|Home | About | Journals | Submit | Contact Us | Français|
Background: This retrospective study determined the level of compliance to rotavirus vaccination guidelines within a large, commercially insured US population, as well as compliance with PI, ACIP and HEDIS measures for rotavirus vaccination.
Methods: Medical and pharmacy claims were obtained from the HealthCore Integrated Research Database. Enrolled children were stratified into PI, ACIP and HEDIS cohorts. The PI cohort was subdivided into RV5 and RV1 cohorts due to the differences in dosing schedules and patients with mixed dosing were excluded from the these two cohorts. Patients identified in the HEDIS cohort were linked to the administering physicians.
Results: Of 162,614 patients in PI cohort, 27% did not receive rotavirus vaccinations, 24% (RV5) and 15% (RV1) had incomplete doses (p < 0.0001; RV1 vs. RV5). A total of 76% of patients completed RV5 series but not on schedule, 54% completed on schedule. A total of 85% of patients completed the RV1 series at any time, 69% completed on schedule. Among health plans, 53% of patients completed the series, 22% (RV5) and 15% (RV1) had incomplete doses (p < 0.0001). Of 2,086 physicians who treated ≥ 10 patients within the plan (regardless of vaccination status), 78% had > 50% of patients complete, 22% had > 90% of patients completed.
Conclusion: Despite both two effective rotavirus vaccines and national immunization recommendations, rotavirus vaccination remains underutilized for infants.
Rotavirus is a leading cause of vomiting and diarrhea, which can range from mild to life-threatening, in children younger than 5 y of age.1 Deaths from rotavirus are more common in developing countries and less prevalent in the US. The pathology of rotavirus infection typically involves fever, vomiting and diarrhea that may persist for 3 to 8 d.2 Within the temperate climate of the United States, rotavirus outbreaks follow a seasonal pattern from December to June.2 Primary infections of rotavirus do not provide sufficient immunity to protect against re-infections; however, the immune response after primary cases of rotavirus generally leads to a milder illness.3
Two rotavirus vaccines are currently approved by the Food and Drug Administration for use in the US: the pentavalent vaccine (RV5) (RotaTeq®, Merck and Co., Inc.), which is a 3-dose series introduced in February 2006, and the monovalent vaccine (RV1) (Rotarix®, GlaxoSmithKline), which is a 2-dose series, introduced in April 2008. Dosing schedules vary by vaccine and the following standards due to the dosing regimens: (1) prescribing information (PI); (2) Advisory Committee on Immunization Practices (ACIP), a government body which provides recommendations to the Center for Disease Control and Prevention (CDC); and (3) Healthcare Effectiveness Data and Information Set (HEDIS) measures from the National Committee for Quality Assurance (NCQA), which is a quality measure for vaccinations used by greater than 90% of health plans. (See Table 1 for dosing schedules.) A recent study by Krishnarajah et al. observed that 83.3% of RV1 patients and 76.4% of RV5 patients were compliant with ACIP recommendations for the vaccine series, but no evaluation of compliance with HEDIS standards was made.6
The burden of illness associated with rotavirus infection can be significant. Prior to introduction of a vaccine in 2006, rotavirus infection accounted for 55,000 hospitalizations annually among children younger than 5 y of age in the US.2 Emergency department utilization as a result of rotavirus infection has been estimated at 260,910 visits annually for children up to age 5 y in the US, based on data from 1999 to 2000,7 with 410,000 annual physician visits.8 Total estimated annual costs, including both direct and indirect expenses, were approximately $893 million (1993 to 2002 data).9 United States deaths related to rotavirus infection in children younger than 5 y ranged between 20 and 60 per year, but worldwide approximated half a million deaths per year.10,11
Within the first year of the introduction of Rotateq, up to 60% of children aged 5 mo or younger received at least one dose of rotavirus vaccine, according to data collected by the CDC at its 4 immunization surveillance sites.12 National Immunization Survey data collected in 2010 showed an increasing trend in rotavirus vaccine coverage. The percentage of children receiving the complete rotavirus vaccine series (either two or three immunizations, depending on the specific vaccine used) rose from 44% in 2009 to 59% in 2010, with coverage rates reaching 70% among children born in the first half of 2009, the most current data available.13
The rapid uptake of the vaccine helped reduce the incidence of rotavirus-related hospitalizations from a median of 101.6 per 10,000 children from 2000 through 2006 to 85.5 in the 2007–2008 season and 55.5 in the 2008–2009 season.14 The onset of rotavirus season was delayed 11 weeks in 2007–2008 and 6 weeks in 2008–2009, and the duration of the season was shortened from a median of 26 weeks from 2000 through 2006 to 14 weeks in 2007–2008 and to 17 weeks in 2008–2009.13 Widdowson et al. have estimated that a national vaccination program for rotavirus would decrease rotavirus related deaths by 44%, hospitalizations by 66%, office visits by 60% and including medical and non-medical costs, would save a net of $216 million dollars annually in the US9
The purpose of this observational, retrospective, longitudinal, cohort study is to determine the level of compliance to the PI, ACIP’s recommendations and HEDIS measures in administration of rotavirus vaccine uptake within a large, commercially insured population in the US, as well as the compliance rates of physicians and health plans in meeting the PI, ACIP and HEDIS measures for rotavirus vaccination. Secondarily, when reviewing the ACIP and PI results, comparisons were made between the RV1 and RV5 cohorts. No comparisons between ACIP, PI or HEDIS cohorts were made.
