Spatial patterns of child health inequalities in the province of Punjab , Pakistan The advantages of GIS in data analysis

Spatial patterns of child health inequalities in the province of Punjab, Pakistan. The advantages of GIS in data analysis In the overall context of the Millennium Development Goals, Pakistan recently embarked on an effort to collect data regarding crucial issues of human development, such as health, poverty, and education. For their effective use in policy formulation, however, governments need to find ways of making the data easily accessible. It is argued that Geographical Information Systems (GIS) can play a major role here. The data compiled by the Federal Bureau of Statistics at district level are used in this article to develop maps on child health in the province of Punjab. Correlation allows identification of major determinants of the health status of children. For Punjab, these appeared to be mother literacy, vaccination coverage, and the incidence of poverty. Interpolation (here using Inverse Distant Weighting) and variable aggregation allow for easy identification of areas of concern at sub-district level. For Punjab, the areas most affected by negative health conditions are in the southwest. It is argued that health disparities can be addressed more effectively with the help of the presented methods because they help to close the information gap in the health system and contribute toward more precise allocation of funds and infrastructure.

Spatial patterns of child health inequalities in the province of Punjab, Pakistan  The advantages of GIS in data analysis Kashif Ur-Rehman, Basel, Anna Zimmer, Bonn 1 Introduction Poverty is a multidimensional issue which trans¬ lates broadly into impairment of human well-being.
The current understanding of poverty is accordingly broad, encompassing classical concerns like food, shel¬ ter, health care, and education as well as immaterial aspects, such as freedom of speech, emancipation of women, security, and empowerment World Bank 2000).
Among other deprivations, poverty in a complex way is a cause of high child mortality rates Food and Agriculture Organization of the United Nations -FAO 2002).Child health status reflects the degree of human development of the entire population as «no indicator captures the divergence in human develop¬ ment opportunity more powerfully than child mortality» United Nations Organisation -UNO 2005: 4).
Child mortality is thus a long-established measure not only of child health, but also of the well-being of a society Canadian Institute for Health Informa¬ tion 2005).
In the overall context of theMillennium Development Goals, Pakistan recently embarked on an effort to collect data regarding crucial issues of human devel¬ opment.The surveys conducted under the Federal Bureau of Statistics 2001-2009) aimed at informing poverty reduction strategies as well as development plans.
Despite the high relevance of this indicator for human development, studies about child health in Pakistan are currently limited.One reason for this is the lack of data at fine resolution below district level) which would allow identification of differentiated patterns of child health inequalities over space.This shortcoming can be bypassed with the help of Geographical Infor¬ mation Systems GIS) Epprecht & Heinimann 2004).With GIS it is possible to map incidences of diseases, thereby contributing toward better planning and loca¬ tion of health infrastructure Schweikart & Kiste¬  mann 2001).Butsch & Sakdapolrak forthcoming) identify public health service providers' lack of ade¬ quate information as one of the main bottlenecks in securing access to health care, especially for the poor.
Thefollowing article therefore argues that child health inequalities between different districts of Punjab can be addressed more effectively with the help of the methods presented.
Child health status shows highly differentiated pat¬ terns over space.The first objective of this study, therefore, is to visualize geographical patterns of child health in one particular region in Pakistan:   Punjab.Considering that health status is related to different sets of socio-economic determinants, be these directly such as nutrition, vaccination, sani¬ tation, and sources of drinking water) or indirectly maternal literacy and economic poverty), the second objective is to present spatial data of these factors.
It was hoped that the degree of correlation between these sets of determinants and the health status of children would become clearer.The third objective is to create a map of child health status at below district level resolution using the aggregate value of the most important determinants of health.It was assumed that this map would highlight the hot spots of poor health in the region.
2 Data and methodological approach 2.1 Data To assess the health status and the relationships between health and other socio-economic factors, this study uses data from four surveys carried through by the Federal Bureau of Statistics Office of Pakistan: 1) Pakistan Social and Living Standards Measure¬ ment PSLM) Survey 2004-05: In this survey, infor¬ mation was collected on education, health, house¬ hold assets/amenities, and vaccination coverage.
2) Household Integrated Economic Survey HIES) number of children who die before turning one per 1000 children.U5MR reflects the number of children who die before their fifth birthday per 1000 children.
