Canada Communicable Disease Report

September 2008

Volume 34

Number 09

Monthly Report

Determinants for immunization coverage by age 2 in a population cohort in the Capital Health region, Edmonton, Alberta

J Zhang, MPH, MSc (1), A Ohinmaa, PhD (1,2), T-H Nguyen, MPH (3), L Mashinter, BSc, RN (4), A Hanrahan, MN (4), J Loewen, BN (4), W Vaudry, MD, FRCPC (5), P Jacobs, PhD (2,6)

1. University of Alberta, School of Public Health, Edmonton, Alberta, Canada
2. Institute of Health Economics, Edmonton, Alberta, Canada
3. Alberta Health and Wellness, Edmonton, Alberta, Canada
4. Capital Health, Edmonton, Alberta, Canada
5. University of Alberta, Department of Pediatrics, Edmonton, Alberta, Canada
6. University of Alberta, Department of Medicine, Edmonton, Alberta, Canada

Introduction

Routine childhood immunization is considered one of the most cost-effective public interventions protecting children from deadly infectious diseases such as diphtheria, measles and bacterial meningitis^(1,2) .

Alberta provides immunization free of charge for all children in the province. The Alberta immunization schedule is consistent with national recommendations⁽³⁾ , recommending five single and component vaccines in the first 2 years of life protecting children from 11 different diseases. The recommended schedule and number of doses needed for children up to 2 years of age depends on the age of the child when immunization was started and, in the case of varicella whether the child has a history of the chickenpox disease.

The Alberta goal for routine childhood immunization coverage is 97% of children will have received four doses of DTaP-IPV-Hib and Pneu-C-7 vaccine, three doses of Men-C vaccine and 98% will have received one dose of measles-mumps-rubelle (MMR) and varicella (Var) vaccine by 24 months of age.

Several studies have investigated determinants for immunization coverage by age 2 years. Parental concerns about safety and effectiveness of vaccines, risk of side-effects, and an unfounded fear that multiple vaccines might overwhelm or weaken a child's immune system have been highlighted as major factors associated with low immunization rates(4-7) . Socioeconomic status, particularly mother's education, income, family size and race are factors that have consistently been shown to influence whether a child receives a vaccine(8-10) . Parental beliefs on the protective effect of breast-feeding, complementary medicine and healthy living to build up immunity are other factors(5) . Ease of access to health care service, such as short clinic waits, and convenient health care location, have also been factors facilitating immunization rates(5,7) .

A better understanding of how different factors affect immunization rates could help public health managers to increase the immunization coverage in their area. The aim of this study is to identify demographic, socioeconomic and other factors that are associated with routine childhood immunization rates for children at 24 months of age in the Capital Health (CH) region of Alberta. Data were obtained from population-wide data sets, which were available due to Alberta's system of public provision of vaccines and its population-wide registry.

Methods

In the CH region, routine childhood immunizations are delivered through a universal public health program. Immunization data in CH is recorded and stored in a centralized database called Caseworks. For every child born in Alberta who will reside in the CH region, an individual record is created at birth including information from the child, mother, and the family. Information on immunization is updated at each contact with the public health system. Immunization records of children known to have moved out of the region are inactivated and are excluded from the calculations. Records of children who move into the region are initiated at the time of first contact with the public health system and the previous immunization events are entered in Caseworks. Children's other health information, including demographic and socioeconomic background, are stored in the provincial database, the Alberta Wellnet Electronic Health Record, which links clinic, hospital, pharmacy and other points of care to patient information.

The study population consisted of all children actively recorded in the Caseworks database in CH who were born between 1 July and 31 December, 2002 (n = 4,988). Information on the immunization events and other relevant data from the Caseworks were extracted for this cohort up to 31 December, 2004. This data was merged with Alberta Health and Wellness (AHW) data on the children's demographic and socioeconomic background using Personal Health Numbers (PHN) in both data sets. After AHW merged the datasets, the PHNs and other identifiers were removed from the data to make it anonymous and it was given to the researchers.

