Users Online: 742 Home Print this page Email this page Small font sizeDefault font sizeIncrease font size


Home About us Editorial board Search Ahead of print Current issue Archives Submit article Instructions Subscribe Contacts Login 

 Table of Contents  
Year : 2021  |  Volume : 65  |  Issue : 4  |  Page : 362-368  

A spatiotemporal geographic information system-based assessment of human immunodeficiency virus/acquired immune deficiency syndrome distribution in Manipur, India

1 Associate Professor, National Institute of Technology, Imphal, Manipur, India
2 Research Scholar, National Institute of Technology, Imphal, Manipur, India
3 Project Assistant, National Institute of Technology, Imphal, Manipur, India

Date of Submission20-Dec-2020
Date of Decision04-Jun-2021
Date of Acceptance09-Nov-2021
Date of Web Publication29-Dec-2021

Correspondence Address:
Vicky Anand
Department of Civil Engineering, National Institute of Technology, Imphal, Manipur
Login to access the Email id

Source of Support: None, Conflict of Interest: None

DOI: 10.4103/ijph.IJPH_1308_20

Rights and Permissions

Background: Geographic information system (GIS) is a versatile tool that assists in health education, planning, research, monitoring, and evaluation of programs related to health. One of the epidemics which threaten the overall human welfare is human immunodeficiency virus/acquired immune deficiency syndrome (HIV/AIDS). In Manipur, the cases of HIV/AIDS have been reported at significant level. Objective: The study aimed to detect the hotspot regions of HIV/AIDS prevalence in Manipur and to identify the significant factors which influence the HIV prevalence. Methods: This study evaluates the spatial variations of HIV/AIDS prevalence in the state of Manipur, India, from 2011 to 2018. In this study, Getis-Ord Gi* statistic was used to detect the HIV/AIDS prevalent regions. The ordinary least square (OLS) spatial statistics embedded in the ArcGIS were employed for exploring the spatial relation between HIV/AIDS occurrence and the predictors. Results: It was observed from the hotspot results that Churachandpur, Ukhrul, and Thoubal are the blocks where HIV/AIDS is more prevalent. Six factors associated with the prevalence of HIV/AIDS were found to be significant. The most obvious factor influencing HIV in the region is illiteracy. The constructed OLS model has the highest value of adjusted R2 statistic equals to 0.67 and the lowest value of the Akaike Information Criterion statistic equals to 474.55. Conclusion: The use of hotspot analysis, regression analysis, spatial autocorrelation, and GIS can aid health planners in properly assessing and identifying spatial prevalence of diseases among the masses to better guide evidence-based health planning decisions.

Keywords: Epidemiology, geographic information system, hotspot, human immunodeficiency virus, regression

How to cite this article:
Oinam B, Anand V, Kajal R K. A spatiotemporal geographic information system-based assessment of human immunodeficiency virus/acquired immune deficiency syndrome distribution in Manipur, India. Indian J Public Health 2021;65:362-8

How to cite this URL:
Oinam B, Anand V, Kajal R K. A spatiotemporal geographic information system-based assessment of human immunodeficiency virus/acquired immune deficiency syndrome distribution in Manipur, India. Indian J Public Health [serial online] 2021 [cited 2022 Jan 22];65:362-8. Available from:

   Introduction Top

Human immunodeficiency virus (HIV) epidemic represents high risk in public health problem in society. It currently spreads in the world at the rate of one new infection every second. Acquired immune deficiency syndrome (AIDS) is an infectious disease caused by the HIV. There are two variants of the HIV virus HIV-1 and HIV-2, both of which ultimately cause AIDS. According to the HIV Estimation 2017 report (National AIDS Control Organization [NACO]), age between 15 and 49 years HIV prevalence is at 0.22% (0.16%–0.30%) and adult HIV prevalence is estimated at 0.25% (0.18–0.34) among males and at 0.19% (0.14–0.25) among females.[1]

The overall prevalence of HIV in the Indian population is approximately 0.3%, which is greater than the world average of 0.2%.[2] The epidemic in the country is uneven with the second highest estimated prevalence of 1.43% (1.17–1.75) in the North-Eastern state of Manipur.[1] Five other states, Mizoram, Karnataka, Telangana, Nagaland, and Andhra Pradesh, have shown prevalence of more than double that of the National average. Rising trends in adult HIV prevalence have been observed in some of the hitherto relatively low-prevalence states/union territories like Assam, Chandigarh, Delhi, Jharkhand, Punjab, Tripura, and Uttarakhand.