Additionally, the number of children under five who are malnourished has been included in this study.This reflects the strong influence that the nutritional status of children has on their health.The main source of data for IMR, U5MR and malnourished children under five is the Multiple Indicator Cluster Survey of Punjab.Three types of data have been used for esti¬ mating IMR and U5MR in this survey: 1) five year intervals for women aged 15-49, 2) the mean number of children ever born MCEB) and 3) the proportion of child deaths to total female population at respec¬ tive ages.
Regarding representation of poverty, two important steps were necessary: 1) estimation of the poverty line and 2) estimation of percentage of population below this line.There are many approaches towards establishing the poverty line Jamal 2007; Moser 2006; World Bank 2009).The poverty line used in this study is the officialpoverty line of Pakistan.Here, households are classified as poor if they do not have sufficient incometo support a consumption pattern of 2350 calories per adult equivalent per day and cover minimum non-food requirements.The main source of data for calculating the incidence of poverty is the Household Integrated Economic Survey HIES) [2004][2005].The available data does not give insight into the degree or severity of poverty Coudouel et   al. 2002: 34ff.).It is also acknowledged that the mul¬ tiple dimensions of poverty cannot be suitably repre¬ sented by figures reflecting relative population num¬ bers under an income-based poverty line Kanbur & Squire 1999: 23).The basic data for the other main socio-economic determinants of child health, like percentages of diarrhea-affected children, vaccina¬ tion coverage, adequate sanitation, and accessibility to improved sources of drinking water have been adapted and modified from Pakistan Living Stand¬ ard Measurement PSLM) 2004-05.Unfortunately, the data collected by PSLM does not follow inter¬ national procedures and definitions set by United Nations Children's Fund -UNICEF & World Health Organization -WHO 2008: 6 ff.).The term «adequate sanitation» here therefore refers to flush toilets and non-flush toilets, including closed and ventilated improved pit latrines and service/bucket latrines.Inadequate sanitation refers to households with no toilets.The term «improved sources of drink¬ ing water » includes households with access to taps, hand or motor pumps and dug wells.No distinction is made here between covered and uncovered wells, an aspect which would have been worth paying attention to United Nations Children's Fund -UNICEF & World Health Organization -WHO 2008: 22).
Definite figures on the literacy rate of married women are not available in PSLM 2004/05.Maternal literacy has therefore been estimated using available data in PSML 2004/05 on literacy of women 15 years and older.
The sample size of the PSLM consists of 76'520 house¬ holds and is considered sufficiently large enough to produce reliable estimates at district level in all prov¬ inces of Pakistan.Indicators are presented disaggre¬ gated by province and district.This is the first time that a district level survey has been undertaken for the entire country.A detailed overview of the meth¬ odological approaches used for the socio-economic variables selected here are presented in the Pakistan Living Standard Measurement PSLM) 2004-2005, annexure-II.

Methodological approach
Methodically, this study was organized in three steps: • In a first step, available related statistical data from different surveys were processed and geo-refer¬ enced.This resulted in a statistical and GIS data¬ base which can be modified and used for further research and for other maps on child health status and related factors.
• In a second step, interrelations between the vari¬ ables were identified and visualized using the geo¬ statistical database.
• In a third step, all related variables with strong cor¬ relation to health status were aggregated into one variable.The result of this step is a map at sub-dis¬ trict level.
To develop a statistical and GIS database for this study, the variables needed to be processed to make them suitable for spatial analysis.All indicators were weighted to allow ranking amongst districts.The district with the most negative values ranked at the bottom of the health status list, while the district with the most positive values came out tops.During the course of the study, all districts in the province of Punjab were geo-referenced and all statistical data was translated into GIS form.This led to a geo database suitable for spatial analysis Fig. 1).
All maps presented herein were derived from this geo database.GIS based maps on each variable of health status and on each socio-economic variable with strong relation to the aggregated health status were generated inorder to keep the relation between health status and its determinants transparent.
By using a linear regression model, correlations were calculated between all indicators.These indicators were divided in two main categories: health status and other socio-economic indicators: Fig. 1: Methodological steps involved in generation of suitable data for GIS maps on child health inequali¬ ties Schritte der Datenverarbeitung in der Erstellung von GIS-Karten zum Gesundheitsstatus von Kindern Etapes méthodologiques conduisant à la production de cartes SIG sur les inégalités de santé infantile Geoprocessing by the authors 1) As discussed above, health status reflects the com¬ bination of IMR, U5MR and percentage of chil¬ dren under five who are underweight.A very close correlation could be found between these vari¬ ables of child health.Due to the close correlation, these three basic indicators were merged to one indicator, health status, allowing data reduction.