The variables used in this study include continuous variables: mother's age at delivery and number of sibling of the babies and binary variables: gender of the babies (male = 1); gestational age < 38 weeks; gestational age > 42 weeks; breastfed only or with formula; delivery method as caesarean section; mother's marital status as single; mother's marital status as common-law; First Nations status; location of hospital delivery (Edmonton = 1); location of permanent health care (Edmonton = 1); delivery by a midwife; recipients of welfare subsidy (if parents were recipients of any of child welfare, general welfare or native welfare, code equals to 1); and mother aged 30 to 40 years old with one other child already. Some interaction terms among independent variables were also considered.

Immunization rates were calculated for each of the antigens used for routine childhood immunization (DTaP-IPV-Hib, Men-C, Pneu-C-7, MMR, and Var). A specific computer program was developed that followed the recommended Alberta immunization schedule and calculated the individual immunization rate for every vaccine. The details of the calculation algorithm are published elsewhere⁽¹³⁾ . Each component of the DTaP-IPV-Hib vaccine was entered into the Caseworks database separately, and therefore evaluated separately.

A child was considered to have complete coverage for a vaccine if he/she received the correct number of doses with adequate spacing between doses, as specified in the schedule. If a child started a series of vaccines but did not complete the recommended number of doses or had inadequate spacing between doses, that child was considered partially complete. A child was considered not immunized if he/she had not received any doses of the vaccine. Finally, a child was considered not completed if he/she was considered either partially complete or not immunized. Applicable only to varicella immunization status, children who had the disease were not eligible for the vaccine under the Alberta guidelines, and were considered immune.

Data

Both the CH and AHW databases were checked for demographic information and socioeconomic background factors for the 4,988 children in the half-year cohort. Of the study cohort, demographic information for 521 infants and socioeconomic background factors for 434 infants were systematically missing in the database. The data was missing because the families of these infants moved out of or into CH during the follow-up period. Due to the missing data, the 521 infants were excluded from the modeling study. The remaining 4,467 children (90%) comprised the final study population.

Demographic and socioeconomic background information is presented in Table 1. Of the 4,988 infants 48.6% were female and 3.8% of the infants were identified as Treaty Indian. Age of the mothers at time of delivery ranged from 14 to 45 years with an average of 29.0 years (standard deviation 5.43). Of the total, 25.4% were born by caesarian-section. Midwifes performed only a small portion (0.8%) of the deliveries. Most (86.3%) of the deliveries occurred within the City of Edmonton area and the Public Health Centre for most infants was also most commonly (69%) located within Edmonton. Of the infants, 84.9% were fed by breast milk only or a combination of breast milk and formula (Table 1).

A majority of families did not receive any premium or welfare subsidy. The most common welfare subsidies were General Welfare received by 3.7%, and Native Welfare subsidy received by 3.8%.

Table 1. Descriptive statistics of background variables

Statistical Analysis

Logistic regression analysis was done to compare statistical associations between complete (value 1) and not complete (value 0) groups. Multicollinearity was checked among demographic and socioeconomic background factors by examining the correlations (continuous and ordinal variables) and associations (nominal variables) between independent variables, as well as by examining variance inflation factors⁽¹¹⁾ . A backward stepwise selection procedure was adopted to obtain determinants that contributed significantly to childhood immunization rates. The p-value of 0.05 was used to determine statistically significant associations. All statistical analysis was performed in Statistical Package for the Social Sciences (SPSS) version 14.0.

The study had received ethical approval through the University of Alberta Health Research Ethics Board.

Results

Table 2 displays routine recommended childhood immunization rates of 4,467 children in the cohort. The rates of completed immunization across the five single an d component vaccines varied from 85.5% (Pneu-C-7) to 94.2% (Men-C). Var and MMR had the highest not-immunized rates, 9.8% and 6.2%, respectively. Most of infants in the cohort started their multiple dose vaccine series, 97.9% for DTaP-IPV-Hib, 97.3% for Men-C and 96.5% for Pneu-C-7 vaccine. However, > 10% did not complete their vaccine series.

