Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Did the H1N1 Vaccine Reduce the Risk of Admission with Influenza and Pneumonia during the Pandemic?

  • Salaheddin M. Mahmud ,

    Salah.mahmud@gmail.com

    Affiliations Community Health Sciences, University of Manitoba, Winnipeg, Manitoba, Canada, Winnipeg Regional Health Authority, Winnipeg, Manitoba, Canada, Manitoba Health, Winnipeg, Manitoba, Canada

  • Songul Bozat-Emre,

    Affiliations Community Health Sciences, University of Manitoba, Winnipeg, Manitoba, Canada, Manitoba Health, Winnipeg, Manitoba, Canada

  • Gregory Hammond,

    Affiliations Medical Microbiology, University of Manitoba, Winnipeg, Manitoba, Canada, Medicine, University of Manitoba, Winnipeg, Manitoba, Canada

  • Lawrence Elliott,

    Affiliations Community Health Sciences, University of Manitoba, Winnipeg, Manitoba, Canada, Medical Microbiology, University of Manitoba, Winnipeg, Manitoba, Canada, Winnipeg Regional Health Authority, Winnipeg, Manitoba, Canada, Manitoba Health, Winnipeg, Manitoba, Canada

  • Paul Van Caeseele

    Affiliations Medical Microbiology, University of Manitoba, Winnipeg, Manitoba, Canada, Cadham Provincial Laboratory, Winnipeg, Manitoba, Canada, Manitoba Health, Winnipeg, Manitoba, Canada

Abstract

Background

The extent to which A(H1N1)pdm09 influenza vaccines prevented hospital admissions with pneumonia and influenza (P&I) during the 2009 pandemic remains poorly understood. We evaluated the effectiveness of the A(H1N1)pdm09 and seasonal influenza vaccines (TIV) used during the 2009 mass vaccination campaign in Manitoba (Canada) in preventing P&I hospitalization.

Methods

A population-based record-linkage nested case-control study. Cases (N = 1,812) were persons hospitalized with influenza (ICD-10:J09-J11) or pneumonia (ICD-10:J12-J18) during the study period. Age-, gender- and area of residence-matched controls (N = 7,915) were randomly sampled from Manitoba’s Population Registry. Information on receipt of A(H1N1)pdm09 vaccine and TIV was obtained from the Manitoba Immunization Monitoring System, a province-wide vaccine registry.

Results

Overall, the adjuvanted A(H1N1)pdm09 vaccine was 27% (95%CI 13–39%) effective against P&I hospitalization ≥ 14 days following administration. Effectiveness seemed lower among older (≥ 65 years) adults (10%; −16–30%), particularly when compared to under-5 children (58%; 30–75%). The number-needed-to-vaccinate to prevent 1 P&I admission was lowest among <4 year-olds (928) and ≥65 years (1,721). VE against hospitalization with laboratory-confirmed A(H1N1)pdm09 was 70% (39–85%) overall and (91%; 62–98%) ≥ 14 days following vaccination.

Discussion

Our data suggest that the adjuvanted A(H1N1)pdm09 vaccine was effective in preventing about 55–60% of P&I hospitalizations among children and younger adults who were at much higher risk of infection. Unfortunately, the vaccine was less effective among 65 or older adults. Despite that the vaccine still had a significant population-based impact especially among the very young (<5) and the older (≥ 65 years).

Key Points

In the recent pandemic, the A(H1N1)pdm09 vaccine was effective in preventing 55–60% of influenza and pneumonia hospitalizations among children and younger adults. Although less effective among ≥65 adults(~10%), the vaccine benefited this group the most as measured by the number-needed-to-vaccinate.

Introduction

Worldwide, more than 26 different monovalent A(H1N1)pdm09 vaccines were used during the 2009 influenza pandemic to control the spread of infection and reduce disease burden[1]. In Manitoba (Canada), we found that both adjuvanted and non-adjuvanted vaccines were effective in preventing laboratory-confirmed A(H1N1)pdm09 infections[2], and similar findings were reported for other jurisdictions.[3] However, the extent to which vaccination prevented more clinically meaningful outcomes, such as acute hospital admission due to influenza and pneumonia (P&I), is not known.

