7%. Among the not deprived, it is again the lowest in Bihar at 38.7% and the highest in Himachal Pradesh at 72.7%. The gap between the two catego-ries is largest in Haryana at 28.2 percentage points and smallest in Maharashtra at 0.6 points, but for Kerala where the percentage of use among the deprived is slightly higher than among the not deprived. With regard to three antenatal visits, in the deprived category, the lowest is in Bihar at 6.4% and the highest in Kerala at 100%. Among the not deprived, the lowest is in Uttar Pradesh at 28.3% and the highest in Kerala at 99.2%. The gap between the two categories is largest in Himachal Pradesh at 65 percentage points and smallest in Kerala at 0.6 points. With regard to safe delivery, in the deprived category, the lowest is in Bihar at 12.2% and the highest is in Kerala at 100%. Among the not deprived, the lowest is in Uttar Pradesh at 30.2% and the highest is in Kerala at 99.7%. The gap between the two categories is largest in Haryana at 40.6 percentage points and smallest in Kerala at 0.3 points.With regard to full immunisation of children, in the deprived category, the lowest is in Uttar Pradesh at 10.2% and the highest is in Tamil Nadu at 71.9%. Among the not deprived, the lowest is again in Uttar Pradesh at 25.3% and the highest in Tamil Nadu at 81.7%. The gap between the two categories is largest in Himachal Pradesh at 74.5 percentage points and smallest in Tamil Nadu at 9.8. The zero value estimated for Himachal Pradesh for the deprived category of households may be due to the very small sample interviewed in this category. 4.3 Linkage with Health OutcomesThe wider and better utilisation of various reproductive and other public health services offered by the government institu-tions in rural and urban areas should eventually be reflected in better health conditions and lower morbidity and mortality levels in the population. It has to be repeated that the provision of public health services alone cannot improve the health conditions of a population beyond a limit. It has to be backed up by better social and economic conditions of the population. In this section, we examine how deprivation levels at the household level affect the health outcome in four dimensions. The health outcome variables are (a) Child malnutrition; (b) Nutritional intake of mothers; (c) Registration of births; and (d) Prevalence of tuberculosis.(a) Child Malnutrition: Table 7 (p 73) shows the proportion of childrenunder the age of three identified as severely under-weight for age (less than 3 SD of the median of the reference category)andunderweight (less than 2 SD of the median of the reference category) by levels of deprivations estimated from the data of the three surveys.Table 5: Utilisation of Reproductive Health Services by Level of Deprivation (in % ,1992-2006) Combined Rural Urban 1992-931998-992005-06 1992-931998-992005-06 1992-931998-992005-06Safe delivery AD 16.315.424.113.812.823.032.137.933.3MD 21.823.524.317.920.822.046.946.040.1JAD 30.9 38.237.525.233.033.9 60.4 66.059.1WAD 60.162.7 67.145.649.855.679.682.884.8Deprived 