Impact of the Change in Polio Vaccination Schedule on Immunization Coverage Rates: A Study in Two Large Health Maintenance Organizations
Objective. In January 1997, one of the most significant changes to United States vaccine policy occurred when polio immunization guidelines changed to recommend a schedule containing inactivated polio vaccine (IPV). There were concerns that parent or physician reluctance to accept IPV into the routine childhood immunization schedule would lead to lowered coverage. We determined whether adoption of an IPV schedule had a negative impact on immunization coverage.
Design. A cohort study of 2 large health maintenance organizations (HMOs), Group Health Cooperative and Kaiser Permanente Northern California, was conducted. For analysis at 12 months of age, children who were born between October 1, 1996, and December 31, 1997, and were commercially insured and covered by Medicaid were continuously enrolled; for analysis at 24 months of age, children who were born between October 1, 1996, and June 30, 1997, and were commercially insured and covered by Medicaid were continuously enrolled. The 3 measures of immunization status at 12 and 24 months of age were up-to-date status, cumulative time spent up-to-date, and the number of missed opportunity visits.
Results. At both HMOs, children who received IPV were as likely to be up to date at 12 months as were children who received oral poliovirus vaccine (OPV), whereas at Group Health, children who received IPV were slightly more likely to be up to date at 24 months (relative risk: 1.12; 95% confidence interval [CI]: 1.05, 1.19). These findings were consistent for children who were covered by Medicaid. At Kaiser Permanente, children who received IPV spent ∼3 fewer days up to date in the first year of life, but this difference did not persist at 2 years of age. At Group Health, children who received IPV were no different from those who received OPV in terms of days spent up to date by 1 or 2 years of age. At Group Health, children who received IPV were less likely to have a missed opportunity by 12 months old (odds ratio [OR] 0.46; 95% CI: 0.31, 0.70), but this finding did not persist at 24 months of age. At Kaiser Permanente, children who received IPV were more likely to have a missed opportunity by 12 months (OR 2.06; 95% CI: 1.84, 2.30), and 24 months of age (OR 1.50; 95% CI: 1.36, 1.67).
Conclusions. The changeover from an all-OPV schedule to one containing IPV had little if any negative impact on vaccine coverage. Use of IPV was associated with a small increase in the likelihood of being up to date at 2 years of age at one of the HMOs and conversely was associated with a small increase in the likelihood of having a missed-opportunity visit in the other HMO.polio, poliomyelitis, vaccination, immunization coverage.
- ACIP =
- Advisory Committee on Immunization Practices •
- IPV =
- inactivated polio vaccine •
- OPV =
- oral poliovirus vaccine •
- HMO =
- health maintenance organization •
- GHC =
- Group Health Cooperative of Puget Sound •
- KPNC =
- Kaiser Permanente of Northern California •
- DTP/DTaP =
- diphtheria and tetanus toxoids and pertussis/acellular pertussis •
- HIB =
- Haemophilus influenzae type b •
- MMR =
- measles-mumps-rubella •
- RR =
- relative risk •
- CI =
- confidence interval •
- OR =
- odds ratio
In January 1997, the Advisory Committee on Immunization Practices (ACIP) recommended adoption of a sequential inactivated poliovirus vaccine–oral poliovirus vaccine (IPV-OPV) immunization schedule.1 The schedule of IPV at 2 months and 4 months of age followed by OPV at 12 to 18 months and again at 4 to 6 years was intended to minimize the risk for vaccine-associated paralytic polio while maintaining population immunity to the potential introduction of wild-type poliovirus.1–5 This switch to an IPV-containing schedule represented one of the most significant changes to United States vaccine policy.1–3 However, there was considerable concern that parents might not want their children to receive numerous simultaneous shots, that physicians might be reluctant to administer multiple injections at a single visit, and that the change to a sequential IPV-OPV vaccination schedule would lead to reduced vaccination coverage of children.1,2,5–14
In a previous report, we suggested that this schedule change had little immediate negative effect on vaccine coverage.15 We now report additional findings that assess more completely whether this change resulted in reduced or delayed vaccination coverage. To do this, we followed children in 2 large west coast health maintenance organizations (HMOs) and evaluated a number of different measures of their immunization coverage at 1 and 2 years of age.