The ranking of the districts according to health status thus reflects the sum of all three indicators.The district with the lowest total sum ranks at the bottom of the table while the district withthe high¬ est sum tops the ranking.
2) All other socio-economic indicators like poverty, diarrhea, vaccination coverage, usage of Oral Rehydration Salts ORS) against dehydration, adequate sanitation, accessibility to improved sources of drinking water, and mother literacy were not aggregated in order to keep their relation to health status transparent.After calculation of correlations, all variables with a strong relation r higher than 0.5) to health status were aggregated using the same method as mentioned above.Vari¬ ables with weak correlations were omitted in the further analysis.
Finally, the data on health status and on socio-eco¬ nomic determinants were aggregated together.On the base of the new aggregated indicator a map at fine resolution see Figure 5) was developed using ArcGIS Geostatistical Analyst Arc Gis2001).Values forareas not recorded were estimated using the known values of surrounding areas.For this purpose, it was assumed that known values closest to those values being esti¬ mated have more influence on the results than those further away.Consequently, the points the distance measurements were taken from the district centroid points) closer to the location of an estimated value are weighted more than those further away.Gener¬ ally speaking, there are two significant drawbacks to using spatial regression models for calculating the spatial correlation: 1) the calculated spatial informa¬ tion between two points does not actually reflect the reality, and 2) due to spatial dependency and auto¬ correlation in the data, any error or misinformation in input data at one point has an influence on results of another point.For this reason, GIS Geostatistical Ana¬ lyst offers six methods for interpolating spatial data.
For this analysis, Inverse Distance Weighting IDW) was chosen to support classification of areas of great¬ est concern below administrative boundaries.are unacceptably high.In the case of Punjab, the IMR lies at 77 per 1000 live births and U5MR at 112 per 1000 live births.Thus, IMR and U5MR in Punjab are higher thanthe national average.Within Punjab, the districts of Rajanpur and Bhakhar 97), Muzaffar Garh 95), and Lodhran 93) are particularlyhard hit concerning IMR.
The latter three districts also have the highest U5MR.
The southwest of Punjab therefore appears to be a major area of concern in terms of health deprivations see Fig. 2).
As mentioned above, child health is strongly affected by nutritional status Hill et al. 1999).Despite high IMR and U5MR, underweight prevalence in Pakistan is lower than in the other large South Asian countries Bangladesh, India, Nepal).Itis higher only in compar¬ ison with Bhutan and Sri Lanka.Still, considering the national average of 38%, it is an aspect which should not be ignored.Punjab is slightly better off than the national average with 34% of children underweight.At district level, highest rates for underweight are found in the southwest with Bhakhar at 47%, Rajanpur at 47%, Muzaffar Garh at 46%, D.G. Khan at 45%, and Lodhran at 45% see Fig. 2).These results on spatial distribution of underweight thus support other results that the southwest of Punjab is particularly affected by health inequalities see chapter 3.2).

Determinants of child health
A good health status is the complex outcome of several structural factors, i.e. individual and household-related socio-economic determinants, as well as agency related factors, i.e. individual strategies to maintain health and cope with ill-health Bohle & Sakdapolrak 2009; Nestle et al. 2007).Leaving aside agency related fac¬ tors, which are best investigated with qualitative meth¬ ods, the following sections describe spatial patterns in structural factors leading to unsatisfactory health outcomes.While «maternal literacy» reflects the edu¬ cation system, and «poverty» refers to economic fac¬ tors, «water and sanitation» as well as «diarrhea» char¬ acterize the physical and infrastructural environment.«ORS therapy» and vaccination coverage are to be read as indicators of the quality of the health system.The national average for maternal literacy in Pakistan is 36%.For Punjab, the average is 40% and thus higher than in other provinces of Pakistan.Within Punjab, rates are worst in the northwest, with 18% in Lodhran, 20% in Mazaffar Garh, 25% in Rajan Pur, 26% in Bhakhar, and 26% in D.G. Khan.
