Missing in space: an evaluation of imputation methods for missing data in spatial analysis of risk factors for type II diabetes

International Journal of Health Geographics

Kerrie Mengersen¹  Nicole White¹  Jannah Baker¹ 

[1] Cooperative Research Centres for Spatial Information, Melbourne, Australia

关键词: Diabetes;    Prevalence;    Spatial;    Missing;    Imputation;   

Others  :  1135991
DOI  :  10.1186/1476-072X-13-47

 received in 2014-08-29, accepted in 2014-11-10,  发布年份 2014

【 摘 要 】

Background

Spatial analysis is increasingly important for identifying modifiable geographic risk factors for disease. However, spatial health data from surveys are often incomplete, ranging from missing data for only a few variables, to missing data for many variables. For spatial analyses of health outcomes, selection of an appropriate imputation method is critical in order to produce the most accurate inferences.

Methods

We present a cross-validation approach to select between three imputation methods for health survey data with correlated lifestyle covariates, using as a case study, type II diabetes mellitus (DM II) risk across 71 Queensland Local Government Areas (LGAs). We compare the accuracy of mean imputation to imputation using multivariate normal and conditional autoregressive prior distributions.

Results

Choice of imputation method depends upon the application and is not necessarily the most complex method. Mean imputation was selected as the most accurate method in this application.

Conclusions

Selecting an appropriate imputation method for health survey data, after accounting for spatial correlation and correlation between covariates, allows more complete analysis of geographic risk factors for disease with more confidence in the results to inform public policy decision-making.

【 授权许可】

   
2014 Baker et al.; licensee BioMed Central Ltd.

【 图 表 】

Figure 1.

Figure 2.

Figure 3.

