Use of Patients With Diarrhea Who Test Negative for Rotavirus as Controls to Estimate Rotavirus Vaccine Effectiveness Through Case-Control Studies

Clinical Infectious Diseases, Apr 2016

Background. Case-control studies are often performed to estimate postlicensure vaccine effectiveness (VE), but the enrollment of controls can be challenging, time-consuming, and costly. We evaluated whether children enrolled in the same hospital-based diarrheal surveillance used to identify rotavirus cases but who test negative for rotavirus (test-negative controls) can be considered a suitable alternative to nondiarrheal hospital or community-based control groups (traditional controls). Methods. We compared calculated VE estimates as a function of varying values of true VE, attack rates of rotavirus and nonrotavirus diarrhea in the population, and sensitivity and specificity of the rotavirus enzyme immunoasssay. We also searched the literature to identify rotavirus VE studies that used traditional and test-negative control groups and compared VE estimates obtained using the different control groups. Results. Assuming a 1% attack rate for severe rotavirus diarrhea, a 3% attack rate for severe nonrotavirus diarrhea in the population, a test sensitivity of 96%, and a specificity of 100%, the calculated VE estimates using both the traditional and test-negative control groups closely approximated the true VE for all values from 30% to 100%. As true VE decreased, the traditional case-control approach slightly overestimated the true VE and the test-negative case-control approach slightly underestimated this estimate, but the absolute difference was only ±0.2 percentage points. Field VE estimates from 10 evaluations that used both traditional and test-negative control groups were similar regardless of control group used. Conclusions. The use of rotavirus test-negative controls offers an efficient and cost-effective approach to estimating rotavirus VE through case-control studies.

A PDF file should load here. If you do not see its contents the file may be temporarily unavailable at the journal website or you do not have a PDF plug-in installed and enabled in your browser.

Alternatively, you can download the file locally and open with any standalone PDF reader:

http://cid.oxfordjournals.org/content/62/suppl_2/S106.full.pdf

Use of Patients With Diarrhea Who Test Negative for Rotavirus as Controls to Estimate Rotavirus Vaccine Effectiveness Through Case-Control Studies

CID Use of Patients With Diarrhea Who Test Negative for Rotavirus as Controls to Estimate Rotavirus Vaccine Effectiveness Through Case-Control Studies Jacqueline E. Tate 0 Manish M. Patel 0 Margaret M. Cortese 0 Daniel C. Payne 0 Benjamin A. Lopman 0 Catherine Yen 0 Umesh D. Parashar 0 0 Centers for Disease Control and Prevention , Atlanta , Georgia Background. Case-control studies are often performed to estimate postlicensure vaccine effectiveness (VE), but the enrollment of controls can be challenging, time-consuming, and costly. We evaluated whether children enrolled in the same hospital-based diarrheal surveillance used to identify rotavirus cases but who test negative for rotavirus (test-negative controls) can be considered a suitable alternative to nondiarrheal hospital or community-based control groups (traditional controls). Methods. We compared calculated VE estimates as a function of varying values of true VE, attack rates of rotavirus and nonrotavirus diarrhea in the population, and sensitivity and specificity of the rotavirus enzyme immunoasssay. We also searched the literature to identify rotavirus VE studies that used traditional and test-negative control groups and compared VE estimates obtained using the different control groups. Results. Assuming a 1% attack rate for severe rotavirus diarrhea, a 3% attack rate for severe nonrotavirus diarrhea in the population, a test sensitivity of 96%, and a specificity of 100%, the calculated VE estimates using both the traditional and test-negative control groups closely approximated the true VE for all values from 30% to 100%. As true VE decreased, the traditional case-control approach slightly overestimated the true VE and the test-negative case-control approach slightly underestimated this estimate, but the absolute difference was only ±0.2 percentage points. Field VE estimates from 10 evaluations that used both traditional and testnegative control groups were similar regardless of control group used. Conclusions. The use of rotavirus