This study included children who were enrolled at Group Health Cooperative of Puget Sound (GHC), a Seattle-based HMO with ∼530 000 members, and Kaiser Permanente of Northern California (KPNC), an Oakland-based HMO with ∼2.8 million members. Both sites have automated immunization tracking systems that allow for assessment of vaccination coverage by region, clinic, and individual patient.
Beginning in April 1997, GHC formally adopted the ACIP guidelines for the sequential IPV schedule as an option for physicians and families. Replacement of OPV with IPV occurred rapidly at GHC (Fig 1). In the year after introduction of IPV, the proportion of GHC children who received their first polio vaccination as IPV increased from 18% to 96%.
This changeover occurred less rapidly at KPNC. Each of its 17 medical centers made a local decision about whether and when to adopt an IPV schedule. In the year after introduction of IPV, the proportion of KPNC children who received their first polio vaccination as IPV increased from 13% to 45% (Fig 1).
Polio Vaccination Schedule
We defined children to be on an IPV schedule if they received their first polio vaccination as IPV; children whose first polio vaccine was OPV were defined to be on the OPV schedule. Fewer than 5% of GHC children and 3% of KPNC children received a mixed schedule (OPV at 2 months followed by IPV at 4 months of age, or IPV at 2 months followed by OPV at 4 months of age). Thus, this definition separated children who received an all-OPV schedule from those who received the all-IPV or IPV-OPV sequential schedule.
We studied children who were born between October 1, 1996, and December 31, 1997, resided in the greater metropolitan areas of Seattle and the Northern California region, were continuously enrolled in the HMO through the first year of life, and received at least 1 polio vaccination (n = 2721 GHC and 25 609 KPNC enrollees). Statistical analyses of 2-year-olds were restricted to children who were born through June 30, 1997, and who were continuously enrolled for the first 2 years of life (n = 1313 GHC; 13 312 KPNC). Among children who were studied at 2 years of age, 22% of those at KPNC received IPV and 65% at GHC received IPV.
Using the same methods at both KPNC and GHC, we also studied immunization rates among children who were covered by Medicaid. We defined Medicaid status according to whether children were ever covered by Medicaid during the first year of life (for analyses of 1-year-olds) or the first 2 years of life (for the analyses of 2-year-olds). A total of 14% of GHC 1-year-olds and 11% of 2-year-olds had ever been covered by Medicaid; at KPNC, 0.6% of 1-year-olds and 0.5% of 2-year-olds had ever been covered by Medicaid.
We specifically excluded from our analyses children who did not receive at least 1 polio immunization. We did this because it was likely that these children were receiving their polio immunizations (and the majority of their other immunizations) at clinics or health departments outside the respective study HMO sites (eg, in the case of a child with dual parental coverage who might receive immunizations outside the HMO). However, we also considered that the proportion of children who had not received at least 1 polio immunization might have increased as a result of the new sequential schedule, and therefore we also assessed whether this proportion changed over time after IPV introduction. Of more than 28 000 children studied from the fourth quarter of 1995 through the fourth quarter of 1997, only 18 children (or <0.1% of the total enrolled population of children at GHC and KPNC) failed to receive at least 1 polio immunization during their first year of life. This proportion did not increase in the time period after introduction of the sequential schedule, and these children, therefore, were excluded from further analyses.
We assessed the impact of the new IPV schedule recommendations by evaluating 3 commonly used measures of immunization status at both 12 and 24 months of age: 1) up-to-date status; 2) cumulative up-to-date time; and 3) total number of missed-opportunity visits.16,17
Up-to-date status at 12 months was defined as having received all of the following: 2 polio vaccinations (IPV or OPV), 3 diphtheria and tetanus toxoids and pertussis/acellular pertussis (DTP/DTaP) vaccinations, 3 Haemophilus influenzae type b (HIB) vaccinations, and 1 hepatitis B vaccination administered after age 3 weeks. Because our data information systems may not have reliably captured hepatitis B vaccinations administered at the time of birth, we restricted our analysis of hepatitis B vaccinations to those administered after age 3 weeks and adjusted the definition of up-to-date accordingly. An additional polio, DTP/DTaP, HIB, and hepatitis B vaccination plus 1 measles-mumps-rubella (MMR) vaccination were required to be up to date at 24 months.