Rates of laboratory-confirmed influenza can vary significantly between jurisdictions and with epidemic phase in the same jurisdiction as a function of clinical guidelines, epidemic phase and laboratory workload. By comparison, decisions for hospitalization are less discretionary, and as a result hospitalization rates are a robust indicator of epidemic severity.[4] Arguably, avoiding hospitalization is a more relevant outcome for influenza control efforts than the mere reduction in numbers of laboratory-confirmed infections.

We evaluated the effectiveness of the A(H1N1)pdm09 and seasonal influenza vaccines (TIV) used during the 2009 mass vaccination campaign in Manitoba (Canada) in preventing P&I hospitalization using a case-control design and data from Manitoba’s provincial administrative and laboratory databases. We also assessed vaccine effectiveness (VE) for different age groups and among high-risk populations, e.g., immunocompromised persons.

Methods

Design

We conducted a nested case-control study using de-identified records obtained by linking the electronic database of the Manitoba Immunization Monitoring System (MIMS) with the Hospital Separation Abstract database and other Manitoba Health (MH) administrative databases, housed at the Manitoba Centre for Health Policy. MH provides publicly-funded universal healthcare insurance to virtually all of Manitoba’s 1.2 million residents regardless of age or income [5]. Insured services include hospital, physician and preventive services (e.g., immunizations). For administrative purposes, MH maintains several centralized electronic databases that can be linked using a unique health services number. The study was approved by the Research Ethics Board of the University of Manitoba and the governmental Health Information Privacy Committee.

Definition of cases and controls

Anyone 6 months or older who was registered with MH during the study period was eligible for inclusion in the study. The study spanned the period from November 2, 2009 (1 week after the start of mass immunization in Manitoba) to February 7, 2010, two weeks following the last reported A(H1N1)pdm09 case in the province.

Any eligible person who was admitted for ≥24 hours to a hospital in Manitoba with a diagnosis of influenza (ICD-10: J09-J11) or pneumonia (ICD-10: J12-J18) during the study period was included in the cases group. Cases were identified using the Hospital Separation Abstract database, which, since 1971, has recorded all services provided by hospitals in the province, including admissions and day surgeries [6]. The data collected include clinical information such as admission and discharge dates and up to 25 diagnoses and 20 services or procedures, coded using the International Classification of Diseases (ICD-10-CA[7]) and the Canadian Classification of Health Interventions (CCI)[8].

Using risk-set sampling, we matched each case to five controls (persons who have not been admitted to hospital by the index date) who were of the same age, gender and area of residence. The date of admission was considered as the “index date” for cases and for their matched controls. Controls were randomly selected from MH’s Population Registry, a continuously updated database that tracks dates and reasons for the initiation and termination of coverage (e.g., death/migration) for all insured persons.

Determination of vaccination status

For all cases and controls, information on the receipt of the pandemic, seasonal influenza and pneumococcal vaccines during and before the 2009/10 season was obtained from MIMS, a population-based province-wide registry of virtually all vaccines administered to Manitoba residents since 1988 [9]. Information, including vaccine type and date of vaccination, is captured either through direct data entry for vaccines administered by public health staff (who administered most influenza vaccines during the pandemic) or using physician claims data for vaccines administered by physicians [10].

In Manitoba, most pandemic vaccines were administered during a mass immunization campaign that began in October 26, 2009 (Week 43), just one week before the peak of the second pandemic wave [2]. Initially, the Canadian-manufactured Arepanrix® (GlaxoSmithKline), an AS03-adjuvanted split virion monovalent vaccine, was used to vaccinate adults and children over 6 months of age. Later on, two nonadjuvanted vaccines, from GlaxoSmithKline and CSL Limited, were offered to pregnant women and children 10 years or older [11]. All vaccines contained the A(H1N1)pdm09 hemagglutinin antigen derived from the influenza A/California/7/2009 strain recommended by the WHO. A single vaccine dose (15μg/0.5 ml) was recommended for those aged >9 years and 2 half doses given 21 days apart were recommended for children 6 months-9 years old.