According to the State Development Report of Manipur-2006, HIV is a major public health problem in Manipur, which has the second highest rate of prevalence in the country. The NACO classifies Manipur as “high-prevalence” with 5% of the high-risk groups and more than 1% of the women in antenatal clinics (ANC) testing HIV positive. HIV infection is declining in Manipur by <10% in last 7 years.[1]

GIS is emerging trend to study health problems, analyze, and help prevent disease. GIS can study disease affected region, associated risk factors, service availability, prevention, and planning of health resources. In GIS, spatial analysis method is used to study the data for disease risk factor, epidemic trends over space and time, and hotspot region.[3],[4] Spatial analysis tools have been used in understanding the health problem associated to HIV/AIDS related studies in sub-Saharan Africa.[5],[6] The spatial statistics tools in GIS contains two major regression techniques-Geographically weighted regression (GWR) and ordinary least square (OLS). GWR is a method that allows in studying how a given phenomenon varies spatially in a particular area.[7] As the GWR takes into account local parameters, it can be considered as a spatial extension of multiple linear regressions. This method takes into account spatial variability of HIV prevalence factors. The method gains increasing attention and has been used in studies of economic, social, and environmental phenomena that show spatial variability. OLS regression is one of the major techniques used to analyze data and forms the basis of many other techniques (for example generalized linear models and analysis of variance).[8] OLS regression is particularly powerful as it relatively easy to also check the model assumption such as linearity, constant variance and the effect of outliers using simple graphical methods.[9] The main objectives of this study includes (1) hotspot analysis of HIV/AIDS affected areas, (2) identification of statistically significant HIV prevalence factors by the formulation of spatial regression model (OLS), (3) assessment of OLS regression model for HIV prevalence.

   Materials and Methods Top

Study area and design

Manipur lies between latitudes 23° 50' N to 25° 41'N and longitudes 92° 58'E to 94° 45' E with an area of 22,327 km2 [Figure 1]. Majority of the population are concentrated in valley region which is about 10% of the total geographical area and the remaining population are settled in the hilly region which is almost 90% of the geographical area of the state [Figure 1].
Figure 1: Location of Manipur with block boundary, road network, and population concentration centers.

Click here to view

Geographically, this region comes under complex terrain (i.e., hilly areas rising up between 300 m to 2994 m from the msl) and is characterized by poor infrastructure, economic underdevelopment, and disturbed area status.[10] According to Census of India, 2011, Manipur has a population of 28,55,794 in which 57.2% lives in Imphal Valley (10% of total area) and the others 42.8% resides in the hilly region (90% of the total area). There are 9 districts in Manipur in which 4 districts are located in valley region while 5 districts make up the hilly region.[11]

In this study, health data for 7 years (2011–2018) provided by State Health Society, Manipur under National Rural Health Mission, Manipur AIDS Control Society, Health Directorate, Manipur and National Health System Resource Centre were used. The health dataset used in this study was acquired from the above-mentioned sources with a proper permission for the use in this scientific study. Various geospatial parameters like population distribution, existing health center, and health data were evaluated using ArcGIS. To analyze the hotspot zone of prevalence of HIV among male, female ANC, and female non-ANC spatial statistics tools were used. 'This research does not involve any human subjects at individual level and only secondary data on health facilities, diseases and GPS locations were utilized for mapping purpose.'-Not required: Exemption from IEC Review (IEC-NIT MANIPUR).

Study tools/techniques, statistical methods

Hotspot analysis

Hotspot is defined as a condition indicating some form of clustering in a spatial distribution.[12] Hotspot analysis tool identifies the features with high and low values. It is calculated by Getis-Ord Gi* statistic, developed by Getis and Ord, to analyze the cluster structure of high or low value of the feature.[13],[14] A feature with a high value may not be significant hotspot, when a high value is surrounded by other high values and if it is statistically significant it becomes a hotspot. The formula is given as:[15]

Where xj is the attribute value for the feature j, wi, j is the spatial weight between feature i and j, n is equal to the number of features,

The Gi* statistics is a z-score. Large and positive value of z-score represents intense clustering of high values which signifies it is a hotspot, and small and negative value of z-score represents intense clustering of low values which signifies it is a cold spot.