At GHC, some children received Comvax (a combination vaccine containing both hepatitis B and HIB vaccine and recommended at 2, 4, and 15 months) instead of monovalent HIB and hepatitis B. These children were considered up to date at 12 months if they received 2 polio, 3 DTP/DTaP, and 2 Comvax vaccinations and up-to-date at 24 months if they received 1 MMR, 3 polio, 4 DTP/DTaP, and 3 Comvax vaccinations. In all analyses, we disallowed vaccinations that were administered outside the minimum between-vaccination intervals recommended by the ACIP (number of disallowed vaccinations at GHC = 515 [1.2% of 41 544 administered vaccinations] and at KPNC = 6607 [1.75% of 376 868 administered vaccinations]).
Cumulative Up-to-Date Time
Cumulative up-to-date time was defined as the total number of days that an individual was up-to-date during a given time interval. Because this measured delay in vaccination coverage (rather than coverage at a single point in time), it likely was more sensitive than the traditional up-to-date status measures. We measured cumulative up-to-date time at birth to 12 and 24 months and calculated cumulative up-to-date time for each individual by assessing up-to-date status for each day over a given time interval and summing the days that were spent up to date.
Because additional vaccines are recommended as children grow older, their up-to-date status may change as well. For example, if a child who was up to date before 15 months of age did not receive his or her recommended MMR immunization by at least 17 months of age (2 months after the allowed 12- to 15-month window), he or she would be counted as up to date before 15 months of age in this analysis, but as not up to date at 15 months of age. In this example, if a child received all of his or her shots at the scheduled time except for MMR by 17 months of age, then the cumulative number of days spent up to date in the first 2 years would be 670 (730 potential days up to date minus 60 days late for MMR).
Finally, because of the concern that IPV would lead to reluctance on the part of physicians to administer multiple injections at a single visit or on the part of parents to have their children receive numerous injections simultaneously, we calculated the total number of missed-opportunity visits for each child by 12 and 24 months of age. Missed-opportunity visits were defined as visits at which at least 1 but not all recommended vaccinations were administered. For example, if a child had a health care visit at which only DTaP, hepatitis B, and IPV were administered—yet this child was also eligible for HIB vaccination—we considered this to be a missed-opportunity visit.
As the follow-up time was the same for each child in the study, we estimated the relative risk of being up to date in the IPV versus OPV groups using Mantel-Haenszel estimation on stratified cumulative incidence data.18 In these analyses, we controlled for the effect of a child's clinic on up-to-date status, reasoning that the type of clinic might be correlated with both up-to-date status and propensity to receive IPV. We did this by dividing clinics into tertiles based on mean up-to-date status in the year before introduction of IPV (October 1995–September 1996) and then assigning children to the clinic that they visited most often in the first year of life. At GHC and KPNC, the mean up-to-date status for clinics in the year before introduction of IPV was 88% and 95% at 12 months of age and 73% and 81% at 24 months of age, respectively. For the final data analysis, we stratified by clinic tertile, calendar time (in quarters), gender, Medicaid status, and, at GHC, whether the child was on the Comvax schedule.
We modeled cumulative up-to-date time by comparing its mean value in the IPV group with the OPV group using analysis of variance. Finally, to examine the effect of IPV on missed-opportunity visits, we modeled the likelihood of having a missed-opportunity visit using logistic regression. In the analyses of cumulative up-to-date time and missed-opportunity visits, we adjusted all results for the same potential confounding variables as in the analysis of up-to-date status.
The percentages of children who were and were not covered by Medicaid in each HMO and who were up to date at 12 and 24 months of age are shown in Fig 2. There were no obvious trends to suggest that implementation of an IPV schedule negatively affected coverage rates at either age studied. At both GHC and KPNC, children who were on an IPV schedule were as likely as children who were on the OPV schedule to be up to date at 12 months (GHC: relative risk [RR], 1.04; 95% confidence interval [CI]: 1.00, 1.08; KPNC: RR, 0.99; 95% CI: 0.98, 1.00); all RRs shown are for children who were on the IPV schedule compared with those who were on the OPV schedule). At 24 months, children at GHC who were on an IPV schedule were slightly more likely to be up to date than were children who were on the OPV schedule (RR 1.12; 95% CI: 1.05, 1.19). However, at KPNC, there was no difference between the 2 groups at 2 years of age (RR 1.00; 95% CI: 0.97, 1.02).
