Abstract
Objective Vaccination against preventable infections is important for the management of rheumatic diseases (RDs). This study assessed the vaccination coverage and predictors among patients with RDs using real-world data from Israel.
Methods This retrospective cross-sectional study, based on a Maccabi Healthcare Services database, included adult patients diagnosed with rheumatoid arthritis (RA), psoriatic arthritis (PsA), and systemic lupus erythematosus (SLE), as of April 30, 2019. Age-specific vaccination coverage for influenza (past year), pneumococcal (23-valent pneumococcal polysaccharide vaccine [PPSV23] and/or 13-valent pneumococcal conjugate vaccine [PCV13]), and live-attenuated herpes zoster (HZ) vaccines (past 5 years) was reported. Logistic regression was used to investigate predictors of vaccination.
Results The study included 14,528 patients (RA: n = 6932; PsA: n = 4395; SLE: n = 1951; > 1 condition: n = 1250). Influenza vaccine coverage among patients with RA, PsA, and SLE was 45.1%, 36.2%, and 33.7%, respectively. For PPSV23, corresponding rates were 19.6%, 16.2%, and 12.6%, respectively. In the elderly population (≥ 65 years), 63.2% had influenza vaccine in the past year and 83.4% had a PPSV23 vaccine in the past 5 years or at age ≥ 65. For PCV13 and HZ, coverage in the overall study population was low at 4.8% and 3.6%, respectively. Central residence and treatment with corticosteroids and biologic or targeted synthetic disease-modifying antirheumatic drugs within the past 5 years were significant predictors of vaccination coverage across all vaccines (P < 0.05). Other predictors varied by vaccine, including female sex (influenza, PPSV23, PCV13), age (influenza, PPSV23), chronic comorbidities (influenza, PPSV23, PCV13), shorter disease duration (PCV13), and high socioeconomic status (PCV13, HZ).
Conclusion This study demonstrated suboptimal coverage of influenza, pneumococcal, and HZ vaccination in patients with RA, PsA, and SLE, in particular among younger adults in Israel.
Patients with autoimmune inflammatory rheumatic diseases (AIIRDs) bear an increased burden of infections and infectious complications in comparison to the general population.1-3 This risk is attributed to the nature of autoimmune diseases, associated comorbidities, and the use of immunosuppressive (IS) treatments.4-6 Consequently, the prevention of infections constitutes an important part of the treatment of patients with AIIRD.
Vaccine-preventable infections caused by influenza virus, Streptococcus pneumoniae or pneumococcus, and the reactivation of the varicella zoster virus leading to herpes zoster (HZ), are particularly increased across various AIIRDs.2 In patients with rheumatoid arthritis (RA), increased prevalence of influenza and a 2.75-fold increase in the incidence of influenza-related complications, including pneumonia, stroke, and myocardial infarction, were reported in a large study from the United States.7 Increased prevalence of invasive pneumococcal disease was found in patients with different AIIRDs.8 RA and systemic lupus erythematosus (SLE) were identified as risk factors for pneumococcal pneumonia in comparison to the general population.9 Importantly, young patients with SLE aged 18 to 49 years were particularly susceptible to invasive pneumococcal disease.10-12 The susceptibility to HZ infection among patients with AIIRD was confirmed by several studies.13,14 Specifically, patients with RA are at a 2-fold risk of developing HZ compared with the age-matched healthy population, with an even higher risk among elderly patients and those treated with Janus kinase inhibitors (JAKi).15 In SLE, the risk of HZ was further increased up to 10-fold higher compared with the general population.16 Young patients with SLE aged 18 to 30 years were particularly susceptible to HZ infection,17 which was linked to high disease activity and consequent use of high-dose corticosteroids (CS) and IS treatments in these patients.18,19
Vaccination is universally considered an important tool to mitigate the risk of preventable infectious diseases in susceptible patients with AIIRD. Reassuringly, most patients with AIIRD develop an adequate immunogenic response to influenza, pneumococcal, and HZ vaccination, except for those treated with a high dose of CS and CD20-depleting therapy such as rituximab.20 With regard to the efficacy of vaccination, inactivated influenza vaccine reduced influenza-like illness, hospitalization for pneumonia and/or chronic obstructive pulmonary disease exacerbation, and death due to pneumonia in patients with AIIRD treated with disease-modifying antirheumatic drugs (DMARDs).21 Limited data are available concerning the efficacy of pneumococcal vaccine in patients with AIIRD.