【 参考文献 】

• [1]Earnest A, Morgan G, Mengersen KL, Ryan L, Summerhayes R, Beard J: Evaluating the effect of neighbourhood weight matrices on smoothing properties of Conditional Autoregressive (CAR) models. Int J Health Geogr 2007, 6:54. BioMed Central Full Text
• [2]Besag J, York J, Mollie A: Bayesian image restoration with two application in spatial statistics. Annc Inst Statist Math 1991, 43(1):1-59.
• [3]Diabetes UK: Diabetes in the UK 2012. Diabetes UK 2012. [http://www.diabetes.org.uk/Documents/Reports/Diabetes-in-the-UK-2012.pdf webcite]
• [4]Holden SH, Barnett AH, Peters JR, Jenkins-Jones S, Poole CD, Morgan CL, Currie CJ: The incidence of type 2 diabetes in the United Kingdom from 1991 to 2010. Diabetes Obes Metab 2010, 15(9):844-852.
• [5]Palmer AJ, Tucker DM: Cost and clinical implications of diabetes prevention in an Australian setting: a long-term modeling analysis. Prim Care Diabetes 2012, 6(2):109-121.
• [6]Harris MI, Eastman RC: Early detection of undiagnosed diabetes mellitus: a US perspective. Diabetes Metab Res Rev 2000, 16(4):230-236.
• [7]Liese AD, Lawson A, Song HR, Hibbert JD, Porter DE, Nichols M, Lamichhane AP, Dabelea D, Mayer-Davis EJ, Standiford D, Liu L, Hamman RF, D’Agostino RB Jr: Evaluating geographic variation in type 1 and type 2 diabetes mellitus incidence in youth in four US regions. Health Place 2010, 16(3):547-556.
• [8]Noble D, Mathur R, Dent T, Meads C, Greenhalgh T: Risk models and scores for type 2 diabetes: systematic review. BMJ 2011, 343:d7163.
• [9]Weng C, Coppini DV, Sonksen PH: Geographic and social factors are related to increased morbidity and mortality rates in diabetic patients. Diabet Med 2000, 17(8):612-617.
• [10]Egede LE, Gebregziabher M, Hunt KJ, Axon RN, Echols C, Gilbert GE, Mauldin PD: Regional, geographic, and racial/ethnic variation in glycemic control in a national sample of veterans with diabetes. Diabetes Care 2011, 34(4):938-943.
• [11]Green C, Hoppa RD, Young TK, Blanchard JF: Geographic analysis of diabetes prevalence in an urban area. Soc Sci Med 2003, 57(3):551-560.
• [12]Bocquier A, Cortaredona S, Nauleau S, Jardin M, Verger P: Prevalence of treated diabetes: Geographical variations at the small-area level and their association with area-level characteristics. A multilevel analysis in Southeastern France. Diabetes Metab 2011, 37(1):39-46.
• [13]Geraghty EM, Balsbaugh T, Nuovo J, Tandon S: Using Geographic Information Systems (GIS) to assess outcome disparities in patients with type 2 diabetes and hyperlipidemia. J Am Board Fam Med 2010, 23(1):88-96. Jan-Feb
• [14]Chaix B, Billaudeau N, Thomas F, Havard S, Evans D, Kestens Y, Bean K: Neighborhood effects on health: correcting bias from neighborhood effects on participation. Epidemiology 2011, 22(1):18-26.
• [15]Congdon P: Estimating diabetes prevalence by small area in England. J Public Health (Oxf) 2006, 28(1):71-81.
• [16]Kravchenko VI, Tronko ND, Pankiv VI, Venzilovich Yu M, Prudius FG: Prevalence of diabetes mellitus and its complications in the Ukraine. Diabetes Res Clin Pract 1996, 34(Suppl):S73-S78.
• [17]Lee JM, Davis MM, Menon RK, Freed GL: Geographic distribution of childhood diabetes and obesity relative to the supply of pediatric endocrinologists in the United States. J Pediatr 2008, 152(3):331-336.
• [18]Noble D, Smith D, Mathur R, Robson J, Greenhalgh T: Feasibility study of geospatial mapping of chronic disease risk to inform public health commissioning. BMJ Open 2012, 2(1):e000711.
• [19]Magalhaes RJ, Clements AC: Mapping the risk of anaemia in preschool-age children: the contribution of malnutrition, malaria, and helminth infections in West Africa. PLoS Med 2011, 8(6):e1000438.
• [20]Stromberg U, Magnusson K, Holmen A, Twetman S: Geo-mapping of caries risk in children and adolescents - a novel approach for allocation of preventive care. BMC Oral Health 2011, 11:26. BioMed Central Full Text
• [21]Joshua V, Gupte MD, Bhagavandas M: A Bayesian approach to study the space time variation of leprosy in an endemic area of Tamil Nadu, South India. Int J Health Geogr 2008, 7:40. BioMed Central Full Text
• [22]Cocco E, Sardu C, Massa R, Mamusa E, Musu L, Ferrigno P, Melis M, Montomoli C, Ferretti V, Coghe G, Fenu G, Frau J, Lorefice L, Carboni N, Contu P, Marrosu MG: Epidemiology of multiple sclerosis in south-western Sardinia. Mult Scler 2011, 17(11):1282-1289.
• [23]Goovaerts P: Geostatistical analysis of disease data: accounting for spatial support and population density in the isopleth mapping of cancer mortality risk using area-to-point Poisson kriging. Int J Health Geogr 2006, 5:52. BioMed Central Full Text
• [24]Hegarty AC, Carsin AE, Comber H: Geographical analysis of cancer incidence in Ireland: a comparison of two Bayesian spatial models. Cancer Epidemiol 2010, 34(4):373-381.
• [25]Cramb SM, Mengersen KL, Baade PD: Developing the atlas of cancer in Queensland: methodological issues. Int J Health Geogr 2011, 10:9. BioMed Central Full Text
• [26]Haque U, Magalhaes RJ, Reid HL, Clements AC, Ahmed SM, Islam A, Yamamoto T, Haque R, Glass GE: Spatial prediction of malaria prevalence in an endemic area of Bangladesh. Malar J 2010, 9:120. BioMed Central Full Text
• [27]Zayeri F, Salehi M, Pirhosseini H: Geographical mapping and Bayesian spatial modeling of malaria incidence in Sistan and Baluchistan province, Iran. Asian Pac J Trop Med 2011, 4(12):985-992.
• [28]Stensgaard AS, Vounatsou P, Onapa AW, Simonsen PE, Pedersen EM, Rahbek C, Kristensen TK: Bayesian geostatistical modelling of malaria and lymphatic filariasis infections in Uganda: predictors of risk and geographical patterns of co-endemicity. Malar J 2011, 10:298. BioMed Central Full Text
• [29]Kang SY, McGree J, Mengersen K: The impact of spatial scales and spatial smoothing on the outcome of bayesian spatial model. PLoS One 2013, 8(10):e75957.
• [30]National Diabetes Services Scheme: Australian Diabetes Map. 2012. [http://www.ndss.com.au/Australian-Diabetes-Map/ webcite]
• [31]Australian Bureau of Statistics: 3218.0 Population Estimates by Local Government Area, 2001 to 2011. 2012. [http://www.abs.gov.au/AUSSTATS/abs@.nsf/DetailsPage/3218.02011 webcite]
• [32]Queensland Government: Queensland self-reported health status 2009–2010: Local Government Area summary report. 2011. [http://www.health.qld.gov.au/epidemiology/documents/srhs0910lgasummary.pdf webcite]
• [33]Besag J: Spatial interaction and the statistical analysis of lattice systems. J Royal Sta Soc Ser B (Methodological) 1974, 36(2):192-236.
• [34]Pascutto C, Wakefield JC, Best NG, Richardson S, Bernardinelli L, Staines A, Elliott P: Statistical issues in the analysis of disease mapping data. Stat Med 2000, 19(17–18):2493-519. Sep 15–30
• [35]The R Project: The R Project for Statistical Computing. 2014. [ http://www.r-project.org/ webcite]
• [36]The BUGS Project: WinBUGS. 2014. [http://www.mrc-bsu.cam.ac.uk/software/bugs/the-bugs-project-winbugs/ webcite]
• [37]Spiegelhalter D, Best NG, Carlin B, Van Der Linde A: Bayesian measures of model complexity and fit. J Royal Sta Soc 2002, 64(4):583-639.