Table 2. Capital Health immunization coverage rates (at 2 years of age, n = 4,467)

Significant factors (Table 3) associated with completion of immunization rates indicated that as the mother ages, immunization rates increase, while the number of sibling increases, the immunization rates decrease. Infants who had a mother aged between 30 and 40 years with one other child were more likely to get immunized. Conversely, a mother's marital status as common-law and single-parent was negatively associated with fully completing the vaccine series. Midwives attending delivery was significantly negatively associated with the immunization rate. Infants delivered by caesarean-section showed a significant positive association with immunization rates. Being a recipient of general welfare or provincial health care premium subsidy was significantly associated with incomplete immunization series. Lastly, babies who had gestation age > 42 weeks were less likely to get fully immunized.

Table 3. ORs of statistically significant variables in the multivariate logistic regression model of five immunization programs: Complete vs. Not Complete (partially complete + not immunized)

All explanatory variables fi tted in the logistic regression were independent.

Discussion

Evaluating immunization rates is important for assessing coverage of immunization programs in the population groups and for developing strategic plans. The logistic regression model showed that mother's age, a mother aged between 30 and 40 years with one other child, services within city area, and delivery by caesarian-section were all associated with increased immunization rates. Living common-law or being a single-parent was negatively associated with fully completing the immunization as was delivery by midwives or having welfare subsidy.

Immunization rates calculated for infants in our sample who had all socio-demographic background information was slightly different from the immunization rates published previously for the whole Capital Health region population(13) . The percentage complete for all five routine immunization programs was slightly higher in this sub-population than those in the full population. Moving in or out of the region may be a key factor for this difference; unfortunately, it was not possible to determine with certainty the reason why inconsistent immunization rates were reported. Immunization rates obtained from the sub-population more precisely describe the true coverage rates in this half year cohort since these infants have more likely stayed in the region for the whole 2 year period.

There is a wide variation in coverage among the recommended vaccines. Complete coverage rates for MMR were higher than for DTaP-IPV-Hib. The number of doses required to complete a DTaP-IPV-Hib series is greater than that required to complete the MMR series. As the number of doses required to complete a series increases, the rate of complete immunizations decreases. There were 11.3% who were partially covered for DTaP-IPV-Hib at 2 years, but some of those children who were partially covered at that age eventually completed their immunization series. Furthermore, there was a discrepancy between the completion rate of Hib and the completion rate of DTaP-IPV doses. This may reflect a late onset of immunization schedules, which require fewer doses for Hib than for DTaP-IPV, making completion rates for Hib higher. Furthermore, the starting age required for a vaccine (Appendix 1) increases, the percentage not immunized increases as well. MMR and Var have higher not-immunized rates than other recommended vaccines.

The general welfare recipients had significantly lower immunization rates in two vaccines (DTaP-IPV and Hib). Since all First Nations children were welfare recipients it indicates that their immunization did not differ from other low income families. However, First Nations children had higher immunization rates in Var vaccine.

Moreover, our analysis shows that there is a negative association between delivery by a midwife and immunization rates. This may reflect a belief in this population that natural methods of health promotion are preferred to interventions like immunization (e.g. use of a midwife being more natural than the contrived, highly technical environment of the hospital). These beliefs are often incongruent with immunization.

In addition to the model that compared the complete coverage and not completed groups we also made three models comparing complete, partially complete and not immunized groups. The results from those models were in line with the logistic regression results of the complete and not complete populations.

