A comparison of missing data procedures for addressing selection bias in HIV sentinel surveillance data

Population Health Metrics

Stephen S Lim¹  Christopher JL Murray¹  Emmanuela Gakidou¹  Marie Ng¹ 

[1] Institute for Health Metrics and Evaluation, University of Washington, Seattle, USA

关键词: Complete-case analysis;    Multiple imputation;    Missing data;    Simulations;    Selection bias;   

Others  :  804238
DOI  :  10.1186/1478-7954-11-12

 received in 2012-01-31, accepted in 2013-07-15,  发布年份 2013

【 摘 要 】

Background

Selection bias is common in clinic-based HIV surveillance. Clinics located in HIV hotspots are often the first to be chosen and monitored, while clinics in less prevalent areas are added to the surveillance system later on. Consequently, the estimated HIV prevalence based on clinic data is substantially distorted, with markedly higher HIV prevalence in the earlier periods and trends that reveal much more dramatic declines than actually occur.

Methods

Using simulations, we compare and contrast the performance of the various approaches and models for handling selection bias in clinic-based HIV surveillance. In particular, we compare the application of complete-case analysis and multiple imputation (MI). Several models are considered for each of the approaches. We demonstrate the application of the methods through sentinel surveillance data collected between 2002 and 2008 from India.

Results

Simulations suggested that selection bias, if not handled properly, can lead to biased estimates of HIV prevalence trends and inaccurate evaluation of program impact. Complete-case analysis and MI differed considerably in their ability to handle selection bias. In scenarios where HIV prevalence remained constant over time (i.e. β = 0), the estimated View MathML"> derived from MI tended to be biased downward. Depending on the imputation model used, the estimated bias ranged from −1.883 to −0.048 in logit prevalence. Furthermore, as the level of selection bias intensified, the extent of bias also increased. In contrast, the estimates yielded by complete-case analysis were relatively unbiased and stable across the various scenarios. The estimated bias ranged from −0.002 to 0.002 in logit prevalence.

Conclusions

Given that selection bias is common in clinic-based HIV surveillance, when analyzing data from such sources appropriate adjustment methods need to be applied. The results in this paper suggest that indiscriminant application of imputation models can lead to biased results.

【 授权许可】

   
2013 Ng et al.; licensee BioMed Central Ltd.

【 图 表 】

Figure 1.

Figure 2.