In a randomized double-blind trial, the 23-valent polysaccharide vaccine (PPSV23) did not prevent against all-cause pneumonia in patients with RA,22 whereas a retrospective study showed that a single administration of PPSV23 provided up to a 10-year protection against the development of pneumococcal pneumonia in patients with RA on methotrexate.23 Live-attenuated HZ vaccine was shown to be protective for about 5 years in immunosuppressed patients.24
In line with present data, the European Alliance of Associations for Rheumatology (EULAR) recommends annual influenza vaccination and pneumococcal vaccination for most patients, and HZ vaccination for high-risk patients with AIIRD.25 According to the 2019 Israeli national health guidelines, patients with AIIRD were eligible for annual influenza vaccination; pneumococcal vaccination, in line with the Centers for Disease Control and Prevention (CDC) recommendations for immunosuppressed patients,26 including the 13-valent pneumococcal conjugate vaccine (PCV13) followed by the PPSV23 within a 2-month interval (called the “PCV13 prime–PPSV23 boost strategy”); and 1 dose of live-attenuated HZ vaccination for immunosuppressed patients aged ≥ 50 years.27 PPSV23 has been included in healthcare services since 1995 and was defined as a national quality control measure for the elderly population (aged ≥ 65 yrs) since 2008.28 Electronic medical record (EMR) alerts for influenza and PPSV23 have been routinely used within the medical system for the target population. PCV13 vaccine was approved in Israel in 2013 for use in adults with high-risk conditions, including immunosuppressed patients, but has been provided only since 2016.29 HZ vaccines have been in use in Israel since 2014. All vaccines were administered in the nationwide outpatient setting and documented in the health maintenance organization centralized EMRs. Influenza and PPSV23 were provided free of charge, whereas PCV13 and live-attenuated HZ vaccine required a medical prescription and a subsidized copay.
Despite these recommendations, a suboptimal uptake of vaccinations among patients with AIIRD has been reported worldwide,30-32 resulting from multiple reasons related to misconception and concerns regarding the safety of vaccines, along with low awareness and implementation barriers.33,34 In view of the lack of data concerning the vaccination coverage in patients with AIIRD in Israel, we investigated real-world vaccination coverage of influenza, pneumococcal (PPSV23 and PCV13), and live-attenuated HZ vaccines in patients with common rheumatic diseases (RDs), RA, psoriatic arthritis (PsA), and SLE within a large healthcare organization in Israel. The secondary aim of the study was to identify factors associated with vaccination uptake within this cohort of patients.
METHODS
Data source. The study used the computerized database of Maccabi Healthcare Services (MHS), 1 of 4 healthcare providers in the Israeli universal healthcare system, with approximately 2.7 million members, such that the database represents 25% of the national population.35 The database automatically integrates information available since 1998, including physician diagnoses, dispensed medications, consultations, hospitalizations, procedures, and individual sociodemographic data. MHS uses the International Classification of Diseases, 9th revision, Clinical Modification (ICD-9-CM) coding systems as well as self-developed coding systems to provide more granular diagnostic information beyond the ICD codes. Medications are coded according to the Israeli coding system with translations to the Anatomical Therapeutic Chemical coding system, and procedures are coded using Current Procedural Terminology codes.
Ethics. This study was performed in accordance with the principles of the Declaration of Helsinki and was approved by the Research Ethics Committee of MHS (MHS-0005-19). Patient consent was not required given the retrospective design of the study.
