A total of 468 relevant citations were identified through a search strategy. Eleven additional studies were detected by a hands-on search. After identification of duplicate citations (n=156 articles removed), title-/ abstract screening (n=323 articles) and full-text screening (n=67 articles), 44 studies met the inclusion criteria. The PRISMA diagram illustrates the selection process of the studies and shows reasons for exclusion (Figure 1).
3.1. Study Characteristics
The study designs are divided into 34 cross-sectional studies [9-42] of which 13 studies were developed by means of a secondary analysis of already existing data [11, 12, 14, 15, 19-21, 28-30, 38, 39, 41] and two cross-sectional studies following a qualitative study design [26, 32], four randomized controlled trials [43-46], two systematic reviews [47, 48], two cohort studies [49, 50] and two theoretical reviews [51, 52]. Sample size ranged from 37 [32] to 13.106.163 [27] participants.
Methods used to evaluate vaccination uptake in the cross-sectional and cohort studies and in the systematic reviews were:
- questionnaires and surveys [13-15, 19, 21, 22, 25, 28, 34-36];
- telephone interviews [20, 23, 27, 29, 30, 33, 39, 41];
- data retrieved from national/regional authorities and health care institutions [12, 38, 40, 49, 50];
- data retrieved from medical records or vaccination registries [9, 10, 31, 37];
- combined methods [11, 17, 18, 24];
- focus group discussions [26, 32];
- face-to-face interviews [42];
- database search [47, 48].
One study did not indicate the detailed survey method used [16], yet was included because of its high relevance to the research topic.
The number of participants older than 65 years of study samples ranges from 11% to 100% of all initially included study participants (n= 19.604.711) from across the globe in countries from four continents and in following care settings: Community-dwelling or non-institutionalized citizens [9, 15, 21, 23, 27-31, 33, 36, 39, 42, 48, 50], nursing homes [12, 38], combined settings [32, 37], outpatient clinics [44], hospitals [18, 22], primary care centres/clinics or practices [24, 26, 43, 45], home-based primary care settings [10] and data bases such as the Medicare registry, national vaccine industry or settings of health services/insurance authorities [13, 40, 46]. In 12 studies, details about the care setting of the participants were not indicated [11, 14, 16, 17, 19, 20, 25, 34, 35, 41, 47, 49].
Target groups for evaluation of factors influencing vaccination uptake at time of inclusion of the studies mentioned were adults from 65 years and older [9, 21, 25, 27, 28, 31, 33, 35, 41, 42, 50] (n= 13.884.163), persons at risk or with a certain (medical) condition over 65 years or with a sub-analysis of subjects older than 65 years [11, 17, 34, 45, 49, 51, 52] (n= 452.901), adults aged ≥ 65 years in a care setting [10, 18, 22, 26, 37, 44] (n= 390.102), healthy adults aged ≥ 18 years with a sub-analysis of subjects older than 65 years [20, 29, 30, 39] (n= 559.966), community-dwelling adults aged 60 years and older [36, 47, 48] (n= 1.056.678), nursing home residents [12, 32, 38] (n= 2.682.324), health care professionals providing care for patients older than 65 years [24] (n= 2.535), Medicare beneficiaries and patients older than 65 years enrolled in a national health insurance program [40, 46] (n= 364.944), combined target groups [23, 43] (n= 49.038), other related authorities [13] (n= 16) and adults of other age groups with a sub-analysis of subjects older than 65 years [14-16, 19] (n= 162.044). Further information on the baseline characteristics of the 44 studies included in this scope review can be drawn from supplementary Table 1 (supplementary material).