test-negative controls offers an efficient and cost-effective approach to estimating rotavirus VE through case-control studies. - Two live attenuated, orally administered rotavirus vaccines, a monovalent vaccine (RV1) (Rotarix, GSK Biologicals) based on a human rotavirus strain and a pentavalent bovine-human reassortant vaccine (RV5) (RotaTeq, Merck & Co), are licensed and recommended for use in all countries globally [ 1–3 ]. These vaccines demonstrated high efficacy (85%–98%) against severe rotavirus disease in prelicensure trials conducted in high- and middle-income countries in the Americas and Europe, but clinical trials in low-income countries of Africa and Asia, where most of the rotavirus mortality burden occurs, showed moderate vaccine efficacy (50%–70%) [ 2–6 ]. In anticipation of vaccine introduction, many countries began sentinel surveillance for rotavirus diarrhea [ 7 ]. Following introduction of rotavirus vaccine into national, routine immunization programs, many of these countries have used their diarrhea surveillance systems to document declines in mortality and/or morbidity due to rotavirus and all-cause diarrhea [ 8–16 ]. However, because vaccines may Correspondence: J. E. Tate, Centers for Disease Control and Prevention, 1600 Clifton Rd NE, MS-A34, Atlanta, GA 30333 ( ). Clinical Infectious Diseases® 2016;62(S2):S106–14 Published by Oxford University Press for the Infectious Diseases Society of America 2016. This work is written by (a) US Government employee(s) and is in the public domain in the US. DOI: 10.1093/cid/civ1014 perform differently under conditions of routine use than under the controlled conditions of randomized clinical trials, determining vaccine effectiveness (VE) postintroduction is important. Case-control studies are often performed to estimate VE by comparing the odds of vaccination in individuals who develop the disease of interest (cases) with the odds of vaccination in individuals who do not have the disease of interest (controls). The selection of an appropriate control group is important, and each type of control group has its own pros and cons and different biases. Controls should be representative of the source population that produces the cases and have the same chance to be vaccinated as the cases. Case-control studies have traditionally selected controls from the same community or neighborhood as the case or from the hospital where the case was treated for a condition unrelated to the exposure (eg, vaccination) or the disease of interest (eg, children with non-vaccine-preventable, nondiarrheal disease). Enrollment of these “traditional controls” from either hospital or community settings can be challenging, time-consuming, and costly. For rotavirus VE evaluations, an alternate resource and time-efficient approach to enrolling control subjects is to use children enrolled in concurrent diarrheal surveillance programs who test negative for rotavirus disease using an enzyme immunoassay (EIA), as these children are identified through the same system that is used to identify cases. These “test-negative controls” have similar healthcareseeking behavior for diarrheal illness as cases with confirmed rotavirus diarrhea and an equal likelihood of being tested for rotavirus as all surveillance specimens are tested. Influenza VE studies have pioneered the use of data for cases and test-negative controls obtained from laboratory testing of patients with influenza-like illness to obtain VE estimates and have shown that these estimates are similar to those obtained by using traditional controls [ 17–22 ]. Rotavirus VE evaluations have also been performed using test-negative controls from diarrhea surveillance data, but a formal evaluation comparing these results with results obtained using traditional controls has not been done. The rotavirus EIA has both higher sensitivity and specificity than the rapid influenza test and thus, in theory, rotavirus test-negative controls should provide even more accurate estimates to those obtained using traditional control groups than have been observed in influenza VE studies. When selecting a potential control group in a case-control evaluation, its strengths and weaknesses should be evaluated. The objective of this article is to describe how the test performance of the rotavirus EIA affects VE estimates when the test-negative control group is used. Specifically, we examine how changing the values of selected inputs affects the theoretical estimates of rotavirus VE and