Study population. A retrospective cross-sectional study was conducted using MHS data collected between January 1, 1998, and April 30, 2019 (assessment date). The study population included MHS members who, on the assessment date, had been continuously enrolled for at least 12 months, were at least 18 years old, and were diagnosed with RA (ICD-9: 714.0), PsA (ICD-9: 696.0), or SLE (ICD-9: 710.0). For each RD of interest (RA/PsA/SLE), to increase the specificity of the case definition, patients were required to have (1) at least 1 diagnosis of the given condition defined by a rheumatologist, hospital, and/or the MHS Medication Approval Committee for the use of biologic agents/JAKi/phosphodiesterase 4 inhibitor indicated for one of these diseases; or (2) at least 2 diagnoses of the given condition defined by a primary care physician (PCP). In addition, patients with RA and SLE who met criteria 1 without criteria 2 and had no record of dispensed DMARDs were excluded. Patients with “mixed” diagnoses met the inclusion criteria for > 1 of these conditions.
Vaccine coverage. Data were obtained on dispensed prescriptions and procedures indicative of vaccination coverage up to 5 years prior to the assessment date (ie, April 30, 2014, to April 30, 2019) with the PPSV23, PCV13, and live-attenuated HZ vaccine. A recombinant zoster vaccine was not available in Israel within the study period. Vaccination against influenza was assessed up to 12 months prior to the assessment date, regardless of vaccine coverage in the previous years. For PPSV23, data were also obtained for patients vaccinated at age ≥ 65 years, or more than 5 years prior to the assessment date, in order to assess vaccine coverage in the prior 5 years or since age 65. For HZ vaccine, analysis was restricted to patients aged ≥ 50 years. All vaccines were administered in the nationwide outpatient setting and documented in the centralized MHS EMRs. All patients in the study population had the same health insurance and the same access to vaccinations as indicated by the local guidelines.
Disease category and medication use. Patients with a diagnosis history of only one of the conditions of interest were classified into mutually exclusive groups: RA, PsA, or SLE. Patients with > 1 different condition of interest were described separately. Disease duration (in years) was estimated as the time between the earliest diagnosis code in the database (RA, PsA, and/or SLE) and the assessment date. Data were obtained on medications dispensed since the earliest diagnosis and until the assessment date. DMARDs available to patients with RA, PsA, and/or SLE during the study period were classified into conventional synthetic DMARD (csDMARD; hydroxychloroquine, leflunomide, methotrexate, sulfasalazine), targeted synthetic DMARD (tsDMARD; apremilast, tofacitinib), and biologic DMARD (bDMARD; abatacept, adalimumab, anakinra, belimumab, certolizumab, etanercept, golimumab, infliximab, ixekizumab, rituximab, sarilumab, secukinumab, tocilizumab, ustekinumab). Other IS treatments included azathioprine, cyclosporine, cyclophosphamide, mycophenolate mofetil, and tacrolimus. Use of oral or intravenous CS was also described. For the main analyses, data obtained on treatments dispensed up to 5 years prior to the assessment date were described. Treatment history with any DMARD/IS was also described using data available since 1998.
Sociodemographic data and comorbidities. Data were collected on patients’ age, sex, socioeconomic status (SES), residence area (North, Central, and South regions), BMI (calculated as weight in kilograms divided by height in meters squared), and smoking. Smoking status was defined by 3 categories: no smoking (never), smoking in the past (ever), and current smoking. SES was based on a score ranked from 1 (lowest) to 10 on an individual’s place of residence (at the neighborhood level). This residential SES measure was originally derived by the Israel Central Bureau of Statistics using national census data and augmented by Points Location Intelligence, using aggregated data on housing prices, motorization level, education, employment, and financial resources.36 BMI was based on the most recent measure up to 5 years prior to the assessment date and categorized according to standard World Health Organization cutpoints. Existing MHS chronic disease registries (based on multiple data sources available since 1998) were used to describe the prevalence of diabetes,37 cardiovascular disease (CVD),38 chronic kidney disease (CKD),39 hypertension,40 osteoporosis,41 and chronic obstructive pulmonary disease. Cancer history was obtained from the Israel National Cancer Registry42 and MHS cancer registry data, which are drawn from pathology reports and diagnoses linked to prior medication approvals for oncology treatments.