All vaccines were offered free of charge, but due to limited supply at campaign start priority was given to certain groups including health care workers, Aboriginal persons, residents of remote communities, pregnant women, 6–60 months-old children, persons under 65 years with chronic medical conditions and all immunocompromised persons. On November 18, 2009, vaccines became available to the whole population [2].

Manitoba’s routine immunization schedule includes seasonal trivalent inactivated influenza vaccines (TIV)—during the study period these were Fluviral® (GlaxoSmithKline) and Vaxigrip® (Sanofi Pasteur)—and several polysaccharide and conjugate pneumococcal (PVs).

Potential confounders

Individuals were assigned to a neighbourhood of residence (neighbourhood clusters within the capital city of Winnipeg and regional health authorities in the rest of the province) based on their postal code as recorded in MH’s Population Registry. Household income quintiles, measured at the level of Census Dissemination areas, were determined using 2006 Canadian census data. Information on pregnancy, comorbidities, propensity to seek health care (measured as the number of hospital and family physician visits in the previous 5 years) was obtained from the Hospital Separation and Physician Claims databases. Previously validated algorithms were used to identify various chronic diseases and other indications for vaccination [1215]. In addition, Charlson comorbidity scores were calculated using an algorithm validated for administrative databases[16]. Information on the use of antivirals and other medications was obtained from the Drug Program Information Network database, the comprehensive database of all out-of-hospital prescriptions dispensed in Manitoba since 1995 [17].

Influenza testing results were obtained from the database of Cadham Provincial laboratory (CPL), the province’s only public health laboratory. During the study period, influenza testing in Manitoba was completed at CPL using a real-time multiplex reverse-transcription polymerase chain reaction (RT-PCR) assay developed by the National Microbiology Laboratory [18].

Statistical analysis

In the primary analysis, we used conditional logistic regression models to estimate the odds ratio (OR) for the association between the receipt of the adjuvanted A(H1N1)pdm09 vaccine and P&J hospitalization while adjusting for matching and confounding variables. Models were adjusted for income, comorbidity, receipt of seasonal and pneumococcal vaccines, use of neuraminidase inhibitors, frequency of contact with healthcare providers and belonging to a vaccine priority group. VE was estimated as (1-OR) x 100. Similar but separate models were used to estimate the VE of nonadjuvanted pandemic, seasonal and pneumococcal vaccines.

In primed healthy adults, the peak serum antibody levels are typically observed >2 weeks after vaccination [19]. To account for differences in effectiveness by time since vaccination, we classified vaccinated individuals into three groups depending on whether vaccination occurred 1–6, 7–13, or ≥ 14 days before the index date, and contrasted the odds of A(H1N1)pdm09 infection in each group with the odds of infection among the unvaccinated.

In addition, we repeated the adjuvanted A(H1N1)pdm09 vaccine analyses after stratifying by age group, place of residence, epidemic phase (admission before and after the peak), presence of high-risk conditions and belonging to a vaccine priority group. We also assessed for possible effect modification by use of seasonal and pneumococcal vaccines. The statistical significance of adding the interaction terms was assessed using the likelihood ratio test [20].

Results

A total of 1,812 persons met the case definition and were matched to 9,060 controls. About 54% of cases were 65 years or older at diagnosis, and 39% resided in the poorest parts of the province (Table 1). As expected, cases tended to be generally sicker than controls with higher Charlson comorbidity scores (average score of 3 compared to 0 for controls), higher overall chance of having ≥1 diagnosed chronic medical conditions (73% compared to 39% among controls) and more frequent hospitalizations and physician consultations. As a result, cases were also more likely to belong to A(H1N1)pdm09 or TIV priority groups (Table 1).

thumbnail
Table 1. Demographic and clinical characteristics of cases and controls.