compare this theoretical estimate to effectiveness estimates obtained though field evaluations of rotavirus vaccine using both test-negative and traditional control groups. METHODS Definitions Calculated VE incorporates sensitivity and specificity of the test into estimates of the true VE. A rotavirus case is defined as an individual who tests positive for rotavirus diarrhea by an EIA. A traditional control is an individual from the same community or neighborhood as the case (community control) or an individual from the hospital where the case was treated for a condition unrelated to the disease of interest (eg, children with nonvaccine-preventable, nondiarrheal disease) (hospital control). A test-negative control is an individual with the same symptoms (eg, diarrhea) as the case but testing negative for rotavirus. Expected Vaccine Effectiveness Estimates In case-control studies, calculated VE is a function of the true VE, the attack rates of rotavirus and nonrotavirus diarrhea in the population, and the sensitivity and specificity as previously described (Supplementary Appendix) [ 18 ]. We calculated expected rotavirus VE estimates using both test-negative controls and traditional controls under the following assumptions. First, we assumed that the attack rate for severe rotavirus diarrhea requiring hospital admission in children age eligible to receive vaccine was 1% and for nonrotavirus diarrhea was 3%. Second, because the true rotavirus VE appears to be inversely related to the level of child mortality in a country [ 23 ], we used a range of 30%–100% effectiveness against severe rotavirus disease to estimate true rotavirus VE in a range of settings. Third, we assumed that the EIA test for rotavirus is 96% sensitive and 100% specific based on manufacturer reports. Fourth, we also assumed that all persons in the population had equal likelihood of developing nonrotavirus diarrhea and being captured in the surveillance system, and that only vaccination affected the risk of developing rotavirus diarrhea and does not affect the chance of developing nonrotavirus diarrhea. The risk of developing rotavirus diarrhea was assumed to be independent of the risk of developing nonrotavirus diarrhea, and children with severe diarrhea requiring hospitalization were assumed to have the same care-seeking behavior regardless of etiology. We conducted a sensitivity analysis by varying the values of sensitivity and specificity of the EIA as well as each of our other input assumptions to identify those inputs that had the greatest influence on the expected VE. Field Vaccine Effectiveness Estimates The expected VE estimates were compared to estimates from published case-control evaluations of rotavirus VE that enrolled both traditional controls as well as test-negative controls. We systematically searched the literature to identify rotavirus VE studies that used traditional and test-negative control groups and provided separate results for each control group. We also identified rotavirus VE studies that presented results using only a test-negative control group. From each study, we abstracted the outcome, source of control, type of analysis performed, the rotavirus test used, and the estimated VE. Vaccine effectiveness estimates obtained using the different control groups were compared. RESULTS Expected Vaccine Effectiveness Estimates Assuming a 1% attack rate for severe rotavirus diarrhea and a 3% attack rate for severe nonrotavirus diarrhea in the population and a test sensitivity of 96% and specificity of 100%, the calculated VE using both a traditional control group and the test-negative control group closely approximates the true VE for all values ranging from 30% to 100%. Under these baseline assumptions, as true VE decreases, the use of traditional controls very slightly overestimates the true VE whereas the use of test-negative controls approach slightly underestimates the true VE, but the absolute difference in either direction is ≤0.2 percentage points (Table 1). Keeping sensitivity at 96% and specificity at 100%, if the attack rates increase by a magnitude of 10 to a 10% attack rate for severe rotavirus diarrhea and a 30% attack rate for severe nonrotavirus diarrhea, the estimated VE for both control groups is <3 percentage points greater than the true VE. Similarly, if the ratio between the 2 attack rates increases or decreases by a factor of 3, the true VE and calculated VE remain similar (Table 2). If the sensitivity and specificity of the