Statistical analysis. Patient characteristics, treatment history, and vaccine coverage were described on the assessment date; descriptive statistics were presented as n (%) or median (IQR). Disease and age-specific vaccine coverage rates for each vaccine type and period of interest were calculated by dividing the number of vaccinated individuals by the total number of patients in the study population in a given disease and/or age subgroup. To investigate factors associated with vaccine coverage for each vaccine type, separate logistic regression models were used to obtain odds ratios (ORs) with 95% CIs for disease category and treatment history, sociodemographic characteristics, and comorbidities. Age group, sex, and disease categories were included in all models, whereas backward elimination (conditional) was used to select other variables for inclusion into each model (ie, starting with inclusion of all variables and using P ≥ 0.10 as removal criterion for excluding variables from the model in a stepwise manner).
RESULTS
Study population. A total of 16,490 adult patients met inclusion criteria on the assessment date, with 12 months’ continuous enrollment. Based on the specific exclusion criteria for RA and SLE, 1962 were excluded (RA, n = 1810; SLE, n = 152) due to lack of record of dispensed DMARDs. The study population included 14,528 patients, with a total of 96.1% and 91.6% of patients continuously enrolled for ≥ 5 and ≥ 10 years prior to the assessment date, respectively. Patients were subgrouped in the following disease categories: RA (n = 6932), PsA (n = 4395), SLE (n = 1951), and patients with > 1 of these conditions (n = 1250; Table 1). The latter group comprised patients primarily with diagnoses of RA with PsA (n = 838), or RA with SLE (n = 345), with the remaining patients having a history of PsA with SLE (n = 41), or all 3 conditions (n = 26). Median age was highest for patients with RA (64.3, IQR 52.6-73.5), compared to PsA (56.8, IQR 46.7-66.3) and SLE (49.7, IQR 41.0-61.3). Women accounted for > 70% of patients in all disease groups, except in PsA. Median disease duration ranged from 8.1 to 14.4 years across the disease categories. Patients with PsA were characterized by a relatively shorter disease duration compared to the other disease groups. The prevalence of age-related chronic comorbidities was highest among patients with RA. Patients with any history of dispensed DMARD/IS accounted for 80.3%, 59.8%, 73.6%, and 89.8% of all patients with RA, PsA, SLE, and > 1 condition, respectively. Patients with PsA had the lowest treatment use rates compared to other subgroups.
Characteristics of prevalent patients with RA, PsA, and/or SLE (MHS, 30/4/2019).
Vaccination coverage. Overall, influenza vaccine coverage rates (1 yr prior to the assessment date) among patients with RA, PsA, SLE, and > 1 condition were 45.1%, 36.2%, 33.7%, and 46%, respectively. For PPSV23 (past 5 yrs), corresponding rates were 19.6%, 16.2%, 12.6%, and 17.7%, respectively. For PCV13 (past 5 yrs), coverage in the overall study population was low (4.8%) independently from vaccination with PPSV23. Patients who had at least 1 pneumococcal vaccine of any type (PPSV23 and/or PCV13) as well as at least 1 influenza vaccine accounted for 37.3% overall and 82.2% among age ≥ 65 years. Corresponding results restricted to the previous 5 years were 15% and 28.6%, respectively. The percentage of patients who had at least 1 vaccine of any type before/on the assessment date was 73.5% overall and 93.2% among patients aged ≥ 65 years.
Age-specific vaccine coverage in the study population is presented in Table 2. Overall, in the elderly population (age ≥ 65 yrs, any disease group), 63.2% had an influenza vaccine in the past year and 83.4% had a PPSV23 vaccine in the past 5 years or since age 65. Age-specific vaccine coverage rates for PPSV23 in the prior 5 years or since age 65 ranged from 78.1% (patients with SLE aged 65-74 yrs) to 100% (patients with SLE or > 1 condition, aged ≥ 85 yrs). Among patients with RA in age groups 65-74, 75-84 and 85+ years, PPSV23 coverage (5 yrs or age 65+) was 81.8%, 86.1%, and 92.7%, respectively. HZ vaccine coverage was low in all age groups and was 3.6% in the overall study population aged ≥ 50 yrs. A total of 3856 (26.5%) patients had no record of vaccination in the database for any of the vaccines investigated. Among patients who had no record of vaccination with a specified vaccine ever in the database, the percentage who received at least 1 vaccine of a different type (ie, potential missed opportunity for vaccination of the specified vaccine in combination with the other vaccine received) was 8.8% among those unvaccinated against influenza (ie, ever vaccinated with a PPSV13/PPSV23/HZ), 55.4% among those unvaccinated for PPSV23, 71.1% among those unvaccinated for PCV13, and 72.8% among those unvaccinated for HZ.