https://doi.org/10.1371/journal.pone.0142754.t001

A similar proportion of cases and controls received the adjuvanted A(H1N1)pdm09 vaccine (about 35%; Table 2), comparable to estimates of vaccine coverage for the entire population from MH (37% [11]) and the 2010 Canadian Community Health Survey (37% [95%CI: 33–41%])[21]. However, only about 28% of both groups received the vaccine ≥14 days before the index date corresponding to a VE estimate (adjusted for matching covariates) of 0.2% (95%CI: −13–12%). After adjusting for important confounders, the corresponding VE estimate was 27% (13–39%). Because of small numbers it was not possible to reliably estimate VE of the nonadjuvanted vaccines. In adjusted models, neither receiving the TIV in the 2008/09 or 2009/10 seasons nor receiving a pneumococcal vaccine at any point before the index date had a significant influence on the risk of hospitalization during the study period (Table 2).

thumbnail
Table 2. Estimates of the effectiveness (VE) of pandemic, seasonal influenza and pneumococcal vaccine against hospitalization due to influenza or pneumonia*.

https://doi.org/10.1371/journal.pone.0142754.t002

In subgroup analyses (Table 3), there was evidence that the adjuvanted A(H1N1)pdm09 vaccine was less effective in preventing P&I hospitalization among 65 or older adults (10% [−16–30%]) compared to younger age groups, particularly under-5 children (58% [30–75%]). The A(H1N1)pdm09 vaccine was also less effective among those with chronic diseases (14% [−11–34%]; Pinteraction = 0.044) and among those who received the 2008/09 TIV (21% [1–37%]), although the latter finding was not statistically significant (Pinteraction = 0.3). On the other hand, there was a statistically significant difference (Pinteraction = 0.016) in A(H1N1)pdm09 VE between those who also received the 2009/10 TIV (12% [−21–36%]) and those who did not (46% [28–60%]).

thumbnail
Table 3. Estimates of the effectiveness (VE) of the adjuvanted A(H1N1)pdm09 vaccine (when received ≥14 days before the index date) against hospitalization due to influenza or pneumonia by certain demographic and clinical characteristics.

https://doi.org/10.1371/journal.pone.0142754.t003

About 8% of the cases and none of the controls tested positive for A(H1N1)pdm09 (Table 4), which is not surprising given provincial guidelines discouraging viral testing unless the patient is very ill. Only 6 controls were tested within 2 weeks of the index date compared to 921 cases. In analyses limited to cases hospitalized with laboratory-confirmed A(H1N1)pdm09, 13% of the cases received an adjuvanted A(H1N1)pdm09 vaccine compared to 25% of the controls (Table 4), corresponding to a VE estimate (adjusted for matching covariates) of 59% (30–76%). The corresponding fully adjusted estimate was 70% (39–85%). VE among those who were vaccinated ≥ 14 days before the index date was higher (91%; 62–98%) than among those who were vaccinated <7 days (46%; −48–81%) or 7–13 days before the index date (61%; −11–86%). Despite small numbers, estimates of VE of the nonadjuvanted A(H1N1)pdm09 vaccine were comparable. On the other hand, there was no evidence that the 2008/09 or the 2009/10 TIVs reduced the risk of hospitalization with laboratory-confirmed A(H1N1)pdm09.

thumbnail
Table 4. Estimates of the effectiveness (VE) of pandemic, seasonal influenza and pneumococcal vaccine against hospitalization with laboratory-confirmed influenza*.

https://doi.org/10.1371/journal.pone.0142754.t004

Discussion

Our data suggest that the adjuvanted A(H1N1)pdm09 vaccine prevented about 55 to 60% of P&I hospitalizations among children and younger adults and a much lower percentage (10–15%) of hospitalizations among 65 or older adults and among those with pre-existing chronic diseases (14%). The vaccine was also effective (70% on average) in preventing hospitalizations with laboratory-confirmed A(H1N1)pdm09 with higher levels of protection achieved >14 days after vaccination.

Excellent immune responses following even one dose of the monovalent split/subunit inactivated pandemic vaccines were documented in several immunogenicity trials [22, 23]. In post-marketing studies, the vaccines were also effective in preventing laboratory-confirmed A(H1N1)pdm09 infections during the pandemic. In a systematic review that was limited to 5 observational studies which met stringent quality criteria, the median VE of monovalent A(H1N1)pdm09 vaccines was about 69% [3]. Similar estimates were observed in studies not included in the review [2, 24].