rotavirus test are decreased by 5 percentage points to 91% and 95%, respectively, the case-control methodology, regardless of the control group used, underestimates the true VE when the true VE is high and the ratio between the 2 attack rates is pronounced (Table 2). The difference between the true and calculated VE diminishes as the true VE decreases or the ratio between the attack rates decreases. However, the absolute difference between the VE estimates calculated using the traditional and test-negative control groups remains small (within 3 percentage points), regardless of the scenario examined. In an extreme scenario of a poorly performing test with sensitivity and specificity of 75%, both traditional and test-negative controls significantly underestimate the true VE, with the test-negative controls resulting in a larger underestimate (Table 2). Field Vaccine Effectiveness Estimates Ten studies were identified in the literature that presented rotavirus VE results separately for a traditional control group and a test-negative control group including 3 that evaluated RV5, 5 that evaluated RV1, and 2 that evaluated both RV5 and RV1 (Tables 3 and 4) [ 24–33 ]. For the RV5 evaluations, 3 studies used children seeking medical care for acute respiratory illness or another illness not related to diarrhea as the traditional control group, and 2 studies selected traditional controls from immunization registries. All studies used an EIA to identify rotavirus-positive cases, and all were performed in high- or middle-income countries (4 studies in the United States and 1 in Taiwan). Compared with the point estimate for the traditional control group, the VE calculated using the test-negative control group was within 4 percentage points, and there was substantial overlap with the 95% confidence intervals in all RV5 studies (Table 3). Sensitivity and Specificity Sensitivity = 96% Specificity = 100% Sensitivity = 91% Specificity = 95% Sensitivity = 75% Specificity = 75% For the RV1 evaluations, 4 studies used children seeking medical care for acute respiratory illness or other nondiarrhea or non-vaccine-preventable conditions for the traditional control group, 2 studies selected traditional controls from an immunization registry, and 1 study used children from the community as traditional controls. Six studies used an EIA to identify rotavirus-positive cases and test-negative controls, and 1 study used an immunochromatographic assay. Five studies were performed in high- or middle-income countries (United States, Australia, Taiwan, Brazil, South Africa), and 2 were M M M n l o t a i e n s s o se n it r & o i d g on te om con ire c d ad tp , t e h e is m t tch itr a b sy ad log n a i id , ne re ce i,t n a lo Mie n c in ireem tacoR I)(;E ioB inC hO A s c io r n l o t a i e n s s o se n it r & o i d g on te om con ire c d ad tp , t e h e is m t tch itr a b sy ad log ro D E d e e iz th l a g I t n R en isp iit A r d o is r li h v fo h 6 e 3 l 6 iv 4 tro :1 ta :1 on iss eg l is o s Ic lya -tn rt ly s n a AR an te co an D su , rE ir v n o a o , t i t ay ro liza ts ro a rt f it o it E p h is G s s v A o f l n o l, a io a n s te it io s a p it re d s d g on ho on re c i c d & t t e h ca i s tch itr a b xe log , , ) A G I A N (E m EE su rah yan R ir SC tav ipo rm A o B e D R R G I ) 0 0 1 – 2 8 ( 7 9 2 7 7 ) 0 0 1 – 4 8 ( 7 9 S r o f 5 V R f o V . 3 e l b ) I C ) % 6 9 95 – ( 0 7 ( % , 9 E 8 G A : s n o i t a i v e r b b A D o w 8 , 2 0 1 6 A Test-Negative Controls to Estimate VE CID 2016:62 (Suppl 2) S109 ) 7 9 – 5 7 ( 2 9 3 7 ) 7 9 – 4 7 ( 1 9 on ,e i s it h l s s irt a e r n r b , io ge fo & edh itdn irtc ted ,no rte tac co is su sa ra n g jd e u m u lo a s q n n a id , i ne re ce i,t n a lo Mie n c in ireem tacoR I)(,EA ioB inC hO S c io r l n a io izp iiton sse r & d g n e on te o r c c d ad , it e h de is tch itr a b co log 2 2 5 n o i t a ry iz t n is u g m re m I 9 7 D su , rE ir v n o a o , t i t ay ro liza ts ro a rt f it o it E p h is G s s v A o ] 0 ne [3 s u 1 te ), J 1 ta ia –y 20 S rg ra – ed eo nu 10 i (G aJ 20 t n I) ) 0 C 7 % to 5 9 9 ( 1 1 % – , ( E 9 V 1 9 1 h t i s e i d u t r o f 1 f o V . 