20.922.524.317.219.923.044.645.029.9Notdeprived 43.1 49.2 54.0 31.9 39.2 42.6 73.0 77.3 60.13 ANC visits AD 24.621.421.123.619.620.1 31.438.5 –MD 30.828.428.928.7 26.427.444.9 45.6 39.8JAD 41.141.441.437.537.538.660.2 63.4 58.7WAD 66.461.567.155.549.956.7 80.6 79.1 82.7Deprived 29.827.528.227.925.526.7 42.7 44.7 39.2Not deprived 51.550.455.7 43.342.147.1 73.6 74.1 77.4Full immunisation AD 12.924.115.112.125.615.618.513.910.0MD 22.625.7 24.7 21.424.323.830.537.9 31.7JAD 34.4 37.336.832.135.035.9 46.4 49.7 41.9WAD 54.552.356.348.746.450.262.561.065.2Deprived 21.225.523.820.024.423.028.834.629.9Not deprived 42.8 44.1 47.7 37.6 39.2 42.6 56.8 57.4 60.1Contraceptive use AD 26.130.833.526.429.832.424.641.240.6MD 32.437.943.531.637.242.637.943.450.1JAD 38.4 45.9 52.4 36.9 44.351.345.7 54.158.0WAD 51.556.7 62.146.352.7 58.7 57.361.7 66.3Deprived 31.537.142.730.836.441.736.043.249.3Not deprived 44.4 51.3 58.3 40.2 47.6 55.1 54.0 59.7 64.8

NATIONAL FAMILY HEALTH SURVEY-3 november 29, 2008 EPW Economic & Political Weekly72The severely underweight group in rural areas declined mar-ginally from 23.2% in 1992 to 19.6% by 1998, and to 17.4% by 2005, and in urban areas, from 15.6% in 1992 to 11.3% by 1998, and to 10.6% by 2005. In the deprived category, the severely un-derweight group in rural areas was 28.7% in 1992, 26.6% in 1998, and 26.0% in 2005, implying practically no decline. In urban areas, the figures were 25.9% in 1992, 24.8% in 1998, and 22.2% in 2005. In the not deprived category, the decline was from 20.1% in 1992 to 15.4% in 2005, relatively steep in percentage points when compared to the deprived group. The differentials be-tween the deprived and not deprived are larger in urban areas than in rural areas.If we consider the underweight group, which includes the moderately underweight, there is again very little decline in the deprived category in rural areas at 69.8% in 1992, 57.3% in 1998, and 54.6% in 2005. In urban areas, it fell from 68.2% in 1991 to 57.8% in 1998, and 53.5% in 2005. Thus even by 2005, in the deprived category, more than half the children under the age of three in rural (55%) and urban (54%) ar-eas were malnourished, compared to 41% and 28% in the not deprived category in ru-ral and urban areas respectively. The not deprived in urban areas seem to be faring better in child nutritional status than their rural counterparts. However, the differen-tials between the deprived and not deprived categories are striking in both areas.(b) Nutritional Intake of Mothers: The health conditions of women in India are very poor with severe underweight, high levels of anaemia and iodine deficiency among them, in addition to chronic com-municable diseases such as tuberculosis, respiratory diseases, and amebiasis. To an extent, all these conditions can be attributed to the poor nutrition intake by women and the lack of appropriate public health facilities. Table 8 (p 73) presents estimates computed from theNFHS-3 data on the frequency of intake of two types of food items, milk or curd and fruits. They are classified by different levels of deprivation for