Age-specific vaccine coverage among patients with RA, PsA, and/or SLE (MHS, 4/30/2019).
Factors associated with vaccination. Table 3 to Table 6 report the factors associated with each type of vaccine based on the backward elimination analysis for selection of the significant factors only (as described in the Methods section). In multivariable analyses, central residence and treatment use of DMARDs/IS vs nontreatment in the past 5 years were significant (P < 0.05) predictors of vaccination coverage across all vaccines, whereas other factors associated with vaccination varied by vaccine type (Tables 3-6). High SES was also significantly associated with vaccination in all models, except for PPSV23. Female sex was significantly associated with vaccination coverage for all vaccines, except for HZ. Older age was significantly associated with vaccination coverage for influenza and PPSV23 vaccines. In addition, influenza, PPSV23, and PCV13 vaccination was significantly associated with having chronic comorbidities. For PCV13, patients diagnosed more recently were more likely to have been vaccinated.
Factors associated with influenza vaccine coverage in the past year among prevalent patients with RA, PsA, and/or SLE (MHS, 4/30/2019).
Factors associated with PPSV23 vaccine coverage in the past 5 years among prevalent patients with RA, PsA, and/or SLE (MHS, 4/30/2019).
Factors associated with PCV13 vaccine coverage in the past 5 years among prevalent patients with RA, PsA, and/or SLE (MHS, 4/30/2019).
Factors associated with herpes zoster vaccine coverage in the past 5 years among prevalent patients aged ≥ 50 years with RA, PsA, and/or SLE (MHS, 30/4/2019).
tsDMARD/bDMARD treatment was a significant predictor of vaccination coverage for influenza (adjusted OR [aOR] 1.5, 95% CI 1.4-1.6), PPSV23 (aOR 2.5, 95% CI 2.2-2.9), PCV13 (aOR 5.5, 95% CI 4.6-6.6), and HZ (aOR 1.4, 95% CI 1.1-1.8) vaccines. As certain comorbidities or treatments did not substantially improve the model according to the threshold for this selection process, they were not included in the multivariable model and as such are not listed within the tables.
DISCUSSION
Vaccination against preventable diseases is of utmost importance for patients with AIIRD, who are susceptible to infectious diseases. According to the 2019 EULAR recommendations for vaccination and the Israeli national guidelines for immunosuppressed patients, annual influenza vaccination and prime-boosting pneumococcal vaccination are recommended for most patients with AIIRD, and HZ vaccination is recommended for high-risk patients aged 50 years and older.25,27 Despite these recommendations, our study found a suboptimal vaccination coverage across all the vaccines among a representative cohort of patients with AIIRD, including RA, PsA, and SLE, who have high accessibility to vaccination through the national health insurance. The suboptimal vaccination coverage was particularly prominent among young and middle-aged patients in all disease groups, whereas elderly patients (≥ 65 yrs) had comparable vaccination coverage to the Israeli general population for influenza (past year: 63.2% vs 61%) and PPSV23 (past 5 years or at least once since age 65: 83.4% vs 77.7%).43 Our findings are in line with an Israeli study conducted within the same timeline (2019) and in same healthcare system on a cohort of immunocompromised individuals (n = 32,637) reporting a suboptimal pneumococcal vaccination rate of 11.3% for PCV13, 39.4% for PPSV23, and 45% for influenza.44 Consistent with our study results, age was associated with higher influenza and PPSV23 vaccination rates. A plausible explanation for this observation relates to the fact that influenza and PPSV23 vaccination constitutes a part of the national program for quality care in the elderly community,28 supported by regular EMR alerts reminding physicians to consider vaccinations in this group of patients and provided without prescription and free of charge. In comparison, PCV13 and HZ vaccines required a referral issued by a physician and a subsidized copay by patients, pointing to potential vaccination barriers such as low referral rates for vaccination among physicians as well as cost barriers for certain patients, especially those with low SES.