Maternal education contributes to the reduction of child mortality in many ways: educated mothers, for example, tend to make use of prenatal care, get their children vaccinated see 3.1.5)and are knowledgeable about better child care practices Bhuiyat & Streat¬ field 1991; Caldwell 1986; Hobcraft 1993; Hos¬ seinpoor 2005; Tulasidhar 1993).The correlation of female literacy withhealth in thisstudy is stronger than for all the other factors mentioned above r 0.912).This strong correlation indicates that female literacy should be given maximum attention in state interven¬ tion with regard to child health.As it is assumed that educated women will pass their knowledge on to the next generation and are likely to give high priority to the education of their daughters, the education of women may be seen as a strong contribution towards forwarding development.

Poverty
The poverty line see section 2.1) served for the esti¬ mation of poverty incidence.About 35% of the popu¬ lation of Pakistan lives below the poverty line.In the province of Punjab, the rate is slightly lower, with 32% living below the poverty line.However, within Punjab itself, striking differences exist: while the districts of Muzaffar Garh 56%), Rajanpur 54%), D.G. Khan 51%), Lodhran 48%), and Rahim yar Khan 46%) show very high incidences of poverty, Rawalpindi 11%), Lahore 12%), Jehlum 12%), Gujrat 13%), and Sialkot 14%) are far better off.The poverty map see Fig. 3) helps to visualize these disparities, show¬ ing a clear prevalence of poverty in the southwest of Punjab.
Poverty can be clearly linked to health inequalities.In this study, there is a strong correlation between pov¬ erty and health status r 0.787).These results cor¬ respond with results from several studies, suggesting that there is a high correlation between poverty and child mortality Caldwell 1986; Gwatkin et al. 2000; Wagstaff 2002).
3.1.3Water and sanitation A better housing environment with safe drinking water und adequate sanitation is one of the major basic necessities for good health.Although infrastruc¬ ture needs to be properly maintained if health is to benefita fact overlooked most of the times when set¬ ting political agendas Zimmer 2009) -the creation of facilities is a first step in the right direction.For this reason, attention is paid here to water and sanitation coverage Fig. 4).La précarité des conditions d'évacuation des eaux usées Foto: A. Zimmer   Within Punjab, results show that access to improved sources of drinking water is lowest in the west and northwestern districts, namely Attock, Rawalpindi, and D.G. Khan, all of which have less than 80% of households with access to safe drinking water.These results correspond to diarrhea data, the incidence of diarrhea also being highest in the same districts.
Regarding access to adequate sanitation the prov¬ ince is severely underserviced except in the northeast; particularly in the west, access rates of under 36% prevail, with only an alarming 15-22% of households in Bhakhar recorded as having adequate sanitation facilities.
The connection between water, sanitation, and health has been established time and again.An estimated 80% of sicknesses in the world are water-related Black & Fawcett 2008: 71).However, in our study, the correlation between health and water supply is r 0.1 and health and diarrhea is r 0.586, which is well below expectations.
Possible reason for these results could be the differ¬ ent approaches used to determe sanitation categories between PSLM data and international studies and the lack of information about the quality of sanitation infrastructure.Infrastructure which is not properly maintained provides no benefit in terms of health out¬ comes.The results presented here on available water supply and sanitation does not reflect safe supply and disposal of water and waste water.Low correla¬ tions thus do not suggest that there is no relationship between health and water and sanitation.Rather, they highlight the need for inclusion of more meaningful data.
3.1.4Diarrhea and use of ORS therapy The nutritional status of a child is determined amongst other factors bythenumber of times it has had diarrhea Arif & Ibrahim 1998; Mahmood 2003).Furthermore, acute diarrhea is one of the leading causes of death among children under five, with almost 2 million vic¬ tims per year worldwide United Nations Children's Fund -UNICEF 2006).The main cause of death in these cases is dehydration.The percentage of chil¬ dren with dehydration caused by diarrhea in Pakistan in 2004-2005 is 91%.Of the country's four provinces, Punjab rates second best in this area with 90%.Within Punjab, the northwest and western districts have the highest percentages of diarrheal induced dehydration: Attock 95%), Rawalpindi 91%), Chakwal 88%), Mianwali 88%), and D.G. Khan 87%) see Fig. 5).As correlation of diarrhea with child health status does not indicate a very strong relation r 0.110), it has been omitted from further analysis.