rural and urban areas, and for all households. As expected, the consumption of these essential proteins and protective foods is very low and infrequent, particularly among women in deprived households compared to those in not deprived households. In rural areas, 19.7% of women never con-sumed milk or curd in deprived households compared to 10.0% in the not deprived category.Daily consumption of milk or curd was reported by 16.3% of the households in the former category compared with 39.7% in the latter. In urban areas, 17.4% of women in the deprived category never consumedmilk or curd compared with 9.2% in the not deprived category. Daily consumption of milk or curd was reported by 21% of the households in the former category compared with 48.7% in the latter. Urban women seem to consume more of Table 6: Differentials in RCH Service Utilisation by Deprivation from Three Rounds of NFHS(1992-2006)State/Deprivation Status Contraceptive Use 3 ANC Visits Immunisation Safe Delivery 1992-931998-992005-06 1992-93 1998-992005-061992-93 1998-992005-06 1992-931998-992005-06Andhra Pradesh Deprived 39.5 53.2 63.3 65.7 65.6 72.6 33.6 40.9 36.4 39.3 41.9 68.3 Notdeprived 50.7 61.6 68.2 80.5 85 88.6 52.2 55.2 47.6 60.2 72.4 78.3Assam Deprived 32.7 33.5 43.3 16.2 17.2 18.4 6.7 13.7 11.3 8.7 9.4 12.3Notdeprived 51.948.5 59.8 35.5 41.343.534.6 17.4 38.529.6 30.438.5Bihar Deprived 13.7 13.221.3 9 7.5 6.4 4.2 6.6 19.69.4 15.121.6 Notdeprived 31.4 30.5 38.7 30.3 23.2 21.9 17.3 15.1 40.2 30.5 32.9 38.5Gujarat Deprived 40.4 51 54.9 44.8 42.6 30.8 32.8 30.2 28.6 22.1 28.2 28.3 Not deprived 51.9 60.4 67.1 67.4 65.6 67.9 55.6 52.9 46.7 53.7 59.2 67.7Haryana Deprived 34.1 47.6 36.7 29.2 21.8 31.9 32.4 45.5 14.3 16.1 25 16.9 Notdeprived 51.4 63.5 64.9 46.4 40 62.1 56 62.3 71.3 34.4 44 57.5Himachal Pradesh Deprived 55.6 66.7 66.7 22.2 33.3 0 66.7 50 0 11.1 16.7 50 Not deprived 58.5 67.7 72.7 42.2 62.8 65 63.5 83.3 74.5 27.1 41.2 50Karnataka Deprived 41.4 45.7 59.2 57.6 45.3 60 35.5 43.4 13.5 31.8 36.3 36.4 Notdeprived 51.9 60.5 64 79.5 79.3 81.6 60 58.8 58.6 61.5 64.9 74.8Kerala Deprived 65.1 70.2 81.7 90 97.5 100 43.3 70.3 20 82 84.9 100 Notdeprived 62.8 62.9 68.2 96.6 99.5 99.2 57.7 72.2 77 93.1 95.3 99.7Madhya Pradesh Deprived 28.5 31.9 45.6 16.9 11.2 25.2 18.1 7 23.2 19 12 17.4 Notdeprived 40.4 46.5 58.2 36.3 31.3 45 36.3 23.5 45.4 38.4 33.4 43Maharashtra Deprived 47.5 54.8 66.2 45.7 40.9 42.5 51.2 59.6 34.2 34.2 24.5 30.6 Notdeprived 55.7 62.3 67 68.7 73.2 79.5 68 76.2 61 60.9 68.6 74.8Orissa Deprived 33.6 39.4 41.2 25.1 36.3 42.6 24.6 32.2 33.3 13.5 21.9 22 Notdeprived 36.2 51.1 54.5 43.7 56.5 67.8 46.4 46.6 58.1 28.2 41.6 57.2Punjab Deprived 45.9 60.9 59.6 30 5.3 34.8 26.1 25 25 35.3 30.4 28.6 Notdeprived 59.5 66.8 63.4 64.7 60.4 75.4 65.6 72.4 61.4 49.4 63.9 70.8Rajasthan Deprived 21.7 25.7 29.9 8.4 12.1 19.4 10.1 3.2 14.7 12 18.3 17.1 Notdeprived 36.2 43.8 50.8 23.5 27.2 47.1 26.7 20.4 29.8 27.6 41.3 50.2Tamil Nadu Deprived 45.8 46.8 60.7 78.1 83.5 94.3 53.9 81.7 71.9 56.4 68.7 85.9 Notdeprived 51.7 53.4 61.5 92.6 93.6 96.8 71.9 90 81.7 80.6 88.4 94.2West Bengal Deprived 50.3 61.5 65.2 38.3 45.6 48.3 19.7 32.7 50 20.3 29.1 25.4 Notdeprived 61.6 68.7 72.6 61.6 64.3 67.6 47.3 49.9 68.7 45.6 52.7 56.9Uttar Pradesh Deprived 12.7 17.7 30.3 11.2 5.9 14.5 11.6 10.3 10.2 8.6 9.6 16.6 Notdeprived 22.2 29.3 45.1 28.6 17.5 28.3 23 21.7 25.3 21.8 25.4 30.2North-eastern states Deprived 27 31.3 33.6 17.2 24.2 31.4 6.4 17.4 23.1 15.5 20.3 14.8 Notdeprived 40 46.3 51 49.2 52.7 60.3 27.3 35.8 42.9 44.6 49.8 51.7