The study findings were in line with most risk factors that were suggested in the literature. Low income families have been found to be significantly associated with low immunization rates in the United States(5,8) and Australia(7) . Gust et al.(8) and Haynes et al.(7) have found that marital status of the mother as common-law and single-parent was negatively associated with fully completing the immunization. Using midwives in the delivery were shown to have negative effect in Hamilton's study(5) conducted in New Zealand. However, in some countries that use mainly midwives in delivery, like Finland and Sweden, the coverage rates are nearly 100%(14) . Babies with fewer siblings had higher probability of being fully immunized than those who had more siblings in the family in Gust's work(8) . Contrast to our findings, Gust et al.(8) manifested that breast-feeding had a significant negative impact on coverage rates and Davis et al.(10) discovered premature babies had low coverage rates.

There are three limitations of this study. First, although the Caseworks database captures the vast majority of children living in the CH region, there is a possibility that some children were not included in the sample particularly if the families had moved to the region after the birth and the infants were not registered into the public health system. Missing demographic and socioeconomic background factors may also hide some potential impact on immunization rates. Second, the possibility of a cohort effect exists, although there is no reason to suspect that the cohort sampled in this study differs significantly from other children in the region. Lastly, other potential risk factors, such as mother's education were not available in the data and thus were not included in this study.

Using our data, we could not determine the reason why immunization rates varied among vaccines. Nevertheless, it appears that we cannot assume that immunization rates for one vaccine will be the same as for another even when administered concurrently. Immunization schedules are becoming more complex when new vaccines are added to routine immunization schedules. Our data suggest that there are differences in uptake and appropriate interventions are needed to reduce differences between vaccines and improve coverage in underserved population groups. In addition the study shows that socio-economic and demographic variables had consistently the same type of association with the immunization rate in different vaccines if they reached a statistically significant level.

Acknowledgements

Funding for this project was provided by the Institute of Health Economics, Research Coordination Committee. Specially, thank Capital Health region and Alberta Health and Wellness for providing the data.

Appendix 1. Schedule of childhood immunizations in Alberta

References

1. Alberta Health and Wellness. Alberta immunization manual, 2002.
2. National Advisory Committee on Immunization (NACI). Guidelines for childhood immunization practices. CCDR 1997;23(ACS-6):1-5.
3. World Health Organization. Immunization against disease of public health importance. http://www.who.int/ immunization_delivery/en/
4. Advisory Committee on Population Health and Health Security. National Immunization Strategy: Final report 2003. Ottawa: Minister of Health, 2004.
5. Samad L, Butler N, Peckham C et al. Incomplete immunisation uptake in infancy: Maternal reasons. Vaccine 2006;24:6823-29.
6. Hamilton M, Corwin P, Gower S et al. Why do parents choose not to immunize their children? The New Zealand Medical Journal 2004;117(1189):11-16.
7. Prislin R, Dyer JA, Blakely CH et al. Immunization status and sociodemographic characteristics: The mediating role of beliefs, attitudes, and perceived control. Am J Public Health 1998;88(12):1821-26.
8. Gust DA, Strine TW, Maurice E et al. Underimmunization among children: Effects of vaccine safety concern on immunization status. Pediatrics 2006;114:16-22.
9. Haynes K, Stone C. Predictors of incomplete immunization in Victorian children. Australian and New Zealand Journal of Public Health 2004;28(1):72-79.
10. Kahane SM, Watt JP, Newell K et al. Immunization levels and risk factors for low immunization coverage among private practices. Pediatrics 2000:105:73-83.
11. Davis RL, Rubanowice D, Shinefi eld HR et al. Immunization levels among premature and low-birth-weight infants and risk factors for delayed up-to-date immunization status. JAMA 1999;282:547-53.
12. Cox DR, Snell EJ. Analysis of binary data (2nd edition). London: Chapman & Hall, 1989.
13. Brown-Ogrodnick A, Hanrahan A, Loewen J et al. Immunization coverage by age 2 for fi ve recommended vaccines in the capital health region (Edmonton), Table 1. CCDR 2006;32(10):117-21.
14. World Health Organization. WHO: vaccine-preventable diseases: Monitoring system, 2006 global summary. http://www.who.int/vaccines-documents/GlobalSummary/GlobalSummary.pdf