To our knowledge, this is the first published study to evaluate the effectiveness of the A(H1N1)pdm09 vaccine against admission with P&I during the pandemic. On the other hand, VE against hospitalization with A(H1N1)pdm09 infection was examined in few studies. Using data from a Scottish general practice sentinel surveillance network, Simpson et al reported that the adjuvanted vaccine was 95% (76–100%) effective against laboratory-confirmed A(H1N1)pdm09 [25]. Emborg, et al. estimated, using Danish health databases, that the adjuvanted monovalent vaccine was 44% (−19%–73%) effective in preventing hospitalization with laboratory-confirmed A(H1N1)pdm09 infection among younger (<65) chronically ill people [26]. Steens, et al. found that among persons with underlying medical conditions or ≥ 60 years of age, a single dose of the MF-59 adjuvanted A(H1N1)pdm09 vaccine had VE of 19% (−28–49%) [27].

Higher estimates of VE against P&I hospitalization among children and young adults in our study may reflect the greater contribution of A(H1N1)pdm09 to P&I hospitalization during the pandemic among this age group compared to older adults who generally were less likely to become infected [28, 29]. Also, in a previous analysis from Manitoba we found that the vaccine was more effective in preventing A(H1N1)pdm09 infection in children compared to older adults.[2] This is also consistent with studies that examined VE against hospitalization with A(H1N1)pdm09 among children. In a Quebec study, VE of a single pediatric dose of the same AS03-adjuvanted vaccine that was used in Manitoba, was 85% (61–94%) among 6 month-9 year olds and slightly lower in 5–9 year-olds at 79% (−31–96%) [30]. In a smaller study conducted in New York, a single dose of the nonadjuvanted vaccine was 82% (0–100%) effective in in preventing hospitalization ≥ 14 days after vaccination in children aged 3–9 years [31]. Based on this evidence, it is reasonable to conclude that that the A(H1N1)pdm09 was effective in protecting younger children against hospitalization with P&I during the pandemic.

For older adults, our A(H1N1)pdm09 VE estimates were lower than estimates obtained from observational studies of the effectiveness of TIVs against P&I hospitalization during non-pandemic seasons. In a comprehensive Cochrane review,[32] 8 such studies had a pooled estimate of 26% (12%-38%) during seasons when the vaccine was well matched to the circulating strain. However, our estimates are more in line with the studies that attempted to control for confounding by the “healthy vaccinee effect” (seniors who get vaccinated are on average healthier than those who do not) which produced estimates between 8 and 14%[3335].

Despite lower VE and lower incidence of A(H1N1)pdm09 among the elderly, our data suggest that only very young children (0–4 years) had a lower number needed to vaccinate (NNV = 928) to prevent one hospital admission for P&I. The NNV among 65 or older was 1,721 compared to 2,273 among 45–64 olds and 7,598 among 25–44 olds. This reflects the much lower overall rate of P&I admissions among the latter two groups (4–5/10,000) compared to 59/10,000 among the ≥65 age group and 19/10,000 among 0–4 year olds. Ignoring potential indirect benefits due to herd immunity, these figures suggest that vaccinating the <5 and ≥65 age groups was more cost-effective than vaccinating other groups.

In our study, the TIVs did not appear to protect against P&I admissions overall or due to A(H1N1)pdm09. This makes sense because in Manitoba the pandemic strain almost entirely replaced previously circulating influenza strains during the 2009/10 season and there was no evidence of significant cross-protective response in previous studies.[36] We observed that the A(H1N1)pdm09 vaccine was less effective among those who also received the 2009/10 TIV. It is unclear whether this reflects a biological effect or confounding by indication because the TIV is indicated to persons at higher risk of developing severe influenza illness and its complications [37]. The issue of whether TIV use before or during the pandemic increased the incidence or severity of A(H1N1)pdm09 remains controversial [38].

Strengths and Limitations

Because of its population-based design and the availability of accurate automated hospital admission and vaccination records, this study is less susceptible to selection and recall biases that commonly afflict conventional case-control studies [10]. Because cases and controls were identified on the basis of comprehensive hospital records with high standards of coding practices,[13] misclassification of hospitalization status is also not a major concern in this study.