NATIONAL FAMILY HEALTH SURVEY-3Economic & Political Weekly EPW november 29, 200873milk or curd than rural women both in the deprived and not deprived categories.Consumption of fruits by women is more infrequent than con-sumption of milk or curd both in rural and urban areas in both categories of households. In rural areas, 75.4% of women in the deprived category reported consumption of fruits only occasion-ally compared with 63.3% in the not deprived category. A mere 2% of the households in the former category reported daily consumption of fruits compared with 7.7% in the latter category. In urban areas, 68.1% of women in the deprived category reported consuming fruits occasionally compared with 37.8% in the not deprived category. A mere 3.5% of households in the former category reported daily consumption of fruits compared with 25.7% in the latter. Thus women in urban areas in the not deprived category seem to consuming more nutritious food than those in the other groups.(c) Registration of Births: The Birth and Death Registration Act was passed by the Parliament of India in 1969, making it compul-sory for births to be registered within 14 days and deaths within seven days. The responsibility for implementation of this Act rests with the state governments. Table 9 provides state-wise data on the percentage of births registered in the deprived and not deprived categories of house-holds across the states. It is distressing to note that more than 30 years after the passage of the Act, only 41% of the births that occurred between 2000 and 2005 were registered, 35% in rural areas and 59% in urban areas. Among the states, the variation is large, ranging from 85% in Goa to 6% in Bihar, 7% in Uttar Pradesh, and 9% in Jharkhand. The official responsibility for the registration of births and deaths is vested with peripheral state government officials such as the village munsif/patel, the ANM or male multipurpose worker, and the panchayat secretary or the thane official, depending on the state department which is re-sponsible for the registration of vital events. And the vital events occur all over the place at all times. Given these, the extent of registration of births can be considered a good proxy for the ef-fectiveness of a state government. In other words, it can be con-sidered a simple but good index of effectiveness of state govern-ance (ESG). Viewed in this context, theESG is highest, more than 80% in Goa, Gujarat, Maharash-tra, Himachal Pradesh, Kerala, Mizoram, Sikkim and Tamil Nadu, and lowest, less than 20%, in Rajasthan, Uttar Pradesh, Bi-har and Jharkhand. None of the states in north (except Himachal Pradesh), central and east India