The prevalence measures the proportion of individuals in a population with a specific disease at a certain point time (T). Prevalence is measured at a specific point in time (T).[16]

Regression analysis

Spatial modeling of HIV/AIDS is a demanding problem due to the complexity of factors behind this process and the consequential spatial heterogeneity of variables. GIS spatial data processing techniques can be used to assist the processing of large and numerous data sets required for prevalence modeling and GIS spatial data analysis methods can also be applied to develop prevalence models and assist studies of this phenomenon. The OLSs is a method which assumes that relationships are consistent geographically. The general equation of OLS regression model may be written as[17]

Where, yi is the dependent variable measured at location i, xi are the independent variables, βi are the model coefficients, ɛi is the error. The theory behind spatial regression has been described by (Brundson et al. 1996; Fotheringham et al. 2002) and its implementation by Charlton and Fotheringham (2009).

Several variable combinations was applied to identify the associated relationship between the dependent and the independent variable, the best fit was found with, total number of workers (refers to all persons engaged in “work” defined as participation in any economically productive activity with or without compensation, wages or profit), total number of main worker (refers to those workers who had worked for the major part of the year were termed as main workers. Major part of the year meant 6 months [183 days] or more), number of main worker-cultivator (a person was considered working as cultivator if he or she was engaged either as employer, single worker or family worker in the cultivation of land owned or held from Government or from private person or institution for payment in money, or in kind or on the basis of sharing of crops), number of agriculture labor (refers to a person who works on another person's land for wages in money, kind or share of crop), main worker others (all workers i.e., those who had been engaged in some economic activity during the year preceding numeration and who were not cultivators or agricultural laborers or household industry workers were termed as “other workers), and number of illiterate persons (a person who can only read but cannot write, is not literate), as the variable combination.

The research methodology consisted of the following steps:

  1. Preparation of spatial representations of HIV/AIDS factors-dependent and independent variables for spatial regression analysis
  2. Exploratory analysis. Analysis of the general spatial regression model. Identification and selection of variables providing the best explanation for dependent variable (HIV prevalence)
  3. Construction and analysis of regression model with OLSs method.

   Results Top

Spatial and temporal trends in human immunodeficiency virus prevalence at block level

In this study HIV/AIDS prevalence has been categorized in three different categories i.e., male, female ANC and female Non-ANC. Hotspot analysis was performed at block level for seven different time periods. High hotspot (99% confidence) zones and medium hotspot (95% confidence) zones among the males for the year 2012-13 is shown in [Figure 2]a. It was observed that HIV among males is more prevalent in Churachandpur (2011–2017), Ukhrul (2012–2017), Porompat (2013), Lamphelpat (2017), Bishnupur (2012–2014), and Thoubal (2012, 2017) blocks. In these years no pattern was observed in the hotspot zones rather it was dispersed except for the year 2012–13 and 2013–14 clustering was observed in the blocks which lies in the valley regions namely Churachandpur, Bishnupur, and Thoubal.
Figure 2: Hotspot distribution of (a) number of human immunodeficiency virus positive on male (2012–2013), (b) number of human immunodeficiency virus positive on female with antenatal clinics (2017–2018), (c) number of human immunodeficiency virus positive on female nonantenatal clinics (2016–2017), and (d) total number of human immunodeficiency virus positive (2017–2018) at block level.

Click here to view

Similarly, in the case of ANC female high hotspot (99% confidence) zones and medium hotspot (95% confidence) zones were detected for different years. The hotspot result of ANC female for the year 2017-18 is shown in [Figure 2]b. It was observed that HIV among females ANC in the year (2011–12) was more prevalent in Churachandpur, Ukhrul; Ukhrul, Chandel in (2012–13); Mao Maram in (2013–14); Churachandpur, Wangoi, Sawombung in (2014–15); Churanchandpur, Thoubal in (2015–16); Mao Maram, Thoubal in (2016–17) and Chandel, Chakpikarong, Mao Maram, Paomata in (2017–18).