We used proxies for access to health care (e.g., frequency of physician encounters) to adjust for factors associated with the likelihood of admission and influenza testing. We also adjusted for confounding by several vaccine indications such as immune status and pre-existing health conditions using information obtained from administrative databases. The completeness and accuracy of the MH database are well established [6, 39], and these databases have been used extensively in studies of post-marketing surveillance of drugs and vaccines [2]. However, it is still possible that some variables were measured with error, which could result in residual confounding. We did not have information on functional status. However, the observed protective effects of A(H1N1)pdm09 vaccination are unlikely to be due to the healthy vaccinee effect, because vaccination was targeted at the higher-risk, and generally less healthy, groups, and we adjusted for vaccine indications in our models. Also, we did not observe any protective effects for the TIVs which are presumably subject to the same bias.

Confounding by herd immunity following the summer wave of the pandemic is also unlikely explanation for our findings, because we adjusted for area of residence in all models and because VE estimates were similar for northern and southern communities despite significant differences in A(H1N1)pdm09 seroprevalence between these communities at the end of the summer wave of the pandemic [28].

Conclusions

Our data suggest that the adjuvanted A(H1N1)pdm09 vaccine was effective in preventing about 55–60% of P&I hospitalizations among children and younger adults who were at much higher risk of infection. Unfortunately, the vaccine was less effective among 65 or older adults. Despite that the vaccine still had a significant population-based impact especially among the very young (<5) and the older (≥ 65 years).

Acknowledgments

The authors acknowledge the Manitoba Centre for Health Policy for use of data contained in the Population Health Research Data Repository under project “Effectiveness of GSK Pandemic H1N1 Influenza Vaccines in Preventing Hospitalizations for Influenza Pneumonia” (HIPC# 2011-2012-30). The results and conclusions are those of the authors and no official endorsement by the Manitoba Centre for Health Policy, Manitoba Health, Healthy Living, and Seniors, or other data providers is intended or should be inferred.

Author Contributions

Conceived and designed the experiments: SMM GH LE PVC. Analyzed the data: SBE. Wrote the paper: SMM GH LE PVC.