Dehydration can be prevented by the use of Oral Rehydration Therapy Baltzar et al. 2002; Victora   et al. 2000).WHO and UNICEF recommend the use of Oral Rehydration Salts ORS) to treat dehydra¬ tion caused by diarrhea.In the case of Pakistan, the average percentage of children with diarrhea who received ORS is 78%.However, the percentage for Punjab is lowest in the country with just 41%.The spatial pattern of ORS coverage is not as clear as that of other health related parameters.Figures are highest in the southern districts of Multan, Rajan¬ pur both 57%), as well as Bhawalnagar 55%) and Rahim Yar Khan 51%), but northern districts like Sialkot 57%), Rawalpindi 50%) and Gujranwala 48%) fare almost equally well.In contrast, fig¬ ures are low in the neighboring northern districts of Jehlum, Kashur, Layyah, Khushab and Mianwali between 24% and 29%) see Fig. 3).Correlation between ORS therapy and health status is weak r 0.146).Identifying the reason for this weak cor¬ relation would need further research and is not part of this study.Consequently, this factor was also dis¬ carded from further analysis.m albtherliokfe c hadiledqreunatpee sra1n0it0a0ti on, Fig. 5: Correlations between IMR number of children per 1000 children who die before their first birthday), U5MR number of children per 1000 children who die before their fifth birthday),malnourishment percentage of malnourished children under five), adequate sanitation, accessibility to improved sources of drinking water, poverty, mother literacy, ORS Oral Rehydration Solution) therapies, diarrhea and vaccination coverage Korrelation zwischen Gesundheitsstatus von Kindern und IMR Anteil der Kinder pro 1000 Lebendgeburten, die vorErreichen des 1. Lebensjahres sterben), U5MR Anteilder Kinder pro 1000 Lebendgeburten, die vor Erreichen des 5. Lebensjahres sterben), Unterernährung Prozentsatz mangelernährter Kinder unter fünf Jahren), angemes¬ sener sanitärer Versorgung, Zugang zu verbesserten Trinkwasserquellen, Armut, Bildung der Mütter, ORS Oral Rehydration Solution)-Therapien, Durchfallerkrankungen und Anteil geimpfter Kinder Corrélations entre les variables IMR nombre d'enfants décédés avant leur premier anniversaire en pour mille), U5MR nombre d'enfants décédés avant leur cinquième anniversaire en pour mille), malnutrition pourcentage d'enfants de moins de cinq ans mal nourris), installations sanitaires adéquates, accessibilité à des sources d'eau potable améliorées, pauvreté, éducation de la mère, thérapies ORS Oral Rehydration Solution), diarrhées et couverture vaccinale WHOsuggests children should be vaccinated in order to significantly reduce the riskof death inthe first five years of life United Nations Children's Fund -UNICEF 2007).The recommended vaccinations are BCG against tuberculosis), three doses of DPT against diphtheria, pertussis and tetanus), three doses of polio vaccine and a measles vaccination.A child is considered fully vacci¬ nated after receiving BCG, DTP1-3 and Polio 1-3.About 49% children in Pakistan are fully vaccinated.
The percentages for Punjab are by far the highest in the country 66%).This may be connected to the high maternal literacy rates in the province see 3.1.1).Yet, while values in the northern districts are high, criti¬ cal rates are found in the southern districts of Rajan¬ pur 19%), D.G. Khan 37%), Muzaffar Garh 41%), Rahim Yar Khan 43%), and Lodhran 56%).Correla¬ tion of vaccination coverage and health status is esti¬ mated at r 0.736.For this reason, vaccination status has been included in the further analysis.

Aggregated map
As may be seen in section 3.1, the data on maternal literacy, poverty and vaccination coverage proved to be the most relevant for determining the health status of children in the province of Punjab.As described in section 2.2, each of these parameters was aggregated in order to establish areas of greatest risk see Fig. 6).
The resulting map provides spatial information on child health inequality at district level and within each district across the province.
Areas of greatest concern in terms of health outcomes ± Areas of greatest concern Fig. 6: Areas of greatest concern regarding child health status in the province of Punjab, Pakistan Problematischste Regionen in Bezug auf den Gesundheitsstatus von Kindern in der Provinz Punjab, Pakistan Régions dans lesquelles l'état de santé infantile est problématique dans la province du Punjab au Pakistan The presented data analysis allows a clear picture to be made of the status of child health in the province of Punjab.In particular, child health in southwest Punjab appears to be most at risk and requires special politi¬ cal attention.
