|Author(s)||Xavier Bosch-Capblanch1, Christian Auer2, Sylvie Olifson3
|Affiliation(s)||1Swiss Centre for International Health, Swiss Tropical and Public Health Institute, Basel, Switzerland, 2Swiss Centre for International Health, Swiss Tropical and Public Health Institute, Basel, Switzerland, 3Division Maladies transmissibles, Office fédéral de la santé publique OFSP, Geneva, Afghanistan, 4|
|Country - ies of focus||Switzerland|
|Relevant to the conference tracks||Governance and Policies|
|Summary||Vaccination in Switzerland may not reach public health targets. The Swiss Federal Office of Public Health is developing a National Immunisation Programme to improve coverage based on an overview of available evidence. 22 studies and 18 systematic reviews were included showing modest effects in general. Interventions included information, vaccine prescription, new vaccination programmes, information and vaccination campaigns, organisational interventions, financial incentives, and combinations of the above. This overview provides evidence to inform a participatory process to interpret and assess its relevance to the specific Swiss policy context.|
|Background||Vaccination is one of the most cost-effective interventions to control disease. For vaccines to show public health effects, high coverage needs to be attained, for example, as high as 95% for measles elimination. However, there are gross differences in vaccination coverage between high- and middle- and low-income countries and within countries. Factors related to poverty and low education level of caregivers have been widely described as determinants of non-vaccination. Surprisingly, though, coverage is still relatively low in some high-income countries even among well-educated populations. Switzerland is not an exception; for example, vaccination coverage for the second dose of measles in children by the age of two was 88% (2011-2012), well below international recommendations.
Cognisant of these issues, the Swiss Federal Office of Public Health has recently launched an initiative to better understand the factors affecting the uptake of vaccinations in the country and to find out which interventions to improve vaccination coverage may be effective and relevant to the Swiss context. A review of these interventions was commissioned to the Swiss TPH (presented here) in order to feed into a national consultation process to translate research evidence into meaningful policies.
|Objectives||To search, retrieve and synthesise evidence on the effects of interventions to improve vaccination coverage potentially relevant to the Swiss health policy context.
To inform a participatory process to incorporate research evidence, alongside implementation and policy issues, in the elaboration of the Swiss National Immunisation Programme (NIP).
|Methodology||A systematic literature review (SR) following commonly accepted methods was devised. Two sources of evidence were searched: primary studies assessing the effects of interventions to improve vaccination coverage, carried out in Switzerland or neighbouring countries over the last 10 years; and published systematic reviews of interventions to improve vaccination coverage without any geographical or time limitation. Participants considered were people of any age targeted by routine vaccination programmes or campaigns, any intervention aiming at improving vaccination coverage was considered and tested against any control or without control, and the primary outcomes were changes in vaccination coverage.
A search strategy was produced by a specialist and applied to a wide range of sources including: Medline, EMBASE, PsycInfo, the Cochrane Library, Web of science and CINAHL. Hits from the search strategy were assessed by looking at titles and abstracts. Relevant studies where assessed against the inclusion criteria. At the stage of data extraction, some references were excluded, mainly because there were no quantitative estimates of the primary outcome. Data from included references was extracted by a single reviewer.
Analyses were carried out based on a list of categories of interventions: organisation / governance, human resources, financial resources (health system and users), vaccines management, information (systems and users), way of delivery, education and combination of interventions. The effects of interventions were summarised using standard measures of coverage change. Where possible, results were presented in relative change. Where more than one change of coverage was reported the median relative changes with ranges were presented. We did not attempt to undertake any meta-analyses, due to the large diversity of interventions, outcomes and vaccines. However, we synthesised the findings categorising the estimates in four groups, using arbitrary criteria: (A) effects do not favour the intervention either because the point estimate is below 1 or because the lower precision estimate is below 1; (B) effects favour the intervention with an estimate below 2; (C) effects favour the intervention with an estimate above 2 but with a lower precision estimate below 2; (D) effects favour the interventions with both a point estimate and an existing lower precision level above 2.
|Results||A total of 10,719 references were retrieved: 4,585 primary research studies and 6,134 SR. A large proportion of these references were judged as irrelevant (e.g. studies of vaccine efficacy, vaccine development or animal studies).
(1) Primary studies
122 references from Switzerland or neighbouring countries were relevant, of which 22 studies were included. Most of the studies showed only modest effects. The most remarkable effect was reported in a study which reported changes in influenza coverage among health care workers between two seasons (relative increase 2.56 and 2.85). Another study assessed information and vaccine prescription for parents and vaccination proposed to parents during hospitalization (absolute changes of +6.63% and +7.11% respectively. The only study which assessed an organisational intervention showed that a vaccination programme introduced in a maternity ward produced a relative increase of BCG coverage by 1.23. The most remarkable study on information provided to users looked at the effects of TV spots, press conferences, information to health professionals, an Internet site, and leaflets and posters upon improved influenza vaccination status in the elderly. Interventions affecting the mode of delivery were more promising in general (e.g. vaccination campaigns).
(2) 93 SR were classified as relevant, of which 36 were excluded in the first round and a further 39 at the stage of data extraction. One review reported on 11 different interventions, showing large effects of regulatory interventions requiring vaccination at school, child care and college entry (+15% change in coverage), use of standing orders (51%), incentives to users (8%) and combined interventions for women, children and infants (9%). An SR assessed a series of interventions to increase adult immunisation. Interventions with significant effects included: organisational interventions (adjusted OR 16), education to providers (3.21), reminders to providers (3.80), reminders to users (2.52) and financial incentives for users (3.42). Interventions targeting human resources showed mixed results. Three SR looked at financial incentives targeting providers with relative changes between 1.26 and 2.22. Financial incentives to users showed positive effects (adjusted OR 3.42). A number of SR looked at information interventions within the system, with heterogeneous results. A review which combined access and user targeted interventions suggested relatively large effects with absolute changes of 14%.
|Conclusion||The interventions most widely reported are those related to information to users (e.g. reminders) therefore, it is more likely to find studies with positive effects in these interventions. A second group of interventions widely addressed are those with combinations of strategies.
• All types of interventions have examples of modest effects, or no effects or even negative effects on vaccination coverage.
• Information-related interventions have been the most widely reported. At least one review showing promising findings dealt with mass media.
• Educational interventions targeting users showed very modest effects except for a single study (educational intervention involving health care staff).
• The most effective ways of delivering vaccinations were those more proactive, such as campaigns or approaching parents in wards.
• Organisational interventions were the least reported albeit with promising findings.What is the relevance of these findings?
Our overview provides not only evidence on the effects of interventions but also offers a comprehensive menu of interventions tested and allows the identification of interventions which may have been hardly tested or not at all. While there is little dissent on the need to use the best available evidence reported in SR to inform policies, this overview highlights the challenges in translating research evidence into policy, especially on health systems interventions. For example the bias in reporting different types of interventions, the lack of standardised interventions, the fact that organizational/structural interventions were the least reported and the heterogeneity of findings.This overview was carried out in the context of a participatory policy making process to support the contents of the upcoming Swiss NIP, and not as a pure research exercise. The interpretation of this evidence is, therefore, embedded into a deliberative process which extends for several months and which includes immunisation stakeholders of several constituencies. The deliberative process is the backbone of policy making and assesses the relevance evidence through several mechanisms: (a) deciding which interventions are suitable in the Swiss context; and (b) the magnitude of effects that can be expected. The deliberative process will culminate into a series of recommendations which will be based, not only on the best available research evidence, but on implementation and context specific considerations as well.