have a goodESG (above 80% reg-istration). In the country as a whole, in the deprived category, only 10.7% of the births that took place between 2000 and 2005 were reported registered, while in the not deprived category it was 60.6%, almost six times more. Table 11 (p 74) provides state-wise figures of the percentage of births registered in the deprived and not deprived categories. In the deprived category, it varied from Table 7: Percentage of Children under Three Years Underweight by Level of Deprivation(1992-2006)Deprivation Level Severe Underweight Underweight1 1992-931998-992005-061992-931998-992005-06Combined AD 33.7 29.8 28.8 73.4 63.1 56.3 MD 27.3 25.9 25.8 68.9 56.5 54.3 JAD 22.0 19.0 19.0 43.7 50.4 46.2 WAD 12.9 10.1 9.7 32.8 34.9 30.8Deprived (AD+MD) 28.3 26.4 25.5 69.6 57.3 54.5Not deprived (JAD+WAD) 18.1 15.0 13.8 39.2 43.3 37.6All 21.517.715.949.546.740.4Rural AD 34.0 29.9 30.2 73.2 63.0 56.2 MD 27.7 26.1 26.1 69.2 56.4 54.4 JAD 22.8 19.7 19.4 44.5 51.0 47.2 WAD 14.8 12.5 11.0 36.8 38.2 34.6Deprived (AD+MD) 28.7 26.6 26.0 69.8 57.3 54.6Not deprived (JAD+WAD) 20.1 17.0 15.4 42.0 46.2 41.2All 23.219.717.652.549.343.8Urban AD 31.7 28.8 15.0 75.4 64.4 56.4 MD 24.8 24.3 23.8 66.9 57.0 53.3 JAD 17.8 15.9 16.4 39.9 47.5 39.8 WAD 10.3 6.5 7.6 27.3 29.9 24.6Deprived (AD+MD) 25.9 24.8 22.2 68.2 57.8 53.5Not deprived (JAD+WAD) 12.9 9.6 9.8 31.6 35.7 28.1All 15.611.410.639.138.330.01 Including moderately underweight.Table 8: Percentage of Women Consuming Milk and Fruits by Level of Deprivation (2005-06)Deprivation Level Milk/Curd Fruits NeverDailyWeeklyOccasionally Never Daily WeeklyOccasionallyCombined AD 20.0 12.8 12.5 54.8 8.9 1.9 13.1 76.2 MD 19.4 17.3 5.1 48.3 7.8 2.3 15.7 74.2 JAD 13.9 30.2 16.4 39.4 4.8 5.3 22.8 67.2 WAD 8.1 50.5 15.3 26.1 1.7 19.4 32.3 46.6 Deprived 19.4 16.9 14.9 48.8 7.9 2.2 15.5 74.4 Notdeprived 10.3 42.9 15.7 31.1 2.9 14.1 28.8 54.3 All 11.4 39.9 15.6 33.2 3.5 12.7 27.2 56.6Rural AD 19.7 12.1 12.8 55.4 9.4 1.9 13.1 75.6 MD 19.8 16.6 15.0 48.7 8.2 2.0 14.3 75.4 JAD 14.1 29.8 16.5 39.6 5.1 4.6 21.4 69.0 WAD 7.9 48.9 15.3 27.9 2.2 10.6 29.2 58.1 Deprived 19.7 16.3 14.8 49.2 8.3 2.0 14.2 75.4 Not deprived 10.9 39.7 15.9 33.5 3.5 7.7 25.4 63.3 All 12.2 36.2 15.7 35.9 4.3 6.9 23.7 65.2Urban AD 21.5 17.2 9.8 51.5 5.5 1.8 12.3 80.4 MD 17.1 21.4 15.9 45.7 5.0 3.6 24.5 67.0 JAD 13.0 32.3 16.0 38.6 3.7 8.6 29.7 58.1 WAD 8.4 52.2 15.3 24.2 1.2 29.4 58.1 33.4 Deprived 17.4 21.0 15.4 46.1 5.0 3.5 23.5 68.1 Not deprived 9.2 48.7 15.4 26.7 1.6 25.7 34.9 37.8 All 9.6 47.4 15.4 27.6 1.8 24.6 34.3 39.3Deprived: AD+MD, Not deprived: JAD+WAD.Table 9: Birth Registration by Deprived Groups(2005-06)Deprivation Levels % Births Registered RuralUrbanCombinedAD 18.811.117.9MD 24.530.525.2JAD 32.7 43.334.2AWAD 43.867.853.4Deprived 11.67.310.7Not deprived 51.6 83.1 60.6Total 35.159.741.4Table 10: Prevalence of Tuberculosis* by Deprived Groups(2005-06)Deprivation Levels Prevalence of Tuberculosis RuralUrbanCombinedAD 4.64.34.6MD 3.1 2.7 3.0JAD 2.61.72.4AWAD 1.61.11.4Deprived 3.22.93.2Not deprived 2.1 1.3 1.8Total 2.41.42.0*Among usual members of the household.