In our study, older age and chronic comorbidities were identified as predictors for influenza and pneumococcal vaccination, reflecting the implementation of the national guidelines for vaccinations in the general population, based on age, presence of chronic diseases, and general immune-suppression criteria. Although these criteria define certain target groups, patients with AIIRD seem to fall out of the target population scope. Indeed, the Israeli Ministry of Health recommends annual free influenza vaccinations for the whole population, following the CDC guidelines. Yet, an annual national campaign is mainly coordinated for school-aged children, individuals aged ≥ 65 years, and patients with certain comorbidities, not specifying RDs or those taking antirheumatic medications, resulting in a low influenza vaccination uptake in young and middle-aged patients with AIIRD.
The national guidelines for pneumococcal vaccination during our study period recommended vaccination in elderly individuals (≥ 65 yrs) and those affected by particular conditions that widely overlap with patients with RD treated with DMARDs, yet autoimmune diseases were not mentioned explicitly. Further, no specific recommendations for a stepwise pneumococcal vaccination, namely a PCV13 prime–PPSV23 boost strategy, for patients with AIIRD were issued on the national level. The very low uptake (< 8%) of PCV13 in our study should be interpreted in view of the fact that PCV13 vaccine was approved for use in adults with high-risk conditions in Israel in 2013 but was not provided until 2016.29 Further, a copay required by patients for the PCV13 administration compared to a free PPSV23 vaccine potentially precluded PCV13 vaccination among individuals with low SES. Indeed, high SES was a significant predictor for PCV13 vaccination in our study. The lowest uptake of the investigated vaccines in our study was noted for live-attenuated HZ vaccine (3.6%), Factors associated with vaccination included the older age (65-74 yrs), high SES, and treatment with bDMARD/tsDMARD. A low vaccination coverage with live-attenuated HZ vaccine of 1.2% was demonstrated by a large administrative claims data US study of 44,115 patients with autoimmune diseases conducted between 2006 and 2009.45 In a single-center study from Germany, none of 331 patients with RA received an HZ vaccine.46 The extremely low rate of HZ vaccine should be most likely attributed to a low awareness for vaccination by physicians and possibly concerns of using a live-attenuated vaccine in immunosuppressed patients. Survey studies in patients with RA showed that physician recommendation was the strongest predictor of vaccine uptake, including HZ vaccine,47 and the lack of recommendation by physicians was the most common reason for nonvaccination.48 Despite the reassuring evidence concerning immunogenicity and safety of the live-attenuated HZ vaccine, including patients under biologic treatment with tumor necrosis factor inhibitors,49 there is a global underuse of live-attenuated HZ vaccine in patients with RDs.32 Currently, a recombinant adjuvanted zoster vaccine is strongly recommended in the 2022 American College of Rheumatology guidelines for vaccination for patients with AIIRD on IS medications.50
The findings of our study are in line with previous reports on suboptimal vaccination coverage and predictors for vaccination (older age and treatment with CS and/or bDMARDs/tsDMARDs) in patients with RDs in different countries, with most studies focusing on the RA population.30,51-53 A Canadian survey of patients with various RDs (n = 352) reported suboptimal vaccination coverage with influenza, pneumococcal, and a particularly low rate of HZ vaccine uptake, ranging from 5.6% in patients with RA to 25% to 28% in patients with spondyloarthropathies and systemic autoimmune diseases, respectively.47 Data from the international Comorbidities in RA (COMORA) cohort that included 3920 patients with RA from 17 countries showed a low rate vaccination coverage for both the influenza (25%) and PPSV23 vaccines (17%), with significant disparities between countries. The influenza vaccination rate ranged from < 1% in Morocco and Egypt to 66.2% in Japan, and the PPSV23 vaccination rate ranged from 0% in Morocco to 56.5% in France.30 Consistent with our results, the predictive factors for vaccination in the COMORA cohort included older age, presence of comorbidities, and use of biologic therapy.30 In a large claims data cohort from Germany, influenza and PPSV23 vaccination rates were higher among patients with RA (n = 111,482) vs matched controls (n = 557,410), with rates of 40.8% vs 32.2%, and 15% vs 10%, respectively.53 In this study, predictors for both vaccinations included older age and the region of residence, whereas comorbidities, rheumatologic care, and biologic treatment were associated only with influenza vaccination.53 More recent studies demonstrated a trend for a higher influenza and pneumococcal vaccination