Among 162,614 patients who met the PI criteria, 56% completed the vaccination series at any time and 77% completed on schedule. Of all patients in the PI cohort, 27% did not receive rotavirus vaccinations and 24% and 15% of patients had incomplete doses in the RV5 and RV1 cohorts, respectively (p < 0.0001; RV1 vs. RV5). Patients who received RV5 vaccination (109,163) greatly outnumbered those who received RV1 vaccination (9,861). In the RV5 cohort, 76% of patients completed the RV5 series at any time, with 54% completing on schedule. In the RV1 cohort, 85% completed at any time and 69% completed on schedule (p < 0.0001 for comparisons between RV1 and RV5 in both categories). Regarding incomplete vaccinations, 17% of children in the RV5 cohort received two of the three scheduled doses and 7% received only one of the three doses. In the RV1 cohort, 15% of children missed one of the two doses. Male and female patients had similar rates of completion within each of the RV5 and RV1 cohorts. Children who had incomplete vaccinations were more likely to have initiated the series later in life in both the RV5 and RV1 cohorts (12 and 15 weeks, respectively), than those who completed the rotavirus vaccination series (9 and 10 weeks, respectively, p < 0.0001). Additionally, vaccination completion rates ranged according to geographic region (Table 2).
A total of 164,596 patients met the ACIP criteria, and 58% of the children completed the vaccination series at any time Of all patients in the ACIP cohort, that met the ACIP criteria, 26% did not receive rotavirus vaccinations. A far greater number of patients received RV5 vaccination (107,022) than RV1 (9,275). A total of 22% and 15% of patients had incomplete doses in the RV5 and RV1 cohorts, respectively (p < 0.0001). Overall, among those in the ACIP cohort who initiated an RV vaccine 64% were vaccinated on schedule and 78% completed the series at any time. A total of 81% and 66% completed on schedule in the RV1 and RV5 cohorts, respectively (p < 0.0001), with 85% and 78% completing at any time in the RV1 and RV5 cohorts, respectively. Patients within the mixed cohort (1 or more doses of RV5 with 1 dose of RV1) completed at a rate of 72% (p < 0.0001). Male and female patients had similar rates of completion within each of the RV5, RV1 and mixed cohorts. Children who had incomplete vaccinations were more likely to initiate the series later in life in both the RV5 and RV1 cohorts (13 and 15 weeks, respectively), than those who completed the rotavirus vaccination series (9 and 10 weeks, respectively, p < 0.0001). By geographical region, the proportion of patients not receiving vaccinations was 41% in the Northeast region, 31% in the West, 25% in the Midwest and 20% in the South, and the total proportion was 4% in the northeast region, 9% in the West, 7% in the Midwest and 6% in the South among all patients. A total of 78% and 85% of the RV5 and RV1 patients, respectively, completed the vaccination series. Among those patients within the RV5 cohort, 16% of patients received two of the three doses and 7% received only one dose. Among patients in the RV1 cohort, 15% received one of the two doses (Table 3).
Among the health plans, 53% of patients completed the rotavirus vaccination series (Table 4). The percentage of completed vaccination series among patients who initiated was varied within regions, ranging from 70% for the Northeast to 82% in the Midwest and 79% for all plans. The proportion of incomplete vaccinations ranged from 13% in the Midwest to 16% in the West and overall, 15% for all plans. Overall, 33% of patients received no rotavirus vaccination, ranging from 30% for the Midwest region to 52% of those in Northeast that did not receive any rotavirus vaccine doses.