【 参考文献 】

• [1]WHO: Towards universal access: scaling up priority HIV/AIDS interventions in the health sector. Geneva; 2010. [progress report 2010]
• [2]UNAIDS: Scaling up access to HIV prevention, treatment, care and support: the next steps. 2006.
• [3]UNAIDS: AIDS epidemic update: special report on HIV prevention. 2005.
• [4]Gouws E, Mishra V, Fowler TB: Comparison of adult HIV prevalence from national population-based surveys and antenatal clinic surveillance in countries with generalised epidemics: implications for calibrating surveillance data. Sex Transm Infect 2008, 84:i17-i23.
• [5]Walker N, Garcia-Calleja JM, Heaton L, Asamoah-Odei E, Poumerol G, Lazzari S, Ghys PD, Schwartländer B, Stanecki KA: Epidemiological analysis of the quality of HIV sero-surveillance in the world: how well do we track the epidemic? AIDS 2001, 15:1545-1554.
• [6]Garcia-Calleja JM, Zaniewski E, Ghys PD, Stanecki K, Walker N: A global analysis of trends in the quality of HIV sero-surveillance. Sex Transm Infect 2004, 80:i25-i30.
• [7]UNAIDS: Report on the global AIDS epidemic. Geneva; 2004.
• [8]UNAIDS, WHO: Report on the global HIV/AIDS epidemic. 2002.
• [9]Walker N, Grassly NC, Garnett GP, Stanecki KA, Ghys PD: Estimating the global burden of HIV/AIDS: what do we really know about the HIV pandemic? Lancet 2004, 363:2180-2185.
• [10]Diaz T, Garcia-Calleja JM, Ghys PD, Sabin K: Advances and future directions in HIV surveillance in low-and middle-income countries. Curr Opin HIV AIDS 2009, 4:253-259.
• [11]National Aids Control Organisation and Government of India Ministry of Health Family Welfare: Report on HIV sentinel surveillance. New Delhi; 1997.
• [12]National Center for HIV/AIDS DaS, Ministry of Health: Report on HIV sentinel surveillance in Cambodia 2006. Phnom Penh; 2006.
• [13]Cheluget B, Baltazar G, Orege P, Ibrahim M, Marum LH, Stover J: Evidence for population level declines in adult HIV prevalence in Kenya. Sex Transm Infect 2006, 82:i21-i26.
• [14]National AIDS and STD Control Programme (NASCOP) MoH. Kenya: Senitnel Surveillance of HIV & STDs in Kenya Report; 2006.
• [15]Ministry of Health and Sanitation SL: Antenatal HIV sentinel surveillance protocol. 2003.
• [16]UNAIDS, UNICEIF, WHO: Epidemiological fact sheet on HIV/AODS and sexually transmitted infections Cameroon. 2005.
• [17]National Aids Control Organisation and Government of India Ministry of Health Family Welfare: HIV sentinel surveillance and HIV estimation in India: A technical brief. Delhi; 2007.
• [18]National Aids Control Organisation and Government of India Ministry of Health Family Welfare: Technical report India HIV estimates. 2006.
• [19]Rubin DB: Multiple Imputation for Nonresponse in Surveys. New York: Wiley; 1987.
• [20]Rubin DB, Schenker N: Interval Estimation from Multiply-Imputed Data: A Case Study Using Census Agriculture Industry Codes. Journal of Official Statistics 1986, 3(4):375-387.
• [21]Moses S, Ramesh BM, Nagelkerke NJD, Khera A, Isac S, Bhattacharjee P, Gurnani V, Washington R, Prakash KH, Pradeep BS: Impact of an intensive HIV prevention programme for female sex workers on HIV prevalence among antenatal clinic attenders in Karnataka state, south India: an ecological analysis. AIDS 2008, 22:101-108.
• [22]Ng M, Gakidou E, Levin-Rector A, Khera A, Murray CJL, Dandona L: Assessment of population-level effect of Avahan, an HIV-prevention initiative in India. Lancet 2001, 378(9803):1643-1652.
• [23]Jonathan ACS, Ian RW, John BC, Michael S, Patrick R, Michael GK, Angela MW, James RC: Multiple imputation for missing data in epidemiological and clinical research: potential and pitfalls. BMJ 2009, 338:157-160.
• [24]Marston L, Carpenter JR, Walters KR, Morris RW, Nazareth I, Petersen I: Issues in multiple imputation of missing data for large general practice clinical databases. Pharmacoepidemiol Drug Saf 2010, 19(6):618-626.
• [25]Little RJA, Rubin DB: Statistical Analysis with Missing Data, Volume John. Second Edition edn. New Jersey: Wiley and Sons; 2002.
• [26]Joseph L, Belisle P, Tamim H: Selection bias found in interpreting analyses with missing data for the prehospital index for trauma. J Clin Epidemiol 2004, 57:147-53.
• [27]Tsegaye A, de Wit TF R, Mekonnen Y, Beyene A, Aklilu M, Messele T, Abebe A, Coutinho R, Sanders E, Fontanet AL: Decline in prevalence of HIV-1 infection and syphilis among young women attending antenatal care clinics in Addis Ababa, Ethiopia: results from sentinel surveillance, 1995–2001. JAIDS Journal of Acquired Immune Deficiency Syndromes 2002, 30:359-362.
• [28]Kuate S, Mikolajczyk RT, Forgwei GW, Tih PM, Welty TK, Kretzschmar M: Time trends and regional differences in the prevalence of HIV infection among women attending antenatal clinics in 2 provinces in Cameroon. JAIDS Journal of Acquired Immune Deficiency Syndromes 2009, 52:258-264.
• [29]Reza-Paul S, Beattie T, Syed HUR, Venukumar KT, Venugopal MS, Fathima MP, Raghavendra HR, Akram P, Manjula R, Lakshmi M, et al.: Declines in risk behaviour and sexually transmitted infection prevalence following a community-led HIV preventive intervention among female sex workers in Mysore, India. AIDS 2008, 22:S91-S100.