3.2. Clustering factors influencing influenza vaccination uptake according to WHO determinants of health
Reviewers (R.R.-W, M.I., C.T., P.D., H.F., S.L.) were asked to cluster factors described to influence influenza vaccination uptake in the selected publications according to WHO social health determinants, such as income, education, occupation, social class, gender and race/ethnicity [2]. This process resulted in defining 41 determinants that affect influenza vaccination uptake in adults ≥ 65 years according to this review. The factors found to affect VU and VH were summarized at different levels: structural and intermediate determinants, all further clustered into policy and governance, provider and health care and patient level. Determinants most analysed in connection with seasonal influenza vaccination are: age (n= 32 articles), gender (n= 30 articles), healthcare utilization or accessibility (n= 23 articles), education (n= 19 articles), income/socioeconomic status (n= 17 articles) and types of chronic diseases (n= 16 articles). Factors mostly lacking evidence for determining influenza vaccination behaviour were attitudes and behaviour of physician providing care (n=3 articles), recommendations released by governmental bodies (n= 3 articles), level of care (n= 3 articles), dietary patterns, social networks and deprivation (each n= 2 articles), self-care (n= 1 article) and self-reported reasons, such as “no time” (n= 3 articles), “forgot” (n= 1 article), allergic reactions (n= 1 article) or “didn’t want it” (n = 2 articles). The remaining factors relate to personal experiences of the citizen, varying from influence of family/friend (n= 4 articles) to household arrangements/children, previous vaccinations, and other health parameter (each n= 14 articles). The final workup of information according to presence in literature included into this review can be seen in supplementary Table 2 (supplementary material).
3.3. Determinants and Ecosystem of factors affecting uptake of influenza vaccination globally in adults older than 65 years
Building on the clustering work presented in section 3.2., we aligned information collected from the publications listed with social health determinants as outlined by WHO in 2010 [2]. Table 1 shows determinants clustered by their likelihood to increase or decrease VU for each article analysed, respectively.
Table 1: Determinants associated with increase or decrease of VU
Author
|
VU increase (+) or decrease (-)
|
Determinants associated with increase or decrease, respectively
|
Byeon 2018
|
+ men
|
- · Having a spouse/being married
- · Former smoker/non-smoker
- · Walking activities
- · Health examination
- · Visit of public health centre
- · Hypertension, diabetes, cardiovascular diseases
|
+ women
|
- · Unemployment
- · Former smoker/non-smoker
- · Health examinations
- · Visit of public health centre
- · Hypertension, diabetes
|
- men
|
- · Good subjective health status
|
- women
|
- · Alcohol consumption
- · Good health status
|
Cha 2016
|
+
|
- · Receiving regular health screenings
|
Chang 2016
|
+
|
- · Receiving vaccination in previous year
- · Frequent use of outpatient departments
- · Undergoing health examinations in previous year
|
Kwon 2016
|
+
|
- · Recent history of health screening
- · Higher age (≥ 70 and ≥ 75)
- · Self-reported health status as unhealthy
|
-
|
- · Smoking
- · Low physical activity
|
Leung 2017
|
+
|
- · Face-to-face patient education and information material
|
Mo 2015
|
+
|
- · Female gender
- · Chronic diseases
- · Participation in community activities
- · Knowledge of the fact that vaccine is required every year
- · Lower perceived side effect
- · Lower IV price
- · Recommendations from healthcare providers
|
Oh 2015
|
+
|
- · Female gender
- · Increasing age
- · Having health insurance
- · Having medical check-up
- · Co-morbidities
- · Worse self-related health
|
Wershof-Schwartz 2013
|
-
|
- · Female gender
- · Rural residency
- · Low socio-economic status
- · Recent immigration
- · Being from/Having physician from former Soviet Union
|
Yu 2014
|
+
|
- · Perceived susceptibility
- · Female gender
- · Multimorbidity
- · Perceived disease severity
- · Perceived benefit from current vaccination
|
-
|
- · Post-vaccination discomfort
|
Dyda 2015
|
+
|
- · Female gender
- · Higher Body-Mass-Index
- · Requiring assistance in daily tasks
- · Reporting chronic diseases
|
-
|
- · Smokers
- · Non-English speaking country of birth
|
Regan 2017
|
+
|
|
Aguilar 2012
|
+
|
- · Major chronic conditions
- · High level of dependence