In the case of non-ANC female HIV was found to be more prevalent in Churachandpur in 2011–12; Churachandpur, Ukhrul in 2012–13; Mao Maram, Ukhrul, Porompat in 2013–2014; Churachandpur in 2014–15; Churachandpur, Thoubal, Ukhrul in 2015–16 and 2016–17; Churachandpur, Ukhrul in 2017–18. Hotspot zones for female non-ANC for the year (2016–17) have been represented in [Figure 2]c.

HIV/AIDS maps clearly show that there is no clear pattern or clustering rather the hotspots areas is dispersed at the block level. It can be analyzed that Churachandpur, Ukhrul and Thoubal are areas where the HIV/AIDS is more prevalent as compared to the other blocks in the state. The overall hotspot zones including male, female ANC, and female non-ANC for the year 2017-18 have been represented [Figure 2]d.

HIV among females ANC in the year (2011–12) was more prevalent in Churachandpur; Churachandpur, Ukhrul and Bishnupur in (2012–13); Mao Maram in (2013–14); Churachandpur in (2014–15); Churanchandpur, Thoubal and Ukhrul in (2015–16); Churachanpur, Thoubal in (2016–17) and Chandel, Chakpikarong, Mao Maram, Ukhrul, Lamphelpat and Thoubal in (2017–18). In overall distribution of total HIV positive cases among male, female ANC, female non-ANC for different years it was observed that overall hotspot zones at the block level is dispersed, except in the year 2017-18 clustering pattern was seen in the northern blocks located in hilly terrains i.e.; Mao Maram and Ukhrul and in the South-Eastern blocks located in hilly terrains i.e.; Chandel, Chakpikarong and Thoubal. The total number of positive HIV/AIDS cases including male, female ANC, female non-ANC at block level has been represented in [Figure 3].
Figure 3: Total number of human immunodeficiency virus positive at block level.

Click here to view

The spatial distribution of HIV is always correlated with socioeconomic factors.[18] It would be helpful to investigate the underlying socioeconomic causes of high prevalence area or the hotspot identified area in this study. In this study the number of HIV patients was classified by year and block as well by sociodemographic characteristics such as gender. In order to identify the factors which influence the HIV prevalence in the hotspot regions, regression analysis was done based on epidemiological data available for the study area.

Regression analysis

Various combinations of explanatory independent variables have been tested. Based on the results of exploratory analysis including tests for redundancy (multi-collinearity), bias (nonlinear relationships and data outliers) and assessment of all possible combination of OLS models, the one with: number of illiterate persons, total number of workers, total number of main worker, number of main worker-cultivator, number of agriculture labor and main worker others as explanatory variables has been found to be the best fitting. The selected variables had significant robust probability statistics, but the variable inflation factor (VIF) of the variables was found to be on the higher side. The OLS model constructed with the six selected variables had the highest value of adjusted R2 statistic equal to 0.67 and the lowest value of the Akaike information criterion (AIC) statistic equal to 474.55. The adjusted R2 and the AIC are statistics derived from the regression equation to quantify model performance.[19] The Chi-squared value 0.0078 of Koenker statistics is statistically significant. This indicates relationship between some or perhaps all the explanatory variables are heteroskedastic or nonstationary across the region. This may be because of some explanatory variable may be important with respect to reported HIV/AIDS cases in some blocks, but in some other blocks may demonstrate weak predictive capability. To further verify the pattern, spatial auto correlation statistics (Global Moran's I) was applied. The Global Moran's I result further indicates residuals have no statistically significant spatial autocorrelation. In this case, all empirical evidence point to the fact that the OLS residuals fit properly. The Global Moran's I statistics is shown in [Figure 4].
Figure 4: Global Moran's I spatial autocorrelation for human immunodeficiency virus/acquired immune deficiency syndrome.

Click here to view

In order to find the most significant or the redundant factor, different sets of variables were again tested among which number of persons illiterate, number of main work-cultivator and number of agriculture labor were found to be redundant with VIF values between 1.73 and 2.80 and coefficient 0.003–0.019. The VIF is a test of multi-collinearity among data, and its values >7.5 indicate that variables are redundant.