References

  1. 1. Girard MP, Tam JS, Assossou OM, Kieny MP. The 2009 A (H1N1) influenza virus pandemic: A review. Vaccine 2010; 28(31): 4895–902. pmid:20553769
  2. 2. Mahmud S, Hammond G, Elliott L, Hilderman T, Kurbis C, Caetano P, et al. Effectiveness of the pandemic H1N1 influenza vaccines against laboratory-confirmed H1N1 infections: Population-based case–control study. Vaccine 2011; 29(45): 7975–81. pmid:21884747
  3. 3. Osterholm MT, Kelley NS, Sommer A, Belongia EA. Efficacy and effectiveness of influenza vaccines: a systematic review and meta-analysis. The Lancet Infectious Diseases 2012; 12(1): 36–44. pmid:22032844
  4. 4. Jhung MA, Swerdlow D, Olsen SJ, Jernigan D, Biggerstaff M, Kamimoto L, et al. Epidemiology of 2009 Pandemic Influenza A (H1N1) in the United States. Clin Infect Dis 2011; 52(suppl 1): S13–S26. pmid:21342884
  5. 5. Singh H, Mahmud SM, Turner D, Xue L, Demers AA, Bernstein CN. Long-term use of statins and risk of colorectal cancer: a population-based study. Am J Gastroenterol 2009; 104(12): 3015–23. pmid:19809413
  6. 6. Roos L, Mustard C, Nicol J, McLerran DF, Malenka DJ, Young TK, et al. Registries and administrative data: organization and accuracy. Med Care 1993; 31(3): 201–12. pmid:8450678
  7. 7. Canadian Institute for Health Information. ICD-10-CA International statistical classification of diseases and related health problems, Tenth Revision: Ottawa, Ontario, Canada, 2009.
  8. 8. Canadian Institute for Health Information. Canadian Classification of Health Interventions: Ottawa, Ontario, Canada, 2006.
  9. 9. Roberts J, Roos L, Poffenroth L, Poffenroth LA, Hassard TH, Bebchuk JD, Carter AO, et al. Surveillance of vaccine-related adverse events in the first year of life: a Manitoba cohort study. J Clin Epidemiol 1996; 49(1): 51. pmid:8598511
  10. 10. Roberts J, Poffenroth L, Roos L, Bebchuk J, Carter A. Monitoring childhood immunizations: a Canadian approach. Am J Public Health 1994; 84(10): 1666. pmid:7943493
  11. 11. Manitoba H. H1N1 Flu in Manitoba: Manitoba's Response, Lessons Learned. Winnipeg, MB, CAN: Manitoba Health, 2010.
  12. 12. Elixhauser A, Steiner C, Harris DR, Coffey RM. Comorbidity measures for use with administrative data. Med Care 1998; 36(1): 8–27. pmid:9431328
  13. 13. Lix L, Yogendran M, Burchill C, Metge C, McKeen N, Moore D, et al. Defining and Validating Chronic Diseases: An Administrative Data Approach. Winnipeg: Manitoba Centre for Health Policy, 2006.
  14. 14. Dublin S, Jackson ML, Nelson JC, Weiss NS, Larson EB, Jackson LA. Statin use and risk of community acquired pneumonia in older people: population based case-control study. BMJ 2009; 338: b2137. pmid:19531550
  15. 15. Hardy JR, Holford TR, Hall GC, Bracken MB. Strategies for identifying pregnancies in the automated medical records of the General Practice Research Database. Pharmacoepidemiol Drug Saf 2004; 13(11): 749–59. pmid:15386720
  16. 16. Quan H, Parsons GA, Ghali WA. Validity of information on comorbidity derived rom ICD-9-CCM administrative data. Med Care 2002; 40(8): 675–85. pmid:12187181
  17. 17. Kozyrskyj AL, Mustard CA. Validation of an electronic, population-based prescription database. Ann Pharmacother 1998; 32(11): 1152–7. pmid:9825079
  18. 18. LeBlanc J, Li Y, Bastien N, Forward K, Davidson R, Hatchette T. Switching gears for an influenza pandemic: validation of a duplex reverse transcriptase PCR assay for simultaneous detection and confirmatory identification of pandemic (H1N1) 2009 influenza virus. J Clin Microbiol 2009; 47(12): 3805. pmid:19794033
  19. 19. Bridges CB, Katz JM, Levandowski RA, Cox NJ. Influenza Vaccine (Inactivated). In: Plotkin S, Orenstein W, Offit P. Vaccines: Elsevier, 2008:259–90.
  20. 20. Clayton D, Hills M. Statistical models in epidemiology. Oxford; New York: Oxford University Press, 1993.
  21. 21. Gilmour H, Hofmann N. H1N1 vaccination. Health Rep Vol. 21: Statistics Canada, Catalogue no. 82-003-XPE, 2010
  22. 22. Yin JK, Khandaker G, Rashid H, Heron L, Ridda I, Booy R. Immunogenicity and safety of pandemic influenza A (H1N1) 2009 vaccine: systematic review and meta-analysis. Influenza Other Respir Viruses 2011; 5(5): 299–305. pmid:21668694