- · More visits to the General Practitioner (GP)
- · IV in the previous season
|
-
|
- · Female gender
- · Age < 80 or > 94 years
- · Immigrant status
- · Previous hospitalization
|
Barbadoro 2016
|
+
|
- · Role of local policy in favouring VU
|
-
|
- · Younger age (65 -79 years compared to ≥ 80 years)
- · Medium level education
- · Absence of chronic diseases
- · Smoking
- · No GP contact in the last 12 months
|
Caille-Brillet 2014
|
+
|
- · Getting vaccinated in previous 2 seasons
|
Domínguez 2016
|
+
|
- · 3 or more GP visits in the previous year
- · IV in any of the previous 3 seasons
- · 23-valent pneumococcal polysaccharide vaccination
|
Ganczak 2017
|
+
|
- · Younger age (< 70 years)
- · Living in urban area
- · Co-morbidities
- · Vaccinated family members
- · Being informed about vaccination
- · Willingness for vaccination next year
|
Giese 2016
|
-
|
- · Regarded as not necessary
- · Not thinking about it
- · Consider themselves not at risk
|
- Health care workers
|
- · Regarded as not necessary
- · Rarely getting influenza disease
- · Consider themselves not at risk
|
Godoy 2015
|
+
|
- · Physician has been vaccinated
|
+ Physician
|
- · Worried about infecting patients
- · Believe in effectiveness
- · Concerned about getting influenza disease
|
Hellfritzsch 2017
|
+
|
- · Higher co-morbidity level
- · Less likely to never have smoked
- · Higher prevalence of physical activity
- · Higher prevalence of major physical limitations
- · Need for assistance with activities of daily living (ADL)
|
Martínez-Baz 2012
|
+
|
- · More physician visits per year
|
-
|
- · Female gender
- · Age (65-69 years or > 95 years)
- · Hospitalized or diagnosed with any major chronic condition in previous year
- · Haematological cancer or dementia
|
Poscia 2016
|
+
|
- · Communication/Awareness campaigns: System of reminders, recalls, information
|
Shah 2012
|
+
|
- · Care home patients with & without dementia
- · Chronic diseases
|
-
|
- · Community-dwelling patients with dementia
- · Area deprivation
|
Spreckelsen 2018
|
+
|
- · Vaccination status before nursing home admission
- · Region (East-Germany compared to West-Germany)
- · Number of co-morbidities
|
Verger 2015
|
No associations found for age group ≥ 65 years
|
Vukovic 2018
|
-
|
|
Blank 2012
|
+
|
- · Good monitoring systems for VU rates
- · Sending personal letters offering free vaccination
- · Additional policy elements (e.g. awareness campaigns)
|
Banach 2012
|
+
|
|
-
|
- · Female gender
- · Black race
- · Living alone
|
Black 2017
|
+
|
- · Increasing age (≥ 85 years compared to 75-84 years and 65-74 years)
- · Female gender
- · Chronic medical conditions associated with higher risk for influenza-related complications
|
-
|
- · Race/Ethnicity: Non-hispanic blacks and Hispanic
|
Farmanara 2018
|
-
|
- · Younger age (65-70 years compared to > 70 years)
- · Lower education level
- · No chronic medical conditions
|
Hurley 2018
|
+
|
- · Receipt of any needed vaccine (tetanus, diphtheria, acellular pertussis or pneumococcal besides influenza)
- · Centralized reminder/recall system
|
-
|
- · Prior refusal
- · Male gender
- · Older age (≥ 85 years)
|
Kaljee 2017
|
Factors affecting VU in general
|
- · Healthcare access and utilization
- · Communication and information sources
- · Social networks
- · Disease experience, knowledge and perceptions
- · Vaccine experience, knowledge and perceptions
|
Khan 2018
|
+
|
|
-
|
- · Race/Ethnicity: Non-hispanic blacks and Hispanic
|
Lu 2014
|
-
|
- · Member of ethnic minority group
- · Lower education
- · Unemployment
- · Chronic conditions
- · Last routine check-up > 1 year ago
- · Absence of personal doctor
|
Lu 2018
|
+
|
- · Doctors visit, receiving provider recommendation
|
McIntyre 2014
|
+
|
- · Recommendation by, and trust in, health professionals
- · Believe in effectiveness
|
-
|
- · Fear of adverse reactions
- · Believe in resilience
|
Pereira 2019
|
+
|
- · High-dose vaccine if free of cost
|
Takayama 2012
|
+
|
- · Higher age
- · Prior diagnoses of chronic conditions (except myocardial infarction and stroke)
|
-
|
- · Non-white race
- · Lower household income
- · Lacking health care coverage
- · Smokers
- · Physical inactivity
- · Reporting days of poor physical health in past month
|
Wooten 2012
|
+
|
- · Believe in effectiveness
- · Higher education
- · Doctor’s visit during flu season
- · Believe in personal susceptibility
- · Little concern of side effects
|
Yokum 2018
|
+
|
- · Receipt of single mailed letters
|
Francisco 2015
|
+
|
- · Male gender
- · Slow gait speed