   Discussion Top

For the peoples in the developing countries HIV/AIDS still remains one of the most dangerous risks, even after the HIV/AIDS incidence is gradually decreasing worldwide.[20] Peoples infected or suffering with HIV/AIDS can potentially live same life as of a healthy person due to the advancement in the field of medicines in the past, but still there is no cure for the disease currently. Due to this the best way to control or fight with the spread of this disease is through preventing transmissions. HIV infection is influenced by so many factors one of which is illiteracy. Adequate knowledge of HIV/AIDS is required in order for the population to use appropriate measures in order to prevent the spread of HIV/AIDS. Education remains an important factor in providing the fundamentals for individuals to learn about diseases like HIV and understand its burdens.[21] Even in the past studies have been conducted in the various countries in South-East Asia, Sub-Saharan Africa regarding the HIV/AIDS related knowledge which suggests that lack of education or illiteracy is strongly related to the transmissions of HIV/AIDS.[22],[23]

In this study, illiteracy, main work-cultivator, and agriculture labors were found to be most significant variable in HIV prevalence. This study shows or indicates about the significant associations between the HIV/AIDS incidence and the literacy in the region. The factors of education and literacy have a direct effect on individual's knowledge of diseases like HIV/AIDS. The literatures suggests that those with higher levels of education tend to be more knowledgeable about HIV transmission and prevention and hence less vulnerable to HIV infection than those with lower levels of education.[24] Illiterate peoples could be vulnerable to HIV/AIDS as they might not have adequate information on the causes and transmission of the pandemic.[25] Adequate education programs are needed to teach individuals in these regions in order to effectively navigate through these risky environments and prevent further transmission of HIV/AIDS.

   Conclusion Top

This study demonstrates the use of spatial statistics and GIS can aid health planners in properly assessing and identifying spatial prevalence of diseases among the masses so as to better guide evidence-based health planning decisions. Hotspot analysis result indicates that the HIV/AIDS cases are mainly concentrated in the Churachandpur, Ukhrul, and Thoubal blocks of Manipur. Information on hotspots as depicted in this study can help to locate the regions vulnerable to HIV/AIDS and the potential risk factors, which in turn could aid in implementing targeted intervention programs in future. The results showed that proposed methods and tools could be beneficial for public health officers to visualize and understand the distribution and trends of diffusion pattern of diseases and to prepare warning and awareness to the masses. GIS can be used as the effective tool to monitor and manage HIV and other related activities. In order to do structure studies and interpretation of results for proper socioeconomic development at various levels, a proper understanding of epidemiological methods and principles are required.


The authors would like to acknowledge the Public Health Foundation of India (PHFI) for their funding, supports and constructive suggestion on this part of project outcome. The authors wish to thanks the Manipur State-National Health Mission, Directorate of Health Services, Government of Manipur and RIMS for sharing location information data on health facilities and statistical data on different types of diseases in different districts of Manipur. The authors wish to thanks Department of Community Medicine, JNIMS and State Medical officers posted at different health centers for their kind assistance during field trip and interaction program.

Financial support and sponsorship

Funded by Public Health Foundation of India (PHFI).

Conflicts of interest

There are no conflicts of interest.