4 e l b S110 CID 2016:62 (Suppl 2) Tate et al D o w n l o a d e d 8 , 2 0 1 6 A ) 9 8 – 6 3 ( 3 7 o t 5 io 3 t le – a to ib (8 in m d , s G u r A i v a m t r o a R h p N io E B y E n R a R , C ) m S IA r e A E ( G D I c i f t t s o c i a g te x o a e l d l, l a n on itap n io io s d s it s e o d re h h n g c o e t & c r a 0 9 h t r i b f o e p y T . o N f o e p y T . o N C f o e c r u o d o i r e m i d e k n lie m y z n e , A S I L E ; y a s s a o n u E G A : . s s n s io e t n ia e v v i e t r c b e b ff performed in low-income countries (Bolivia and Malawi). For 4 of the studies, the VE calculated using a test-negative control group was within 3 percentage points of the point estimate for the traditional control group. Greater variability was observed in 3 studies. The VE estimates using the 2 different control groups were within 6 percentage points for a study in the United States, 8 percentage points for a study from Bolivia, and 12 percentage points for a study from South Africa. Howfi ever, for all studies, the 95% con dence intervals for the 2 controls groups overlapped substantially (Table 4). Five studies presented results using only test-negative controls (Table 5) [ 34–38 ]. Four studies examined both vaccines (RV1 and RV5), although 1 study did not present the results for each vaccine separately, and 1 study examined RV1 only. All but 1 study used an immunochromatographic assay to identify rotavirus-positive cases, and the remaining study used an EIA. All were performed in high- or middle-income countries (Spain, Israel, United States, and Colombia). These studies presented VE estimates that are comparable with the clinical trial estimates of vaccine efficacy in high- and middle-income countries. DISCUSSION Estimated VE as a function of rotavirus and nonrotavirus diarrhea attack rates and EIA test performance was almost identical to the true VE regardless of the control group used (traditional vs test-negative). Although the true VE in any given setting is fi unknown, the eld VE estimates from evaluations that used both the traditional and test-negative control groups were similar regardless of the control group used, and studies where only the test-negative control groups were used yielded results comfi patible with the clinical trials. These ndings indicate that the case-control methodology using EIA test-negative controls is a resource-efficient design to estimate the effectiveness of rotavirus vaccine. Test-negative controls have several advantages over traditional community and hospital control groups. First, children with endemic diarrheal illness who test negative for rotavirus have similar health-seeking behavior for severe diarrhea illness as children who test positive for rotavirus. Second, vaccination status can potentially be ascertained blinded to whether the patient is a case or a control. Finally, use of the test-negative control group can simplify enrollment and data collection and is resource and cost-efficient, but careful attention to study design and implementation is still needed. Test-negative controls have previously been used to estimate fl VE for seasonal in uenza vaccines [ 17–19 ]. However, rapid fl in uenza tests have substantially lower sensitivity than the rotafl virus EIA. Thus, high VE obtained using rapid in uenza testnegative controls suggests that the vaccine has been effective fl in reducing in uenza illness, but low VE may be a function of poor test performance. With an EIA test that has high D o w m n l o a d e d f r o h t t p i d . o x f o r d j o u r n a l s . o r g / b y g u e s t o n p r i l 8 2 0 1 6 A le n le d iligb 2 itao ise iligb ine -ee (e< iccn m irts -ee ;e tabo rtpo h t R f o l l u F a f o V . 5 e l b a T S112 CID 2016:62 (Suppl 2) Tate et al 0 6 ; ) y a s u ss ,t r i a d v o y ta n h s t e p s d ra rm eT /a g a s s tao ,D iru iru v v om AG a a r o tro ts ch rm t 5 7 e d n i c an 5V c a 1 R V V R s u r l i a av e ic t m d ro o e r c t m fo u t O hg rae EG u c A o y ] d ra 4 u [3 o i r e D E l; a v r e t n i e c n e d i f n o c I, C ; s i t i r e t n e o r t s a g D o w 8 , 2 0 1 6 A sensitivity and specificity as in our baseline scenario, a high VE obtained using rotavirus EIA test-negative controls will not only inform that the program has been effective in reducing rotavirus disease, it will also closely approximate the true VE. Additionally, the rotavirus EIA test is quick and easy to use, so all surveillance specimens can be tested in a timely and efficient manner. However, if the test used to identify cases of severe rotavirus diarrhea has lower sensitivity or, in particular, lower specificity than our assumption of 96% sensitivity and 100% specificity of the EIA, the observed field VE calculated using the case-control methodology will be substantially lower than the true VE, although the difference in the VE estimates