A total of 22,147 physicians, treating an average of 7 patients each, were included in the assessment of compliance with HEDIS rotavirus vaccine measures. Of those, the majority (16,997, 77%) were pediatricians. Of the 2,086 physicians who treated at least 10 health plan patients (regardless of vaccination status), 78% had more than half their patients complete the rotavirus vaccination series, with 22% having more than 90% of patients with a completed series (Table 5). A greater percentage of pediatricians (22%) had completed vaccine series in > 90% of patients compared with primary care physicians (15%; p < 0.0001). A total of 7% (138 physicians) had none of their eligible patients vaccinated with the rotavirus vaccine. Far fewer pediatricians had no vaccination among members (6%) compared with PCPs (23%; p < 0.0001).
The purpose of the current study was to determine rotavirus vaccination rates among children covered under a large commercial health plan, and to establish whether these children were being vaccinated according to the PI, ACIP and HEDIS quality measures and secondarily, to compare compliance with the ACIP schedule among the two rotavirus vaccines licensed in the US.
In our study population, within the guidelines of the prescribing information, 55% of patients completed the series on schedule. According to the ACIP recommendation, which recommends completion of the rotavirus immunization series by 8 mo, 64% of children were vaccinated on schedule. When assessing completion rates among children who initiated the vaccination series, completion rates were 77% of the PI cohort and 78% of the ACIP cohort completing the series at any time. Using the total cohort as the denominator, rather than those who were initiated, 56% of children in the PI cohort and 58% of children in the ACIP cohort completed the vaccination series at any time. This rate of completion is comparable to the one reported by the CDC. The CDC reported that 59% of children aged 19 to 35 mo were fully vaccinated against rotavirus in 2010 and among children born in the first 6 mo of 2009, rotavirus immunization coverage reached a high of 70%.13 While this is a benchmark, it is important to note that our study measures compliance (on schedule) and CDC measures overall completion, regardless of timing. Compliance in the RV1 cohort (two dose) was higher than in the RV5 (three dose) cohort, both for completion at any time, and for completion on schedule. While it may seem obvious that compliance with a two vaccine series is easier for patients and providers to accomplish than with a three vaccine series, the issue of vaccine schedule compliance is complex. Factors which have been associated with poor vaccine compliance are varied, and include parental beliefs about the importance of vaccine schedules, provider beliefs about false contraindications, older age of the child, younger maternal age, large family size, late birth order, lack of knowledge about disease and vaccination, negative beliefs/attitudes about immunization, fears of adverse events, not remembering vaccination schedules/appointments, delayed well child visits and seasonal variability.15-17
Our study found that 26% of the ACIP cohort and 27% of PI cohort did not receive any rotavirus vaccinations. This finding is comparable to that reported by CDC where 30% of children remain unvaccinated.13 In contrast to our study, national vaccination rates include children who have health plan coverage as well as those who are uninsured.
According to the data derived from the HIRD, compliance with the HEDIS quality measure, which requires that the rotavirus vaccination series be completed by the child’s second birthday, was 52%, which is 7% lower than the overall national rotavirus vaccination rate of 59% reported by the CDC.13 Further, our study results demonstrate that compliance with HEDIS quality measures was inconsistent, varying by regional plans and among physicians. The proportion of children who remained unvaccinated by their second birthday ranged from less than 30% to more than 51%, reflecting a wide disparity in implementation of vaccine guidelines. The National Immunization Survey (NIS) also reported wide variance in the vaccination coverage rates among states within children aged 18–36 mo. Although the 2010 national coverage rate for rotavirus vaccine was 59.2%, individual state rates ranged from a low of 42.1% in Maine to a high of 82.1% in Delaware.13
With regards to overall vaccination, when assessing physician data, there was a similar pattern of inconsistency regarding adherence to HEDIS quality measure. Whereas 78% of physicians fully vaccinated more than half of their patients by the time they reached 2 y of age, only 22% had 90% or more of their patients receiving complete rotavirus vaccinations by the time they reached 2 y of age. Previous research found wide variance in general immunization rates among providers, ranging from 51% to 97%, leading to the conclusion that individual provider behavior was the single most important predictor of a child’s vaccination status.18 In addition, vaccination rates differed greatly between physician specialty, with pediatricians being more likely to fully immunize their patients according to vaccine guidelines than primary care physicians. This difference in vaccination rates between pediatricians and primary care physicians is not surprising. A previous study found although 85% of pediatricians surveyed routinely offered the rotavirus vaccine, only 45% of family medicine physicians did so.19 According to the previous study, the primary reason for the discrepancy in vaccination rates was a greater concern among family medicine physicians about the safety of the rotavirus vaccine and reluctant to add another vaccine to the schedule.19 Although explanations for the difference in immunization rates between pediatricians and primary care physicians were beyond the scope of our study, it is possible that further study into this difference could provide further insight into rotavirus vaccination rates.