Among children who were covered by Medicaid, the results were similar. No differences in up-to-date status were seen at either GHC or KPNC when we looked at Medicaid children's status at 12 months of age (GHC: RR 1.09; 95% CI: 0.94, 1.27; KPNC: RR 1.05; 95% CI: 0.92, 1.19) or at 24 months of age (GHC: RR 0.79; 95% CI: 0.59, 1.06; KPNC: RR 0.91; 95% CI: 0.42, 1.98).
Cumulative Up-to-Date Time
Figure 3 demonstrates up-to-date status as a function of age for children at GHC and KPNC who were on IPV and OPV schedules. At GHC, children who received IPV had more cumulative up-to-date time at 12 and 24 months, although this difference was not statistically significant (at 12 months, an adjusted mean difference of 6.99 days spent up to date for children who were on IPV compared with OPV, 95% CI: –1.63, 15.6; at 24 months, an adjusted mean difference of 8.40 days, 95% CI: –12.8, 29.6). At KPNC, children who were on IPV spent on average 3 fewer days up to date in the first year of life, and this difference was statistically significant (at 12 months, an adjusted mean difference of −3.43 days spent up to date for children who were on IPV compared with OPV, 95% CI: –5.43, −1.43). At 24 months of age, children who were on IPV did not differ significantly from those who were on OPV in terms of total time spent up to date (adjusted mean difference of 0.97 days, 95% CI: –5.87, 7.81).
Children who were not covered by Medicaid at GHC and who were on an IPV schedule were less likely than those who were on the OPV schedule to have had a missed opportunity by 12 months of age (odds ratio [OR] 0.46; 95% CI: 0.31, 0.70) but equally likely to have had 1 by 24 months of age (OR 0.83; 95% CI: 0.60, 1.14; Table 1). At KPNC, children who were on an IPV schedule were more likely to have had at least 1 missed opportunity by 12 months of age (OR 2.06; 95% CI: 1.84, 2.30) and by 24 months of age (OR 1.50; 95% CI: 1.36, 1.67). Among children who were covered by Medicaid, there were no statistically significant differences in missed opportunities at GHC or at KPNC at either 12 months or 24 months of age.
In 1997, the United States undertook the most profound change in polio vaccination policy in more than 35 years, dating from the introduction of OPV as the preferred method in 1961 to 1962.1–3 An all-IPV schedule now has been recommended for routine childhood vaccination in the United States.19 In response to the original change, some researchers as well as some public health groups raised concerns that the recommended schedule change would have an impact on vaccine coverage rates and negatively affect a vaccine policy that had led to polio eradication in the northern and western hemispheres.1,2,5,6,9,11–14 Some of these concerns centered around the potential for parents or guardians to refuse the extra injection imposed by adding IPV to an already crowded vaccination schedule. For some visits, the total number of injections ranged as high as 4, including DTP/DTaP, HIB, hepatitis B, and polio vaccination. Although recent studies have found that parents are receptive to multiple injections at a single visit, there were concerns that providers might not be willing to administer all of the necessary injections at a single visit and that the new schedule would require several extra office visits to complete.6–8 Still others pointed out that the higher cost of IPV might represent a barrier to immunizing children fully and lead to incomplete adoption of the new schedule. Together, these problems could have led to an increase in the number of susceptible children in the population and the potential for the reappearance or increase in the number of vaccine-preventable diseases such as pertussis or measles.11,12 These concerns were particularly high for the Medicaid population, in which immunization coverage historically has lagged behind the coverage of the general population.6,10,14
Our study presents strong evidence that in 2 large HMOs, the changeover from an all-OPV schedule to one containing IPV was not associated with a negative impact on immunization coverage. Using traditional methods of assessing vaccination status (eg, up-to-date status), we found little difference in coverage among children who were on an IPV schedule compared with those who were on OPV at 12 months or at 24 months of age. This was true both for the Medicaid and for the non-Medicaid population. At 1 of the HMOs, children who were on IPV were slightly more likely to be up to date than were those who were on OPV and had an 11% greater likelihood of being up to date at 24 months of age. This difference was statistically significant even after adjusting for clinic characteristics that might have influenced both the choice of polio vaccine and the propensity to be vaccinated on time, as well as other potential confounding influences such as time, gender, and socioeconomic status.