obtained using traditional and test-negative control groups remains relatively small. A lower test specificity deteriorates the case group and therefore affects case-control studies regardless of which control group is used, but a poorer sensitivity misclassifies true cases as controls and therefore only biases the testnegative design. Our study had several limitations. First, the true VE and attack rates for rotavirus and nonrotavirus diarrhea are unknown. However, our sensitivity analysis showed that with a highly sensitive and specific test, even a 10-fold increase in the attack rate had little impact on the estimated VE for a wide range of true VE. If the sensitivity and specificity of the test were lower, the true VE was high, and the difference in attack rates pronounced, the estimated VE would underestimate the true VE, but both the traditional and test-negative control groups would provide similar underestimates. Second, in field evaluations, the VE estimated using the test-negative controls was usually similar but not always identical to the VE estimated using traditional control groups. Although it is unknown which estimate is closer to the true VE, the difference between these 2 estimates may be explained in part by different types of analyses (matched vs unmatched) or by unmeasured confounders that may be present in one of the control groups. In the studies where only testnegative controls were used, the results in these particular studies were similar to the expected results based on the clinical trials. However, lack of similarity between the observed field VE and the expected VE, such as the lower VE observed in the study in Australia, does not necessarily mean that the test-negative control group is performing poorly, as the true vaccine performance may be lower than expected due to host factors such as high maternal antibodies, malnutrition, or viral factors such as substantial strain heterogeneity. Third, the majority of field evaluations that included both a traditional and test-negative control group were conducted in high- and middle-income countries. Studies in Bolivia and South Africa had the largest absolute difference in estimated VE using the traditional and test-negative control groups with the test-negative estimate lower than that observed with the traditional control group. The EIA has been shown to perform well in both low- and high-income settings with regards to identifying symptomatic children with high viral loads measured by reverse transcription–polymerase chain reaction [ 39–41 ]. Differences in VE using the 2 different control groups can still occur if other factors in addition to vaccination influence the risk of developing rotavirus diarrhea or if any factors affect the probability of developing nonrotavirus diarrhea. Moreover, the nonrotavirus diarrhea group are likely to be of bacterial etiology, particularly in developing country settings, and this population may be fundamentally different from the case population. However, the evidence collected to date indicates that potential for bias with the test-negative controls has limited impact on the rotavirus VE estimates. In summary, VE estimates calculated using traditional control groups and rotavirus test-negative control groups both closely approximate the true VE when a highly sensitive and specific test is used to identify rotavirus-positive cases. Thus, the use of rotavirus test-negative controls enrolled through the same diarrhea surveillance platform used to identify rotavirus cases offers an efficient and cost-effective approach to estimating VE for rotavirus vaccines. As rotavirus vaccines continue to be introduced into national routine immunization programs globally, monitoring the impact and effectiveness of these programs is important to ensure continued support and commitment for the vaccination program. Supplementary Data Supplementary materials are available at http://cid.oxfordjournals.org. Consisting of data provided by the author to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the author, so questions or comments should be addressed to the author. Notes Disclaimer. The findings and conclusions of this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention (CDC). The views expressed by the authors do not necessarily reflect the views of PATH, the CDC Foundation, the Bill and Melinda Gates Foundation, or GAVI, the Vaccine Alliance. Supplement sponsorship. This article appears as part of the supplement “Health Benefits of Rotavirus Vaccination in Developing Countries,” sponsored by PATH and the CDC Foundation through grants from the Bill and Melinda Gates Foundation and GAVI, the Vaccine Alliance. Potential conflicts of interest. All authors: No reported conflicts. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed. 1. Conclusions and recommendations on rotavirus vaccination; WHO Strategic Advisory Group of Experts in Immunization meeting ( April 2009 ). Wkly Epidemiol Rec 2009 ; 84 : 197 . 2. Ruiz-Palacios GM , Perez-Schael I , Velazquez FR , et al. Safety and efficacy of an attenuated vaccine against severe rotavirus gastroenteritis . N Engl J Med 2006 ; 354 : 11 - 22 . 3. Vesikari T , Matson DO , Dennehy P , et al. Safety and efficacy of a pentavalent human-bovine (WC3) reassortant rotavirus vaccine . N Engl J Med 2006 ; 354 : 23 - 33 . 4. Armah GE , Sow SO , Breiman RF , et al. Efficacy of pentavalent rotavirus vaccine against severe rotavirus gastroenteritis in infants in developing countries in subSaharan Africa: a randomised, double-blind, placebo-controlled trial . Lancet 2010 ; 376 : 606 - 14 . 5. Madhi SA , Cunliffe NA , Steele D , et al. Effect of human rotavirus vaccine on severe diarrhea in African infants . N Engl J Med 2010 ; 362 : 289 - 98 . 6. Zaman K , Dang DA , Victor JC , et al. Efficacy of pentavalent rotavirus vaccine against severe rotavirus gastroenteritis in infants in developing countries in Asia: a randomised, double-blind, placebo-controlled trial . Lancet 2010 ; 376 : 615 - 23 . 7. Rotavirus surveillance worldwide- 2009. Wkly Epidemiol Rec 2011 ; 86 : 174 - 6 . 8. Richardson V , Hernandez-Pichardo J , Quintanar-Solares M , et al. Effect of rotavirus vaccination on death from childhood diarrhea in Mexico . N Engl J Med 2010 ; 362 : 299 - 305 . 9. Quintanar-Solares M , Yen C , Richardson V , Esparza-Aguilar M , Parashar UD , Patel MM . Impact of rotavirus vaccination on diarrhea-related hospitalizations among children <5 years of age in Mexico . Pediatr Infect Dis J 2011 ; 30 ( suppl 1 ): S11 - 5 . 10. do Carmo GM , Yen C , Cortes J , et al. Decline in diarrhea mortality and admissions after routine childhood rotavirus immunization in Brazil: a time-series analysis . PLoS Med 2011 ; 8 : e1001024 . 11. Yen C , Armero Guardado JA , Alberto P , et al. Decline in rotavirus hospitalizations and health care visits for childhood diarrhea following rotavirus vaccination in El Salvador . Pediatr Infect Dis J 2011 ; 30 ( suppl 1 ): S6 - 10 . 12. Raes M , Strens D , Vergison A , Verghote M , Standaert B. Reduction in pediatric rotavirus-related hospitalizations after universal rotavirus vaccination in Belgium . Pediatr Infect Dis J 2011 ; 30 : e120 - 5 . 13. Paulke-Korinek M , Rendi-Wagner P , Kundi M , Kronik R , Kollaritsch H . Universal mass vaccination against rotavirus gastroenteritis: impact on hospitalization rates in Austrian children . Pediatr Infect Dis J 2010 ; 29 : 319 - 23 . 14. Field EJ , Vally H , Grimwood K , Lambert SB . Pentavalent rotavirus vaccine and prevention of gastroenteritis hospitalizations in Australia . Pediatrics 2010 ; 126 : e506 - 12 . 15. Msimang VM , Page N , Groome MJ , et al. Impact of rotavirus vaccine on childhood diarrheal hospitalization following introduction into the South African public immunization program [manuscript published online ahead of print 8 August 2013] . Pediatr Infect Dis J 2013 . 16. Leshem E , Tate JE , Steiner CA , Curns AT , Lopman BA , Parashar UD . Acute gastroenteritis hospitalizations among US children following implementation of the rotavirus vaccine . JAMA 2015 ; 313 : 2282 - 4 . 17. Kelly H , Carville K , Grant K , Jacoby P , Tran T , Barr I. Estimation of influenza vaccine effectiveness from routine surveillance data . PLoS One 2009 ; 4 : e5079 . 18. Orenstein EW , De Serres G , Haber MJ , et al. Methodologic issues regarding the use of three observational study designs to assess influenza vaccine effectiveness . Int J Epidemiol 2007 ; 36 : 623 - 31 . 19. Skowronski DM , Masaro C , Kwindt TL , et al. Estimating vaccine effectiveness against laboratory-confirmed influenza using a sentinel physician network: results from the 2005-2006 season of dual A and B vaccine mismatch in Canada . Vaccine 2007 ; 25 : 2842 - 51 . 20. De Serres G , Skowronski DM , Wu XW , Ambrose CS . The test-negative design: validity, accuracy and precision of vaccine efficacy estimates compared to the gold standard of randomised placebo-controlled clinical trials . Euro Surveill 2013 ; 18 . 