When compared with other childhood immunizations measured within the NIS (2006–2010), rotavirus vaccinations fell short of the completion rates attained by other vaccines. Greater than 90% vaccine series completion rate was attained for measles, mumps and rubella vaccine, poliovirus vaccine, hepatitis B vaccine and varicella zoster vaccine. Vaccination coverage rates were slightly lower for diphtheria, tetanus toxoid and acellular pertussis vaccine (84%) and pneumococcal conjugate vaccine (83%).13 According to the NCQA report of HEDIS measures for 2009, completed vaccination series in children 2 y of age with commercial insurance were achieved in greater than 85% for diphtheria, tetanus toxoid and acellular pertussis vaccine; polio vaccine; measles, mumps and rubella vaccine; Hemophilus influenzae type b vaccine; hepatitis B vaccine; varicella zoster vaccine; and pneumococcal conjugate vaccine.20 In this study, patients who initiated, but failed to complete the rotavirus vaccine series tended to have initiated the vaccine series several weeks later than those who did complete the series. This could be an indicator that could be utilized by physicians in estimating which patients would require more attentive follow-up to encourage series completion. It is possible that given the history of rotavirus vaccines and the apparent association of intestinal intussusception with earlier versions, that parents and physicians are reluctant to fully adopt the rotavirus vaccine series.21
Although this study included data from a large and diverse patient population, limitations should be noted. First, due to the fact this was a retrospective study, selection bias may have been introduced during the study cohort identification. Only patients with continuous eligibility for the specified cohort of interest (i.e., at least 8 mo of continuous eligibility for assessing PI requirements) were selected. While capturing the data in this manner allowed for a more meaningful comparison among patients, those who had intermittent medical coverage during a 2-y period would not be included. Second, records were identified in the database using Current Procedural Terminology (CPT) codes, which may have included coding errors or omissions, although research to test the positive predictive value of CPT codes for rotavirus vaccination have been shown to be better than 85%.22 This specific information derived from billing codes may be subject to errors based on submission by physicians or insurance companies. Third, while the database utilized included patients from 14 health plans, patients were not included from all 50 states within the United States, so the data may not be broadly applicable to the US population. Finally, all patients in the study cohort were members of a commercial health plan and the results may not be generalizable to those with public health insurance or the uninsured.
Within the PI group, all patients receiving doses of both RV5 and RV1 were excluded since none of this group would be considered to have completed vaccination according to the PI. We did, however, conduct sensitivity analysis by adding patients in the mixed cohort to the RV5 and RV1 cohorts based on which vaccination was first administered, similar to an intent-to-treat analysis in a clinical trial, and assuming that none of these patients completed the series according to the PIs. This resulted in 73% compliance in the RV5 cohort and 56% compliance in the RV1 cohort. The completion rates for the RV1 group under this analysis are reduced substantially because the number of patients in the mixed cohort that started on the RV1 vaccine and switched is substantial compared with the number of patients initially in that cohort (approximately 20% of the original cohort). We note that the fact that the RV1 vaccine was suspended from the market for approximately months during 2010 may have contributed to this result. For the RV5 cohort, the mixed patients starting on RV5 are only a small percentage of the total (approx. 3%) and so had little impact. The authors do not believe this type of analysis is the correct approach and suggest caution in interpreting this data.
Rotavirus vaccination remains underutilized with both US-infant licensed RV vaccines. In our study of commercially insured children as many as 26% of eligible children were unvaccinated, with rates varying by region and type of physician. Initiatives to improve vaccination initiation and completion rates may help to ensure children are provided adequate protection against rotavirus infection and its complications.
This observational, retrospective, longitudinal cohort study utilized administrative claims data, including medical and pharmacy claims, from the HealthCore Integrated Research Database (HIRDSM). The HIRD contains a broad, clinically rich and geographically diverse spectrum of longitudinal claims data from 14 health plans in the southeastern, mid-Atlantic, central and western regions of the United States. The database includes claims information from the largest commercially-insured population in the United States and includes various lines of business, such as health maintenance organizations (HMOs), point-of-service plans, preferred provider organizations (PPOs) and indemnity plans. This study included claims submitted from August 1, 2008 through January 31, 2011, with the exception of one plan for which the data endpoint was November 20, 2010 (less than 5%).
The data were accessed in a manner that complied with federal and state laws and regulations, including those related to privacy and security of individually identifiable health information. Pursuant to the Health Insurance Portability and Accountability Act of 1996 (HIPAA), HealthCore used these data for research as a limited data set. All patient information was de-identified prior to use for this research.