NATIONAL FAMILY HEALTH SURVEY-3 november 29, 2008 EPW Economic & Political Weekly74household deprivation and of course the spread of infectious bacteria, themycobacterium tuberculii. In spite of vigorousTB control programmes implemented by the central and state governments for more than four decades, the disease is still widely found in India, which is considered the major reservoir for the spread of TB in the world. This is mainly because of the deprivation of basic amenities at the household level. Table 10 (p 73) presents data on the prevalence of TB at different levels of household deprivation in 2005-06, and Table 11 the figures reported in the different states. It is significant to note that in deprived households, 3.2% hadTB compared to 1.8% in not deprived households. Across the states, in deprived house-holds, it varied from a high of 5.4% in Bihar and 4.5% in Punjab to a low of close to 0% (reported) in Jammu and Kashmir, Himachal Pradesh and Delhi. The estimates for Delhi are sur-prising and can be attributed to the small number of households in this category in theNFHS-3 sample. In not deprived house-holds, prevalence was the highest in Bihar (3.2%), Jharkhand (3.1%) and Uttar Pradesh (2.5%). The range of variation in the prevalence ofTB is lower in the not deprived category of households than in the deprived category. 4.4 QuantitativeAssessmentsThis section discusses the quantitative assessments of the effects of the changes in deprivation.(a) Decomposition of Changes in RCH Parameters: One of the simple methods used in demography for assessing the effects of changes in the levels of one or more contributing factors on a de-pendent variable over time is the method of decomposition of the changes in the dependent variable attributable to changes in the various contributing factors by the method of standardisation. In principle, this method assumes that there is no change in the dis-tribution of a particular factor over time, allowing others to vary, and estimates the effect of the changes on the dependent variable. We applied this method to five variables where the changes dur-ing 1992-98 to 1998-99 and 1998-99 to 2005-06 were substan-tial, and estimated the effects of changes in the deprivation levels during these two periods. The RCH variables on which this analy-sis was done were safe delivery, full immunisation of children, contraceptive use, severe malnourishment, and malnourishment of children. The findings are presented in Table 12. The last two columns of the table provide the percentage of changes in each of the above variables attributable to changes in household depriva-tion levels, that is, reductions in the percentage of the deprived during 1992-98 and 1998-2005. The first panel provides the fig-ures for the combined population, the second for rural, and the last for urban populations. From Table 12 it can be seen that for most of the variables in both the time periods, between 30% and 40% of the changes in theRCH variables are attributable to changes in the percentage of the deprived. Only the remaining percentage changes can be attributed to the impact of specific programmes. For example, in the case of mothers who had safe deliveries, 39% of the improve-ments during 1992-98 and 36% during 1998-2005 were attribut-able to declines in the percentages of the deprived in these periods. Table 12: Decomposition of Dependent Variable by Deprivation Levels Dependent Overall Overall Part Due to Part Due to Percentage Percentage Variables Change in Change in Change in Change in Change Due Change Due DependentDependentDeprivationDeprivation to Decline in to Decline in VariableVariableProportion,Proportion,DeprivationDeprivation during during 1992-98( D) 1998-2005( D) Levels Levels 1992-98 1998-2005 during during 1992-981998-2005(Combined) Safedelivery 7.5 6.44 2.94 2.35 39.26 36.44 Immunisation 4.4 4.2 2.05 1.89 46.67 44.56 Contraceptive use 8 7.9 1.56 1.23 19.50 15.60 Severeunderweight -3.9 -2.1 -1.25 -0.92 31.84 45.01 Underweight -3.1 -6.4 -1.54 -1.34 49.45 21.02Rural Safedelivery 7.72 4.92 2.3 1.61 29.75 32.66 Immunisation 4.5 3.6 1.76 1.61 39.11 44.40 Contraceptive use 8 8 1.33 1.1 16.66 13.80 Severeunderweight -3.8 -2.2 -1.14 -0.87 29.79 39.87 Underweight -3.9 -5.4 -1.32 -1.10 33.57 20.22Urban Safedelivery 6.3 -16.2 2.71 0.79 42.88 -4.84 Immunisation 3.4 2.8 1.92 0.79 55.83 28.0 Contraceptive use 7.4 5.6 1.39 0.4 18.86 7.20 Severeunderweight -4.2 -0.4 -1.28 -0.32 30.48 84.84 Underweight -4.4 -8.7 -2.63 -2.08 59.15 23.86a high of 54.6% in Tamil Nadu and 53.3% in West Bengal to a low of below 2% in Delhi, Jharkhand and Uttar Pradesh. In not deprived households, it varied from a high of over 90% in Kerala, Tamil Nadu, Gujarat, Himachal Pradesh and Maharashtra to a low of 11.8% in Bihar and 16.2% in Jharkhand, indicating the low levels of governance in these two states.