Correlation analysis of socio-economic factors shows very close correlations between health and poverty, mother literacy, and vaccination coverage.By combining visualization techniques with statistical methods, under¬ lying causes of health decline could be revealed.In the presented case study, female literacy was shown to be the single most important factor in enhancing child health, thus indicating options forpossible policy interventions.
The aggregated map demonstrates how GIS can support policy makers to go beyond administra¬ tive boundaries in identifying ill-health hot spots.
Although fine resolution data are not yet available, the approach makes patterns of socio-economic dep¬ rivation more visible, an aspect which may be useful for decision makers seeking to implement alleviation programs with more precise allocation of efforts and funds.
Further, the study reveals several other benefits of a GIS approach to analyzing health issues: Firstly, the maps of each variable selected for this analysis see Fig. 2 and Fig. 3) make existing spatial patterns of inequalities clearly visible and provide informa¬ tion on background factors of child health inequali- The data processed for these maps can also be used and modified for cognitive and monitoring pur¬ poses.In this way, a GIS approach allows easy com¬ prehension of data and contributes towards closing the information gap in the health system Butsch & Sakdapolrak forthcoming).When combined with statistical correlation it can help overcome narrow approaches towards issues such as health by allow¬ ing complex multiple dimensions to be included in the analysis.
Secondly, a GIS approach results in a more precise identification of areas of concern and allows for better planning of health infrastructure Schwei¬ kart & Kistemann 2001).This could have an impact on health governance as government officials at dis¬ trict level and below would have access to necessary arguments for fund allocation.In this way, decen¬ tralized governance structures can be strengthened vis-à-vis higher levels of bureaucracy and govern¬ ment.
If the maps are made accessible to the citizens, they offer advantages beyond an «informing» of the public sector Berntzen et al. 2005).Even if accessibil¬ ity through the internet is restricted to certain social classes and groups McCall 2003), the visual aspect of mapping statistical data alone allows even poorly edu¬ cated people to apprehend inequalities and monitor the outcomes of political initiatives more easily.GIS can therefore promote transparency if maps become devices of political activism and are used to hold the government accountable.
UGECViewpoints 2009/09: 9-12; http://www.ugec.org/tiki-index.php?page=ReportsNewsletters 01.11.2009.Abstract: Spatial patterns of child health inequalities in the province of Punjab, Pakistan.The advantages of GIS in data analysis In the overall context of the Millennium Development Goals, Pakistan recently embarked on an effort to col¬ lect data regarding crucial issues of human develop¬ ment, such as health, poverty, and education.For their effective use in policy formulation, however, govern¬ ments need to find ways of making the data easily accessible.It is argued that Geographical Information Systems GIS) can play a major role here.The data compiled by the Federal Bureau of Statistics at district level are used in this article to develop maps on child health in the province of Punjab.Correlation allows identification of major determinants of the health status of children.For Punjab, these appeared to be mother literacy, vaccination coverage, and the inci¬ dence of poverty.Interpolation here using Inverse Distant Weighting) and variable aggregation allow for easy identification of areas of concern at sub-district level.For Punjab, the areas most affected by negative health conditions are in the southwest.It is argued that health disparities can be addressed more effec¬ tively with the help of the presented methods because they help to close the information gap in the health system and contribute toward more precise allocation of funds and infrastructure.

Fig. 1 :Fig. 2 :
Fig. 1:Maps on districts locations and status of child health at districts level in the province of Punjab Data source: Pakistan Demographic Survey PDS) 2005 Multi Indicator Cluster Survey MIC) Punjab 2004 Geoprocessing: By the authors Data source: Government of Pakistan 2005a), Pakistan Social and Living Standards Measurement Survey PSLM) 2004/05; Government of Pakistan 2005b), Pakistan Demographic Survey PDS) 2005; Government of Punjab 2005), Multi Indicator Cluster Survey MICS) Punjab 2004/05 3.1.5Vaccination coverage Data source: Government of Pakistan 2005a), Pakistan Social and Living Standards Measurement Survey PSLM) 2004/05; Government of Pakistan 2005b), Pakistan Demographic Survey PDS)2005;Government of Punjab 2005), Multi Indicator Cluster Survey MICS) Punjab 2004/05; Geoprocessing by the authors 4 Conclusion Although figures for child health are worse in Paki¬ stan than in any other South Asian country Gwatkin et al. 2000), studies on infant and child mortality in this region are few among othersArif & Ibrahim 1998;