Using variable importance measures from causal inference to rank risk factors of schistosomiasis infection in a rural setting in China
Tóm tắt
Schistosomiasis infection, contracted through contact with contaminated water, is a global public health concern. In this paper we analyze data from a retrospective study reporting water contact and schistosomiasis infection status among 1011 individuals in rural China. We present semi-parametric methods for identifying risk factors through a comparison of three analysis approaches: a prediction-focused machine learning algorithm, a simple main-effects multivariable regression, and a semi-parametric variable importance (VI) estimate inspired by a causal population intervention parameter. The multivariable regression found only tool washing to be associated with the outcome, with a relative risk of 1.03 and a 95% confidence interval (CI) of 1.01-1.05. Three types of water contact were found to be associated with the outcome in the semi-parametric VI analysis: July water contact (VI estimate 0.16, 95% CI 0.11-0.22), water contact from tool washing (VI estimate 0.88, 95% CI 0.80-0.97), and water contact from rice planting (VI estimate 0.71, 95% CI 0.53-0.96). The July VI result, in particular, indicated a strong association with infection status - its causal interpretation implies that eliminating water contact in July would reduce the prevalence of schistosomiasis in our study population by 84%, or from 0.3 to 0.05 (95% CI 78%-89%). The July VI estimate suggests possible within-season variability in schistosomiasis infection risk, an association not detected by the regression analysis. Though there are many limitations to this study that temper the potential for causal interpretations, if a high-risk time period could be detected in something close to real time, new prevention options would be opened. Most importantly, we emphasize that traditional regression approaches are usually based on arbitrary pre-specified models, making their parameters difficult to interpret in the context of real-world applications. Our results support the practical application of analysis approaches that, in contrast, do not require arbitrary model pre-specification, estimate parameters that have simple public health interpretations, and apply inference that considers model selection as a source of variation.
Tài liệu tham khảo
WHO: Report of the Scientific Working Group meeting on Schistosomiasis. WHO: Geneva; 2006.
King CH, Dickman K, Tisch DJ: Reassessment of the cost of chronic helminthic infection: a meta-analysis of disability-related outcomes in endemic Schistosomiasis. Am J Trop Med Hyg 2005,70(4):443–448.
Leenstra T, Coutinho HM, Acosta LP, Langdon GC, Su L, Olveda RM, McGarvey ST, Kurtis JD, Friedman JF: Schistosoma japonicum reinfection after praziquantel treatment causes anemia associated with inflammation. Infect Immun 2006,74(11):6398–6407.
Ross AG, Bartley PB, Sleigh AC, Olds GR, Li Y, Williams GM, McManus DP: Schistosomiasis. N Engl J Med 2002,346(16):1212–1220.
Clennon JA, King CH, Muchiri EM, Kariuki HC, Ouma JH, Mungai P, Kitron U: Spatial patterns of urinary schistosomiasis infection in a highly endemic area of coastal Kenya. Am J Trop Med Hyg 2004,70(4):443–448.
Li Y, Sleigh AC, Williams GM, Ross AG, Forsyth SJ, Tanner M, McManus DP: Measuring exposure to Schistosoma japonicum in China. III. Activity diaries, snail and human infection, transmission ecology and options for control. Acta Trop 2000,75(3):279–289.
Seto EY, Lee YJ, Liang S, Zhong B: Individual and village-level study of water contact patterns and Schistosoma japonicum infection in mountainous rural China. Trop Med Int Health 2007,12(10):1199–1209.
Nojima H, Santos AT, Blas BL, Kamiya H: The emergence of Schistosoma japonicum cercariae from Oncomelania quadrasi. J Parasitol 1980,66(6):1010–1013.
Li YS, Sleigh AC, Ross AG, Williams GM, Tanner M, McManus DP: Epidemiology of Schistosoma japonicum in China: morbidity and strategies for control in the Dongting Lake region. Int J Parasitol 2000,30(3):273–281.
Remais J, Hubbard A, Wu ZS, Spear RC: Weather-driven dynamics of an intermediate host: mechanistic and statistical population modelling of Oncomelania hupensis. Journal of Applied Ecology 2007,44(4):781–791.
Remais J, Liang S, Spear RC: Coupling hydrologic and infectious disease models to explain regional differences in schistosomiasis transmission in southwestern China. Environmental Science & Technology 2008,42(7):2643–2649.
Matthys B, Tschannen AB, Tian-Bi NT, Comoe H, Diabate S, Traore M, Vounatsou P, Raso G, Gosoniu L, Tanner M, Cisse G, N'Goran EK, Utzinger J: Risk factors for Schistosoma mansoni and hookworm in urban farming communities in western Cote d'Ivoire. Trop Med Int Health 2007,12(6):709–723.
Rudge JW, Stothard JR, Basanez MG, Mgeni AF, Khamis IS, Khamis AN, Rollingson D: Micro-epidemiology of urinary schistosomiasis in Zanzibar: Local risk factors associated with distribution of infections among schoolchildren and relevance for control. Acta Trop 2008,105(1):45–54.
Robins JM, Ritov Y: Toward a curse of dimensionality appropriate (CODA) asymptotic theory for semi-parametric models. Statistics in Medicine 1997, 16:285–319.