  23. 23. Manzoli L, De Vito C, Salanti G, D'Addario M, Villari P, Ioannidis JPA. Meta-Analysis of the Immunogenicity and Tolerability of Pandemic Influenza A 2009 (H1N1) Vaccines. PLoS ONE 2011; 6(9): e24384. pmid:21915319
  24. 24. Manzoli L, Ioannidis JP, Flacco ME, De Vito C, Villari P. Effectiveness and harms of seasonal and pandemic influenza vaccines in children, adults and elderly: a critical review and re-analysis of 15 meta-analyses. Hum Vaccin Immunother 2012; 8(7): 851–62. pmid:22777099
  25. 25. Simpson CR, Ritchie LD, Robertson C, Sheikh A, McMenamin J. Vaccine effectiveness in pandemic influenza—primary care reporting (VIPER): an observational study to assess the effectiveness of the pandemic influenza A (H1N1)v vaccine. Health Technol Assess 2010; 14(34): 313–46. pmid:20630126
  26. 26. Emborg HD, Krause TG, Hviid A, Simonsen J, Molbak K. Effectiveness of vaccine against pandemic influenza A/H1N1 among people with underlying chronic diseases: cohort study, Denmark, 2009–10. BMJ 2012; 344: d7901.
  27. 27. Steens A, Wijnans EG, Dieleman JP, Sturkenboom MC, van der Sande MA, van der Hoek W. Effectiveness of a MF-59-adjuvanted pandemic influenza vaccine to prevent 2009 A/H1N1 influenza-related hospitalisation; a matched case-control study. BMC Infect Dis 2011; 11: 196. pmid:21767348
  28. 28. Mahmud SM, Becker M, Keynan Y, Elliot L, Thompson LH, Fowke K, et al. Estimated cumulative incidence of pandemic (H1N1) influenza among pregnant women during the first wave of the 2009 pandemic. CMAJ 2010; 182(14): 1522–4. pmid:20823167
  29. 29. Thompson LH, Mahmud SM, Keynan Y, Blanchard JF, Slater J, Dawood M, et al. Serological survey of the novel influenza A H1N1 in inner city Winnipeg, Manitoba, 2009. Canadian Journal of Infectious Diseases and Medical Microbiology 2012; 23(2): 65–70. pmid:23730311
  30. 30. Gilca R, Deceuninck G, De Serres G, Boulianne N, Sauvageau C, Quach C, et al. Effectiveness of pandemic H1N1 vaccine against influenza-related hospitalization in children. Pediatrics 2011; 128(5): e1084–91. pmid:21987710
  31. 31. Hadler JL, Baker TN, Papadouka V, France AM, Zimmerman C, Livingston KA, et al. Effectiveness of 1 dose of 2009 influenza A (H1N1) vaccine at preventing hospitalization with pandemic H1N1 influenza in children aged 7 months-9 years. J Infect Dis 2012; 206(1): 49–55. pmid:22551808
  32. 32. Jefferson T, Di Pietrantonj C, Al-Ansary LA, Ferroni E, Thorning S, Thomas RE. Vaccines for preventing influenza in the elderly. The Cochrane database of systematic reviews 2010; (2): CD004876. pmid:20166072
  33. 33. Baxter R, Lee J, Fireman B. Evidence of Bias in Studies of Influenza Vaccine Effectiveness in Elderly Patients. J Infect Dis 2010; 201(2): 186–9. pmid:19995265
  34. 34. Fireman B, Lee J, Lewis N, Bembom O, van der Laan M, Baxter R. Influenza Vaccination and Mortality: Differentiating Vaccine Effects From Bias. Am J Epidemiol 2009; 170(5): 650–6. pmid:19625341
  35. 35. Wong K, Campitelli MA, Stukel TA, Kwong JC. Estimating influenza vaccine effectiveness in community-dwelling elderly patients using the instrumental variable analysis method. Arch Intern Med 2012; 172(6): 484–91. pmid:22371873
  36. 36. Mahmud SM, Van Caeseele P, Hammond G, Kurbis C, Hilderman T, Elliott L. No Association between 2008–09 Influenza Vaccine and Influenza A (H1N1) pdm09 Virus Infection, Manitoba, Canada, 2009. Emerg Infect Dis 2012; 18(5): 801. pmid:22516189
  37. 37. Psaty BM, Koepsell TD, Lin D, Weiss NS, Siscovick DS, Rosendaal FR, et al. Assessment and control for confounding by indication in observational studies. J Am Geriatr Soc 1999; 47(6): 749–54. pmid:10366179
  38. 38. Viboud C, Simonsen L. Does Seasonal Influenza Vaccination Increase the Risk of Illness with the 2009 A/H1N1 Pandemic Virus? PLoS Med 2010; 7(4): e1000259. pmid:20386730
  39. 39. Young T, Roos L, Gelskey D. Estimating the burden of disease: comparing administrative data and self-reports. Med Care 1997; 35(9): 932–47. pmid:9298082