   References Top

National AIDS Control Organization & ICMR-National Institute of Medical Statistics. HIV Estimations 2017: Technical Report. New Delhi: NACO, Ministry of Health and Family Welfare, Government of India; 2018.  Back to cited text no. 1
UNAIDS. Report on the Global AIDS Epidemic; 2014. Available from: < Report_2014>. [Last accessed on 2019 September 26].  Back to cited text no. 2
Lyseen AK, Nøhr C, Sørensen EM, Gudes O, Geraghty EM, Shaw NT, et al. A review and framework for categorizing current research and development in health related Geographical Information Systems (GIS) studies. Yearb Med Inform 2014;9:110-24.  Back to cited text no. 3
Kandwal R, Garg PK, Garg RD. Health GIS and HIV/AIDS studies: Perspective and retrospective. J Biomed Inform 2009;42:748-55.  Back to cited text no. 4
Tanser F, Bärnighausen T, Cooke GS, Newell ML. Localized spatial clustering of HIV infections in a widely disseminated rural South African epidemic. Int J Epidemiol 2009;38:1008-16.  Back to cited text no. 5
Kalipeni E, Zulu L. Using GIS to model and forecast HIV/AIDS rates in Africa, 1986–2010. Prof Geogr 2008;60:33-53.  Back to cited text no. 6
Fotheringham S, Brunsdon C, Charlton M. Geographically Weighted Regression: The Analysis of Spatially Varying Relationships. Hoboken: Wiley; 2002.  Back to cited text no. 7
Rutherford A. Introducing ANOVA and ANCOVA: A GLM Approach. London: Sage Publications; 2001.  Back to cited text no. 8
Mahaboob B, Bhumireddy V, Narayana C, Sankar J, Balasiddamuni P. A treatise on ordinary least squares estimation of parameters of linear model. Int J Eng Technol (UAE) 2018;7:518-22.  Back to cited text no. 9
Oinam B, Oinam J, Kajal R. A geospatial approach to assess health coverage and scaling-up of healthcare facilities. Curr Sci 2020;118:5.  Back to cited text no. 10
GoM. Approach Paper on New Land Use Policy Project of Manipur. Report by Planning Department; 2014. Available from: http:// Manipur%202014.pdf. [Last accessed on 2019 May 11].  Back to cited text no. 11
Osei FB, Duker AA. Spatial and demographic patterns of cholera in Ashanti region – Ghana. Int J Health Geogr 2008;7:44.  Back to cited text no. 12
Getis A, Ord JK. The analysis of spatial association by use of distance statistics. Geogr Anal 1992;24:189-207.  Back to cited text no. 13
Ord JK, Getis A. Local spatial autocorrelation statistics: Distributional issues and an application. Geogr Anal 1995;27:286-306.  Back to cited text no. 14
ESRI. (Gi* de Getis-Ord); 2018. Available from: es/pro-app/tool-reference/spatial-statistics/h-how-hot-spot-analysis-getis-ord-gi-spatial-stati.html. [Last accessed on 2019 December 20].  Back to cited text no. 15
Dekker S, Lee NC, Howard-Jones P, Jolles J. Neuromyths in education: Prevalence and predictors of misconceptions among teachers. Front Psychol 2012;3:429.  Back to cited text no. 16
Charlton M, Fotheringham S. Geographically Weighted Regression. White Paper. Maynooth: National Centre for Geocomputation, National University of Ireland Maynooth; 2009.  Back to cited text no. 17
Ezike-Dennis UN. The Spatial Distribution of HIV and AIDS in Gauteng: South Africa. Dissertation, University of South Africa, Ms. Science; 2007. p. 166.  Back to cited text no. 18
ESRI. ArcGIS desktop 10.2 help library. In: Spatial Regression. California: ESRI; 2015.  Back to cited text no. 19
World Health Organization. UNAIDS Global AIDS Report. Geneva: World Health Organization; 2016.  Back to cited text no. 20
Zimmerman EB, Woolf SH, Haley A. Understanding the relationship between education and health. In: Population Health: Behavioral and Social Science Insights. Rockville, MD: Agency for Healthcare Research and Quality; 2015.  Back to cited text no. 21
Fay H, Baral SD, Trapence G, Motimedi F, Umar E, Iipinge S, et al. Stigma, health care access, and HIV knowledge among men who have sex with men in Malawi, Namibia, and Botswana. AIDS Behav 2011;15:1088-97.  Back to cited text no. 22
Nachega JB, Lehman DA, Hlatshwayo D, Mothopeng R, Chaisson RE, Karstaedt AS. HIV/AIDS and antiretroviral treatment knowledge, attitudes, beliefs, and practices in HIV-infected adults in Soweto, South Africa. J Acquir Immune Defic Syndr 2005;38:196-201.  Back to cited text no. 23
David JK, Pant R, Allam RR, Priya VM, Aridoss S, Arumugam E. The relationship between educational attainment and HIV prevalence among pregnant women attending antenatal clinics in six states of India: Sentinel surveillance from 2010 to 2017. Indian J Public Health 2020;64:S15-21.  Back to cited text no. 24
Matthew O. Illiteracy and Vulnerability to HIV/AIDS: The Case of Igueben Women in Nigeria. Pretoria, South Africa: University of South Africa; 2011.  Back to cited text no. 25


  [Figure 1], [Figure 2], [Figure 3], [Figure 4]


    Similar in PUBMED
   Search Pubmed for
   Search in Google Scholar for
 Related articles
    Access Statistics
    Email Alert *
    Add to My List *
* Registration required (free)  

  In this article
    Materials and Me...
    Article Figures

 Article Access Statistics
    PDF Downloaded84    
    Comments [Add]    

Recommend this journal