Enrolled children were stratified into three cohorts: (1) PI; (2) ACIP; and (3) HEDIS cohorts. Due to the difference in dosing schedule with RV1, which requires 2 doses and RV5, which requires 3 doses, the PI cohort consisted of two subcohorts, the RV5 cohort and the RV1 cohort. The RV5 cohort consisted of all children who had one or more claims for only RV5 (CPT code: 90680) and had 8 mo continuous health plan eligibility after their date of birth. The RV1 cohort consisted of all children who had one or more claims for only RV1 (CPT code: 90681) and had 6 mo continuous health plan eligibility after their date of birth. In each patient cohort, the index date was defined as the date of first vaccination or date of birth in patients with no medical claim for vaccine. Within the PI dosing regimen, the following guidelines needed to be satisfied for a patient to be considered completed on schedule: RV5 schedule: dose 1 must have been administered between weeks 6 and 12; dose 2 must have been administered 4–10 weeks after dose 1; and dose 3 must have been administered 4–10 weeks after dose 24; RV1 schedule: dose 1 must have been administered between weeks 6 and 20; and dose 2 must have been administered ≥ 4 weeks after dose 1.5 Subjects who had more than one brand of vaccine in their vaccination series were excluded from the RV1 and RV5 cohorts and excluded from the analysis. Patients with the required number of vaccinations but not within the correct dosing schedule were considered “complete not on schedule,” or “complete at any time.” Vaccinations administered prior to 6 weeks of age were considered invalid and did not count toward a completed dosing regimen.
The ACIP cohort consisted of all children who turned 8 mo of age within the study period and had continuous eligibility since birth. Patients were flagged if they had at least one medical claim for RV5 (CPT code: 90680) or RV1 (CPT code: 90681) and if they had 8 mo of continuous eligibility after their date of birth. Patients who received a mixed combination of both RV5 and RV1 vaccinations were excluded from the RV5 and RV1 cohorts and completion calculated separately by counting those patients who had a total of three doses were administered. If the vaccination series were completed by age 8 mo, the series was considered to be complete on schedule.
The HEDIS cohort consisted of all children, regardless of vaccination received, whose second birthday was within the study period. Patients were flagged if they had at least one medical claim for RV5 (CPT code: 90680) or RV1 (CPT code: 90681) and if they had 24 mo of continuous health plan eligibility after their date of birth.
To assess physician compliance with successfully completing the rotavirus vaccination series, patients identified in the HEDIS cohort were linked to the administering physicians. Physicians who had at least 10 health plan patients in the cohort were included in the analysis. Physicians who had treated less than 10 health plan patients were excluded to avoid introducing error based on the fact that physicians may have a limited number of health plan patients and would not be properly reflected in the analysis. Patients who had vaccine administered by more than one physician were linked to each physician who had administered vaccine. Additionally, a power calculation was performed to assess if the sample provided would be adequate to detect a significant difference. Given the sample size in this study,
To assess region compliance rates in meeting HEDIS measures for rotavirus vaccination, health plan level data were grouped into four regions: Northeast region includes CT, ME, NH and NY; Midwest region includes MO, IN, OH and WI; South region includes GA, DE, VA and KY; West region includes NV and CA. Within the region, rates of vaccination completion were assessed.
Descriptive statistics were used to evaluate the percentage of children, physicians and health plans meeting PI, ACIP and HEDIS requirements. Main analysis for PI and ACIP cohorts is based on children who received a single brand and sensitivity analysis included mixed vaccination (i.e., children received both RV1 and RV5). Percentages of physicians and plans that were adherent and non-adherent to HEDIS measures for rates of vaccination series completion were also assessed using descriptive statistics.
Infants were not randomly assigned to study cohorts, but were grouped according to the vaccine brand that was actually administered; this resulted in disproportionate cohort sizes (RV5: n = 109,163; RV1: n = 9,861 in PI cohort). Although the cohort sizes were disproportionate, this did not affect the conclusions as the study was sufficiently powered to detect differences in the percent of infants completing the series by cohort. There was more than 99% power to detect the 2% or more difference between the completion rate in RV5 and RV1, assumed α = 0.05 and 76% completion rate in RV5 cohort.
G.K. is an employee and stockholder of GSK. D.E. and T.G. are employees of HealthCore, an independent consulting group, hired to collaborate on this research study. We would like to acknowledge Cheryl Jones and Betsy Marks for their contribution of review, as HealthCore employed medical writers.
Previously published online: www.landesbioscience.com/journals/vaccines/article/22877