(d) Prevalence of Tuberculosis:Tuberculosis (TB) is a commu-nicable disease that is widely prevalent in the context of poverty, Table 11: Birth Registration and Prevalence of Tuberculosis by Deprived Groups (All India and States,2005-06) States/ India Percentage Births Registered Percentage Suffering from TB Deprived Not Deprived All Deprived Not Deprived AllAndhra Pradesh 8.70 58.9 40.30 2.20 1.60 1.70Assam 17.659.7 443.6 2.9 3.1Bihar 2.2011.86.205.403.203.90Gujarat 37.9094.588.303.102.302.30Haryana 21.8087.772.603.601.601.00Himachal Pradesh 10.00 97 90.30 0.00 0.80 0.80Jammu and Kashmir 2.60 66.5 36.80 0.00 0.50 0.50Karnataka 11.50 76.559.70 1.300.600.60Kerala 19.4096.789.402.801.101.20Madhya Pradesh 9.20 45.7 29.80 1.90 1.60 1.70Maharashtra 32.20 9281.50 1.701.701.40Orissa 35.9072.558.602.801.101.60Punjab 9.6094.177.504.501.001.00Rajasthan 2.9032.116.403.501.401.80Tamil Nadu 54.60 91.6 85.60 2.70 1.60 1.70West Bengal 53.30 86.3 76.20 3.80 2.30 2.60Uttar Pradesh 0.40 15.3 7.3 2.80 2.50 2.50New Delhi 1.80 86.9 62.00 0.00 1.00 1.00Chhattisgarh 31.4083.573.001.201.501.40Jharkhand 1.0016.29.605.203.103.50Uttaranchal 3.0064.440.102.701.501.60North-eastern states 22.00 66.7 52.30 3.90 2.90 3.00India 10.70 60.641.40 3.201.802.00

NATIONAL FAMILY HEALTH SURVEY-3 november 29, 2008 EPW Economic & Political Weekly76deprived households have 32% less chance of not being severely malnourished than children in deprived households. In Uttar Pradesh, the reduction in risk of severe malnutrition among de-prived households is 34%, in Maharashtra, it is as high as 58% and in Tamil Nadu, it is 41%. Statistically these are highly significant reductions. Among the reproductive health variables, medically assisted delivery is statistically significant in reducing child mal-nutrition in Maharashtra and Tamil Nadu, and the provision of three or more antenatal visits in Uttar Pradesh. The nutrition variable, especially the consumption of fruits and vegetables during preg-nancy, is significant in all the three states and at the all-India level. Among the 12 predictors, seven turn out to be statistically significant in Uttar Pradesh, five in Maharashtra, and three in Tamil Nadu. So, as a state develops its health services, the differ-ential impact of the many factors on child health tend to dimin-ish, possibly because of the greater availability of basic facilities, nutrition and healthcare. It is interesting that household depriva-tion remains a significant factor in its relation to child malnutri-tion in all the three states. 5 Summary of Major FindingsOur analysis based on the HDS for 1992-93, 1998-99 and 2005-06, which was worked out by analysing the NFHS-1, NFHS-2 and NFHS-3 data, reveals that the percentage of households termed deprived on a uniform definition of the term substantially de-clined during 1992-2005, the period of economic liberalisation. It fell from 41% to 21% in rural areas and from 18% to 7% in urban areas. There are large inter-state variations in the quantum of decline but all states have recorded a decline. Our classification of deprived households broadly agrees with the percentage be-low the poverty line calculated by the Planning Commission from MPCE data collected by the NSSO. The proportion of households experiencing abject deprivation in 2005 was one-third the level in 1992. In rural areas, this declined from 6.7% in 1992 to 3.7% in 1998, and to 2.1% in 2005, and in urban areas, this fell from 2.0% in 1992 to 1.1% in 1998, and to 0.7% in 2005. This calls for a more disaggregated analysis of the income and expenditure data on households below the poverty line collected by the NSSO. Our major findings from analysing differentials in the utilisa-tion of various reproductive healthcare services and the health outcomes of households in the deprived and not deprived catego-ries during 1992-2005 are the following.