coverage in patients with RA in certain countries compared to our study.31,54 In an electronic health record–based study (2000-2013) from the United Kingdom, in patients with RA treated with nonbiologic therapy (n = 15,724), 80% received at least 1 influenza vaccination and 50% patients received a pneumococcal (PPSV23) vaccination during 5 years of follow-up, with a higher vaccination rate in patients aged over 65 years, with 91% for influenza and 61% for PPSV23.31 In this study, older age, presence of comorbidities, and high frequency of visits to general practitioners were associated with vaccination.31 Consistently, a large survey from the UK undertaken in 2014, based on 929 patients with RA, reported that over 85% of patients were vaccinated against influenza, yet only 44% were vaccinated with PPSV23.54 The vast majority of vaccination in that study was undertaken in primary care. In contrast to most studies, older age did not appear to be predictive of vaccination but the use of biologic therapies was.54 Data from the Canadian Early Arthritis Cohort (CATCH) between 2017 and 2021, based on 431 newly diagnosed patients with RA, showed that influenza vaccine coverage increased by only 8% (from 38% to 46%) after RA diagnosis and remained below national goals.55 Several studies from the UK,31 Germany,53 and Canada55 found a regional difference in the rate of vaccination, as was shown in our study. The highest vaccination rate was demonstrated in the Central region of Israel, associated with a higher SES compared to the North and South regions. In addition to regional socioeconomic differences, there is a disparity between the scope of healthcare services available in Central Israel compared to the periphery due to a shortage of medical infrastructures in general, and of healthcare professionals (HCPs) in particular.56
Several factors may explain the suboptimal vaccination coverage rates observed in our study, including lack of proactive healthcare policies, low access to PCV13 and HZ vaccines, low awareness of guidelines57 and perceived responsibility for vaccination among HCPs, vaccination hesitancy by patients, and implementation barriers.58 A recent international survey among HCPs (n = 371) constituting mainly of rheumatologists (n = 330) evaluated a perspective on the 2019 EULAR vaccination guideline. Although the participants agreed with most of the guideline’s principles and recommendations, there was a discrepancy between the level of agreement and the low implementation of the guideline in routine clinical practice.57 Importantly, several studies consistently showed that recommendation by the treating physician was the strongest independent predictor of vaccination and the main factor associated with vaccine uptake in patients with RDs.34,47,59,60 Based on understanding the barriers and facilitators to the implementation of the vaccination policy, a multifaceted intervention at a single academic practice in the US modestly improved pneumococcal vaccination rate and only slightly improved HZ vaccination rate in patients with RA.61 The intervention combined physician auditing and feedback, electronic reminders with linked order sets, and patient outreach. Another single-center rheumatology department used an implementation vaccination strategy based on patient recalls, nurse-administered vaccines, and physician reminders, resulting in a 14% increase in influenza vaccination rate in patients with RA, which was especially prominent in elderly patients and those treated with biologics.62 Both studies demonstrated fundamental barriers related to the implementation of the vaccination policy, even following targeted intervention programs, calling for action on a national level. In Israel, a system-wide intervention to increase pneumococcal vaccination within the MHS was recently developed and included clinician- and patient-targeted alerts as well as reminders implemented in immunosuppressed patients,63 who were identified using an automated registry.44 During the intervention, a substantial increase in pneumococcal vaccination was observed among the target population: PCV13 vaccination rates increased from 11.9% to 50.1% (P < 0.001), and the PPSV23 vaccination rate increased from 39.4% to 57.1% (P < 0.001).63 A similar approach based on the automated registry to identify the target AIIRD population can be used to promote the implementation of the vaccination policy in these patients. Further, prioritization of vaccination of patients with AIIRD at the national level would be the most effective step for improving the recommended vaccination coverage in patients with AIIRD in Israel, based on a successful example of the national Pfizer-BioNTech coronavirus disease 2019 (COVID-19) vaccination policy implementation among the Israeli population.64 This step should be in conjunction with educating rheumatologists and PCPs on the important role they play in the appropriate vaccination of patients with AIIRD.