- · Social involvement
|
-
|
- · Higher level of education
|
Sato 2015
|
+
|
- · Higher age (≥ 80 years compared to 70-79 years)
|
Doherty 2016
|
-
|
- · Negative attitudes and beliefs regarding vaccination
- · Failure of health care provider to recommend vaccination
- · Lack of knowledge of vaccine safety and effectiveness
- · Perceived susceptibility
- · Lack of awareness of national recommendations
|
Kan 2018
|
Factors affecting VU in general
|
- · Demographic factors
-
- - Age
- - Sex
- - Living with others
- · Health promotion factors
-
- - Health status and self-perceived health status
- - Health habits and medical service use
- · Knowledge/information and its sources
- · Health behaviour factors
-
- - Threat perception
- - Perceived barriers
- - Cues to action
- - Behavioural beliefs
- - Subjective norms
- - Past behaviour
|
Thomas 2018
|
+
|
- · Reminder/recall by letter plus leaflet or postcard
- · Patient outreach by retired teachers
- · Invitations by clinic receptionists
- · Patient education by nurses/pharmacists
- · Patient counselling by medical students
- · Patient vaccination by nurses
- · Multiple recall questionnaires
- · Payments to physician
- · Physician reminders
- · Posters in clinics
- · Chart reviews/benchmarking
|
Table 1 illustrates the determinants associated with increase or decrease of VU for each article analysed, respectively. Green rows equipped with a plus sign indicate VU increase, red rows with a minus sign display determinants decreasing VU. Some studies did not clearly figure out results of VU increase or decrease but present factors affecting VU in general.
We found sizeable evidence highlighting a role for several factors at individual level, such as increasing age [11, 12, 20, 28, 34, 36, 39] and decreasing individual health status. The latter included declining functional status or having a chronic diseases, comorbidities or disabilities [9-12, 14, 19, 20, 22, 25, 27-29, 33, 34, 37-39, 42], that supported VU among older people. Results related to gender were divergent, as some studies reported higher VU in females [12, 19, 33, 34, 42, 43] and others presented higher VU rates in males [9, 10, 21, 31, 40]. Besides those epidemiological and health parameters, health beliefs and experiences with recent vaccinations seem to impact VU for seasonal influenza vaccination [9, 15, 17, 18, 23, 26, 32, 33, 38, 41-43, 47, 51]. Not surprisingly, older citizens with positive attitudes towards VU also reported on having other vaccinations such as pneumococcal vaccination [18]. Furthermore, life-style factors as smoking, low physical activity levels, inadequate diet and alcohol consumption seem to be negatively associated with VU [11, 14, 19, 28, 39]. Some studies point to the direction that a higher educational level or higher socio-economic status may support VU in the elderly [20, 29, 39-41, 50]. Strong evidence was found in the field of healthcare utilization, showing that older citizens with more GP visits, health examinations or screenings and medical check-ups are more likely to receive influenza vaccination [9, 11, 14, 16-18, 26, 28-31, 34, 41]. In addition, interventions such as reminders, patient information/education or recommendations by health professionals seem to positively affect VU [22, 32, 33, 44-46, 48, 51, 52]. Moreover, Godoy et al. [24] highlighted that patients whose physicians were vaccinated had a higher VU than those whose physicians were not. An important result gathered in the review is the impact of social inclusion into family or informal social networks, which has been shown to positively affect VU [21, 22, 26, 33, 50]. Only a few publications were found reporting results of interventions on system level, one study found out that countries with good monitoring systems regarding VU rates exhibit higher vaccination coverage on average [13]. Additional policy elements also have the potential to increase VU rates [11, 13]. Figure 2 summarizes the single elements influencing seasonal influenza vaccination uptake among older citizens detected during our search in an ecosystem.
Figure 2 shows the model of social health determinants adapted for the results obtained in this scope review on factors affecting vaccination uptake for seasonal influenza among citizens older than 65 years. The framework does not reflect numbers of publications found or numbers of participants included in the respective studies. It just gives an overall view on factors described for vaccination uptake and hesitancy currently described in literature.