Hubbard AE, Ahern J, Fleischer NL, van der Laan M, Lippman SA, Jewell N, Bruckner T, Satariano WA: To GEE or not to GEE: comparing population average and mixed models for estimating the associations between neighborhood risk factors and health. Epidemiology 2010,21(4):475–8. discussion 479–81
Ahern J, Hubbard A, Galea S: Estimating the effects of potential public health interventions on population disease burden: a step-by-step illustration of causal inference methods. American Journal of Epidemiology 2009,169(9):1140–1147.
Fleischer NL, Fernald LC, Hubbard AE: Depressive symptoms in low-income women in rural Mexico. Epidemiology 2007,18(6):678–685.
Hubbard AE, van der Laan M: Population Intervention Models. Biometrika 2007, 95:35–47.
Greenland S, Drescher K: Maximum likelihood estimation of the attributable fraction from logistic models. Biometrics 1993,49(3):865–72.
Spear RC, Seto E, Liang S, Birkner M, Hubbard A, Qiu D, Yang C, Zhong B, Xu F, Gu X, Davis GM: Factors influencing the transmission of Schistosoma japonicum in the mountains of Sichuan Province of China. Am J Trop Med Hyg 2004,70(1):48–56.
Mosteller RD: Simplified calculation of body-surface area. N Engl J Med 1987,317(17):1098.
The Office of Endemic Disease Control MoH: Handbook of Schistosomiasis Control. Shanghai:Shanghai Science & Technology Press; 2000.
Katz N, Chaves A, Pellegrino J: A simple device for quantitative stool thick-smear technique in Schistosomiasis mansoni. Rev Inst Med Trop Sao Paulo 1972,14(6):397–400.
R version 2.10.0, Copyright (C) 2009 The R Foundation for Statistical Computing [http://www.r-project.org]
Breiman , Friedman , Olshen , Stone : Classification and Regression Trees. Wadsworth. 1984.
Atkinson EJ, Therneau TM: An Introduction to Recursive Partitioning Using the RPART Routines. Technicial Report 61, Mayo Clinic, Section of Statistics. 1997.
Sinisi SE, Polley EC, Petersen ML, Rhee SY, van der Laan MJ: SuperLearning: an application to the prediction of HIV-1 drug resistance. Statistical Applications in Genetics and Molecular Biology 2007,6(1):Article 7.
Efron B: The Jackknife, the Bootstrap and Other Re-sampling Plans. CBMS-NSF Regional Conference Series in Applied Mathematics 38. Capital City Press; 1982.
Huber PJ: The behavior of maximum likelihood estimates under nonstandard conditions. In Proceedings of the Fifth Berkeley Symposium on Mathematical Statistics and Probability. Volume 1. University of California Press; 1967:221–223.
White H: A heteroskedasticity-consistent covariance matrix estimator and a direct test for heteroskedasticity. Econometrica 1980, 48:817–830.
Stata 10, StataCorp LP, College Station, TX
Rubin DB: Bayesian inference for causal effects: the role of randomization. Ann Statist 1978, 6:34–58.
Rubin DB: Comment on a paper by P.W. Holland. J Am Statist Assoc 1986, 81:961–2.
Cole SR, Hernán MA: Constructing inverse probability weights for marginal structural models. Am J Epidemiology 2008,168(6):656–664.
Mortimer KM, Neugebauer R, van der Laan M, Tager IB: An application of model-fitting procedures for marginal structural models. Am J Epidemiology 2005,162(4):382–388.
Messer LC, Oakes JM, Mason S: Effects of socioeconomic and racial residential segregation on preterm birth: a cautionary tale of structural confounding. Am J Epidemiology 2010,171(6):664–73.
Breiman L: Random forests. Machine Learning 2001,45(1):5–32.
Sinisi SE, van der Laan MJ: Deletion/Substitution/Addition algorithm in learning with applications in genomics. Statistical Applications in Genetics and Molecular Biology 2004.,3(1):
Kooperberg C, Bose S, Stone CJ: Polychotomous regression. Jour Am Stat Assoc 1997, 92:117–127.
Spear RC, Zhong B, Mao Y, Hubbard A, Birkner M, Remais J, Qiu D: Spatial and temporal variability in schistosome cercarial density detected by mouse bioassays in village irrigation ditches in Sichuan, China. Am J Trop Med Hyg 2004,71(5):554–557.
Hung YW, Remais J: Quantitative detection of Schistosoma japonicum cercariae in water by real-time PCR. PLoS Neglected Tropical Diseases 2008, 2:e337.
Liang S, Seto EY, Remais JV, Zhong B, Yang C, Hubbard A, Davis GM, Gu X, Qiu D, Spear RC: Environmental effects on parasitic disease transmission exemplified by schistosomiasis in western China. Proc Natl Acad Sci USA 2007,104(17):7110–5. Epub 2007 Apr 16
Robins JM: Marginal structural models versus structural nested models as tools for causal inference. In Statistical Models in Epidemiology, the Environment, and Clinical Trials. Edited by: Halloran ME, Berry D. Springer, New York; 2000:95–113.
van der Laan MJ, Rubin D: Targeted maximum likelihood learning. The International Journal of Biostatistics 2006.,2(1):
van der Laan MJ, Gruber S: Collaborative double robust targeted penalized maximum likelihood estimation. [http://www.bepress.com/ucbbiostat/paper246] U.C. Berkeley Division of Biostatistics Working Paper Series Working Paper 246 2009.