The strength of our study is its basis on accurate vaccination rates, representing a methodological advantage compared to survey-based studies. Several limitations of the study should be considered. Although the average income level of MHS members is slightly above the national average, this population shares similar sociodemographic characteristics,35 and can be considered representative of the national population. The design of our study precluded the investigation of the reasons for nonvaccination. Several additional limitations were intrinsic to the study design. The study definition of RDs was based on documented ICD-9 codes prone to potential misclassification, as EMRs were not available for review and validation of diagnoses. The inaccuracy of ICD-9 coding could lead to ascertainment bias, such as under- or overascertainment of the study population with AIIRD conditions. We aimed to increase specificity with a strict definition of cases as described in the Methods. Nonetheless, certain patients in the study were diagnosed with multiple RDs. We included these patients as they most likely had one of the conditions of interest, although their exclusion may have increased the specificity at the cost of introducing selection bias. In the regression models, this group of patients was included as a separate category of disease group. Despite the inaccuracy of the diagnosis, we believe that the effect of this limitation on vaccination uptake–related variables was nonsignificant. Further, there were no data on RD activity. Although the data addressed CS exposure, it was not possible to account for the dose or intermittent therapy. Finally, no adjustment was made for unmeasured cofounders, such as the severity of reported comorbidities.
In summary, our study represents, to the best of our knowledge, the first epidemiologic report from Israel related to the vaccination coverage of influenza, pneumococcal, and live-attenuated HZ vaccines in the pre-COVID-19 period in a large representative cohort of patients with the prevalent AIIRDs of RA, PsA, and SLE. The study results provide real-world evidence of suboptimal vaccination coverage of inactivated influenza and PPSV23 vaccines, especially in the young and middle-aged patient populations, and of vaccination rates in the elderly patients that were comparable to those of the general population. The coverage with PCV13 and live-attenuated HZ vaccines were particularly low in all disease groups regardless of age, reflecting lack of an efficient vaccination policy and outreach to the AIIRD population. Common predictors for vaccination were central residence, potentially indicating a high SES and/or a higher access to healthcare services, and treatment with CS and bDMARDs or tsDMARDs. The findings of the study are in line with global real-world evidence and can be generalized to the whole Israeli population, given the small size of the country and a similar insurance policy among other healthcare service providers. Thus, our study highlights the vaccination gap, especially among young and middle-aged patients with RDs, underlining the significant barriers to implementing the vaccination policy recommended by rheumatology societies. There remains a significant scope to improve the uptake of vaccinations in patients with AIIRD.
ACKNOWLEDGMENT
We acknowledge and thank Maccabi Healthcare Services and AbbVie for overall support and forming real-world research collaboration.
Footnotes
The study was funded by AbbVie. The design and study conduct were performed in collaboration of AbbVie and Maccabi Healthcare Services, including participation in the interpretation of data, review, and approval of the publication.
VF has served as a consultant to AbbVie and received speaker fees from AbbVie. SAS, SNB, and MB are employees of AbbVie. YFS is a former employee of AbbVie. SNB, MB, and YFS own AbbVie stock. OE has served as a consultant to AbbVie and has received research funding and speaker fees from AbbVie. CW and GC declare no conflicts of interest relevant to this article.
- Accepted for publication January 23, 2024.
- Copyright © 2024 by the Journal of Rheumatology