At 1 site, the use of IPV was associated with an increase in the number of missed opportunities. In some respects, these missed opportunities might be considered intentional. That is, after the introduction of the sequential schedule, some clinics elected to adjust their own schedule and administer the third hepatitis B shot at 9 months of age or later (rather than, as previously, at 6 months of age). Therefore, in our study, these children were assessed as being more likely to have experienced a missed opportunity. We believe that these types of schedule changes must be acknowledged as potential sequelae of adding injections to an immunization schedule that many providers consider to be already overburdened. It was encouraging, nevertheless, that there were no differences in up-to-date levels at 12 or 24 months of age, suggesting that the necessary visits required to complete the immunization schedule were not inordinately delayed. In addition, at this site, children who were on IPV spent a small (but statistically significant) amount of time undervaccinated (∼3 days in the first year of life) relative to children who were on OPV. The opposite effect was seen at the other site; although likely attributable to a smaller sample size, the difference seen was not statistically significant.
This study had a number of strengths. We were able to gather information on a large birth cohort (totaling more than 28 000 children) at 2 HMOs. By restricting our study to children who were enrolled continuously from birth, we were able to reconstruct entire vaccination histories in an unbiased manner from automated systems (rather than by survey or by medical record review). These automated systems have been shown previously to be highly accurate and complete. However, our study did not completely capture vaccinations that might have been administered outside the HMO provider systems, although past studies show these to be relatively rare among our populations.20
A limitation of this study is that it does not provide information on the impact of the schedule change in non-HMO settings or in geographic regions other than the west coast. However, our results are consistent with similar data presented from other areas, suggesting that the impact of an IPV schedule is similar in different regions of the United States.21,22 In addition, the most recent national data show that immunization rates are at their highest level ever, suggesting little (if any) negative effect from the new IPV schedule.23 A second limitation is that we did not have vaccine information on children whose families disenrolled within the first 2 years of life. This occurrence was more pronounced among Medicaid participants, and data on the effect of the new IPV schedule on this population will need to be gathered from other sources, including survey and medical record reviews. Finally, we did not assess the interaction of minority status and receipt of IPV on immunization rates, as race and ethnicity information is not routinely gathered by GHC or KPNC. Additional studies of the impact of the IPV schedule within minority populations that are at high risk for missed immunization opportunities are needed.
Our study indicates that in 2 large HMOs, the changeover from an all-OPV schedule to a schedule containing IPV occurred relatively quickly and that this changeover had little if any negative impact on vaccine coverage. We did not find any evidence that children who were covered by Medicaid were especially affected by the new vaccination schedule. Use of IPV was associated with a small increase in the likelihood of being up to date at 2 years of age at one of the HMOs and conversely was associated with a small increase in the likelihood of having a missed-opportunity visit in the other HMO.
This work was supported by Grant R95–074 from the Centers for Disease Control and Prevention's Comprehensive Linked Data Collection of Medical Events and Immunization.
- Received June 6, 2000.
- Accepted August 30, 2000.
Reprint requests to (R.L.D.) Immunization Studies Program, Center for Health Studies, Group Health Cooperative of Puget Sound, 1730 Minor Ave, Suite 1600, Seattle, WA 98101-1448. E-mail:
- Poliomyelitis prevention in the United States: introduction of a sequential vaccination schedule of inactivated poliovirus vaccine followed by oral poliovirus vaccine. Recommendations of the Advisory Committee on Immunization Practices (ACIP). MMWR Morb Mortal Wkly Rep. 1997;46:(RR-3)1–25 (published erratum appears in MMWR Morb Mortal Wkly Rep. 1997;28:183)
- Katz SL
- American Academy of Pediatrics, Committee on Infectious Diseases
- Judelsohn R
- Schneider S
- Adler P
- Judelsohn R
- Paradiso P. The future of polio immunization in the United States: are we ready for change? Pediatr Infect Dis J. 1996;645–649
- Impact of the sequential IPV/OPV schedule on vaccination coverage levels—United States, 1997. MMWR Morb Mortal Wkly Rep. 1998;47:1017–1019
- Szilagyi PG, Rodewald LE, Humiston SG, et al. Missed opportunities for childhood vaccinations in office practices and the effect on vaccination status. Pediatrics. 1993;91:1–7 (published erratum appears in Pediatrics. 1993;91:545)
- Breslow NE, Day NE. Statistical Methods in Cancer Research, II: The Design and Analysis of Cohort Studies. Lyon, France: IARC Scientific; 1987. Publ. No. 82
- Copyright © 2001 American Academy of Pediatrics