21. Belongia EA , Kieke BA , Donahue JG , et al. Effectiveness of inactivated influenza vaccines varied substantially with antigenic match from the 2004-2005 season to the 2006-2007 season . J Infect Dis 2009 ; 199 : 159 - 67 . 22. Foppa IM , Haber M , Ferdinands JM , Shay DK . The case test-negative design for studies of the effectiveness of influenza vaccine . Vaccine 2013 ; 31 : 3104 - 9 . 23. Rotavirus vaccines: an update . Wkly Epidemiol Rec 2009 ; 84 : 533 - 40 . 24. Boom JA , Tate JE , Sahni LC , et al. Effectiveness of pentavalent rotavirus vaccine in a large urban population in the United States . Pediatrics 2010 ; 125 : e199 - 207 . 25. Cortese MM , Leblanc J , White KE , et al. Leveraging state immunization information systems to measure the effectiveness of rotavirus vaccine . Pediatrics 2011 ; 128 : e1474 - 81 . 26. Staat MA , Payne DC , Donauer S , et al. Effectiveness of pentavalent rotavirus vaccine against severe disease . Pediatrics 2011 ; 128 : e267 - 75 . 27. Snelling TL , Andrews RM , Kirkwood CD , Culvenor S , Carapetis JR . Case-control evaluation of the effectiveness of the G1P[8] human rotavirus vaccine during an outbreak of rotavirus G2P[4] infection in central Australia . Clin Infect Dis 2011 ; 52 : 191 - 9 . 28. Correia JB , Patel MM , Nakagomi O , et al. Effectiveness of monovalent rotavirus vaccine (Rotarix) against severe diarrhea caused by serotypically unrelated G2P [4] strains in Brazil . J Infect Dis 2010 ; 201 : 363 - 9 . 29. Chang WC , Yen C , Wu FT , et al. Effectiveness of 2 rotavirus vaccines against rotavirus disease in Taiwanese infants . Pediatr Infect Dis J 2014 ; 33 : e81 - 6 . 30. Cortese MM , Immergluck LC , Held M , et al. Effectiveness of monovalent and pentavalent rotavirus vaccine . Pediatrics 2013 ; 132 : e25 - 33 . 31. Patel MM , Patzi M , Pastor D , et al. Effectiveness of monovalent rotavirus vaccine in Bolivia: case-control study . BMJ 2013 ; 346 : f3726 . 32. Groome MJ , Page N , Cortese MM , et al. Effectiveness of monovalent human rotavirus vaccine against admission to hospital for acute rotavirus diarrhoea in South African children: a case-control study . Lancet Infect Dis 2014 ; 14 : 1096 - 104 . 33. Bar-Zeev N , Kapanda L , Tate JE , et al. Effectiveness of a monovalent rotavirus vaccine in infants in Malawi after programmatic roll-out: an observational and casecontrol study . Lancet Infect Dis 2015 ; 15 : 422 - 8 . 34. Castilla J , Beristain X , Martinez-Artola V , et al. Effectiveness of rotavirus vaccines in preventing cases and hospitalizations due to rotavirus gastroenteritis in Navarre, Spain . Vaccine 2012 ; 30 : 539 - 43 . 35. Martinon-Torres F , Bouzon Alejandro M , Redondo Collazo L , et al. Effectiveness of rotavirus vaccination in Spain . Hum Vaccin 2011 ; 7 : 757 - 61 . 36. Muhsen K , Shulman L , Kasem E , et al. Effectiveness of rotavirus vaccines for prevention of rotavirus gastroenteritis-associated hospitalizations in Israel: a casecontrol study . Hum Vaccin 2010 ; 6 : 450 - 4 . 37. Payne DC , Boom JA , Staat MA , et al. Effectiveness of pentavalent and monovalent rotavirus vaccines in concurrent use among US children <5 years of age, 2009 - 2011 . Clin Infect Dis 2013 ; 57 : 13 - 20 . 38. Cotes-Cantillo K , Paternina-Caicedo A , Coronell-Rodriguez W , et al. Effectiveness of the monovalent rotavirus vaccine in Colombia: a case-control study . Vaccine 2014 ; 32 : 3035 - 40 . 39. Bennett A , Bar-Zeev N , Jere KC , et al. Determination of a viral load threshold to distinguish symptomatic versus asymptomatic rotavirus infection in a highdisease-burden African population . J Clin Microbiol 2015 ; 53: 1951 - 4 . 40. Tate JE , Mijatovic-Rustempasic S , Tam KI , et al. Comparison of 2 assays for diagnosing rotavirus and evaluating vaccine effectiveness in children with gastroenteritis . Emerg Infect Dis 2013 ; 19 : 1245 - 52 . 41. Phillips G , Lopman B , Tam CC , Iturriza-Gomara M , Brown D , Gray J . Diagnosing rotavirus A associated IID: using ELISA to identify a cut-off for real time RT-PCR . J Clin Virol 2009 ; 44 : 242 - 5 .


This is a preview of a remote PDF: http://cid.oxfordjournals.org/content/62/suppl_2/S106.full.pdf

Jacqueline E. Tate, Manish M. Patel, Margaret M. Cortese, Daniel C. Payne, Benjamin A. Lopman, Catherine Yen, Umesh D. Parashar. Use of Patients With Diarrhea Who Test Negative for Rotavirus as Controls to Estimate Rotavirus Vaccine Effectiveness Through Case-Control Studies, Clinical Infectious Diseases, 2016, S106-S114, DOI: 10.1093/cid/civ1014