(1) There are large and statistically significant differentials be-tween deprived and not deprived households at all the three time points, and there has been no significant decline in the differen-tials over time. This has been observed in terms of the proportion of mothers who received medical attention at the time of delivery, pregnant women who received three antenatal visits, and children under the age of three who were fully immunised. In some cases, such as skilled attention at the time of delivery, the differentials between the two groups have actually increased. Thus while the proportion of population in the deprived category has come down over time, the reproductive health services received by them have not improved.(2) The deprived suffer disproportionately more from conditions such as severe malnutrition andTB. For example, in 2005 in rural areas, 26% of the children from deprived households were se-verely malnourished compared with 15% in the not deprived cat-egory. Such differentials are observed in most of the states and in urban areas as well. With regard to TB, in deprived households, 3.2% suffered from the disease compared with 1.8% in the not deprived category. Again, there are large differentials between the states, but the differentials between the deprived and not de-prived categories have persisted in the states.(3) The effects of deprivation at the household level, as defined in this study, seem to influence the extent of utilisation of various reproductive and child health services and health outcomes to the extent of 30% to 40%, with less than 20% on the variable of contraceptive use. This is true even after control-ling for the availability of health services in different areas by way of a logit analysis. It suggests that increasing the funding for public health services cannot by itself improve the utilisation of such services or improve health outcomes beyond a limit. There is the need to initiate developmental programmes aimed at a more rapid reduction of deprivation in basic amenities at the household level that are closely linked with healthcare utilisation.(4) Our study emphasises the need for a more detailed investiga-tion of the factors that influence the relatively poor utilisation of public health services by deprived households. And the need to formulate appropriate policies and programmes to improve their economic conditions while providing the health programmes required by them. The health outcomes of various programmes implemented by the central and state governments through their public health departments aimed at reducing infant and child mortality levels, general morbidity levels and the prevalence of diseases, can only succeed to a limited extent. Beyond this, im-provements depend on the economic condition of the population, especially in the matters of housing, water supply, sanitary condi-tions, literacy, and consumption of nutritious foods. A depriva-tion of these basic amenities at the household level seem to have a greater negative impact than any other factor on the utilisation of various health services and health outcomes in the population in states as divergent as Tamil Nadu and Uttar Pradesh. The simple measure of the HDS used in this analysis is significantly linked to health outcomes under different public health conditions.ReferencesGovernment of India (1946):Report of the Health Survey and Development Committee, Vol 1 (Delhi: Manager of Publications). – (1961): Health Survey and Planning Committee Re-port (New Delhi: Ministry of Health).IIPS (1994):National Family Health Survey (NFHS 1), 1992-93: India (Mumbai: International Institute for Population Sciences). – (2000):National Family Health Survey (NFHS 2), 1998-99: India (Mumbai: International Institute for Population Sciences).IIPS and Macro International (2007): National Family Health Survey (NFHS-3), 2005-06: India, Vol 1 (Mumbai: International Institute for Population Sciences).National Sample Survey Organisation (1998): “Morbi-dity and Treatment of Ailments”, NSS Fifty-second Round, July 1995-June 1996, Government of India.Srinivasan, K and S K Mohanty (2004): “Health Care Utilisation by Source and Levels of Deprivation in Major States of India: Findings from NFHS-2”, Demography India, Vol 3, No 2, pp 107-26.Srinivasan, K, Chander Shekhar and P Arokiasamy (2007): “Reviewing Reproductive and Child Health Programmes in India”, Economic & Political Weekly, Vol XLII, Nos 27 and 28, pp 2931-39.