In this study, older people with a medical card were twice as likely to be vaccinated against influenza, controlling for age and clinical need. Uptake rates were consistently higher among those with a medical card across all risk groups. In our study, 74% of those aged 65 years and older had received the flu vaccination in the past. However, given that the study assessed having ever received the flu vaccination, coverage in the most recent campaign is likely to be much lower.

The results suggest that the partial public reimbursement policy is leading to inequalities in vaccination coverage among those with established clinical need recommended to receive the flu vaccination. The findings suggest a step-like access to services in the Irish primary care system; those on the lowest incomes qualify for free access to services in general practice while those whose incomes are just above the threshold for medical card eligibility face difficulties accessing services, regardless of clinical need. The results may also reflect the wider influence of medical card status on behaviour. Medical card patients have significantly higher probability of visiting their GP compared to those with private insurance, even when differences in age and health status are considered [22]. A recent study estimated that people with a medical card aged 70 years and over attended almost 10 times a year on average [23]. In this study, private health insurance was not significantly associated with vaccination coverage illustrating the lack of additional benefit conferred by private health insurance in terms of accessing primary care services in Ireland.

Countries that reimburse healthcare practitioners for administering the vaccine or those that provide vaccination within their public health insurance schemes tend to have higher vaccination rates [6,24]. In Ireland, GPs are paid a nominal fee by the Health Service Executive for administering the vaccine to patients with a medical card. Results from the European SHARE project indicated that preventive service use was higher in health systems where doctors were paid a fee-for-service, however the effect was not significant for flu vaccination [7].

Advice from health care professionals has a significant impact on vaccination uptake [6,13,25,26], however it was not possible to control for this and other physician or system-related confounders in the analysis. Few studies have examined these factors which may contribute to variability in flu vaccination rates. A recent study by Blank et al. suggested an association between features of European vaccination policies and coverage rates in the elderly [27]. Countries which issued personal letters offering free vaccination showed higher coverage rates. This approach is not universally implemented in Ireland [27] and the practice may be restricted to those with a medical card who are easily identifiable within a practice computer system. Other organisational interventions shown to have a positive impact on vaccination rates include standing orders in the hospital or primary-care settings, a patient mailing system [28], patient reminder and recall systems in primary care settings [29] and providing the vaccination free of charge [30].

The association between age, chronic illness and vaccination coverage found in this study is well established in previous research [6,7,13-15]. Within the clinical risk group, we found that those in younger age group (50–64) had lower vaccination coverage (53%) than those in the older age group (81%) (≥65 years). While it could be argued that the older group have a longer lead-in time to ever have been vaccinated, there is evidence that age-based immunisation policies are associated with higher uptake among clinical risk groups than standard risk-based policies. In a study of vaccination uptake among clinical risk groups, Sammon et al. found that uptake was higher among age groups where universal flu vaccination was recommended compared to age groups in which only clinical groups were recommended for vaccination [5]. The Centers for Disease Control and Prevention (CDC) suggest that this may be due to the simpler health message attached to age-based policies, eliminating the need for GPs to identify high risk individuals [31]. It is estimated that implementing a 100% vaccination rate programme for all risk groups in the five biggest European countries would require an additional investment of €1.52 billion euro and would result in estimated savings of €39.45 million from reduced primary care visits and further savings of €1.59 billion from reduced hospitalisations [32]. The Irish government has proposed the introduction of universal health insurance including free primary care services. Plans to phase in this reform by providing free GP care to those with chronic illnesses in the first instance were replaced with proposals to start with children aged 6 years or younger. Our study highlights some of the potential pitfalls of a piecemeal approach to access based on income.

A nationally representative random sample participated in TILDA, based on stratified cluster sampling to minimize selection bias [19]. The main limitation of this study is the absence of information on the timing of vaccination, making it necessary to use having ever been vaccinated as the outcome measure. Studies in other countries examined coverage in the previous 12 months [7,33] or the last vaccination campaign [14]. It was not possible in this study to ascertain whether an individual was entitled to a medical card or whether he/she was in a high risk group at the time of vaccination. However, past insurance status has been found to be a significant predictor of current insurance status in Ireland [34]. Similarly, several studies have demonstrated the positive impact possession of a medical card has on GP use in Ireland [22,35]. While we cannot assert definitively that possession of a medical card now will increase the likelihood of vaccination now, the observed persistence in insurance status and numerous studies detailing the relationship between current medical card status and current GP use strongly indicate that this is a reasonable inference to draw from these results. This paper has important policy implications as it highlights the weaknesses of a health system where access to services is based on income rather than need. The pattern detected in this study endures after adjustment for age, income and clinical need.

Self-reported vaccination was above 70% for those with a medical card in high risk clinical groups, although this rate varied between illness groups. While a self-reported outcome measure is open to recall bias, studies have found high sensitivity and a high level of agreement between medical records and self-reported vaccination [36,37]. It was not possible to account for all high risk groups in the dataset, for example data were not recorded on renal disease and identification of carers was based on receipt of state allowance, which represents only a minority of carers in Ireland [38].