The Effect of AUDIT and Feedback on Pneumococcal Vaccination Coverage

Background and rationale:

In March 2020, the INTEGO group published a study on pneumococcal vaccination status in Flanders (De Burghgraeve et al. The incidence of lower respiratory tract infections and pneumococcal vaccination status in adults in Flemish primary care. 2020, Acta Clin Belg. (DOI: 10.1080/17843286.2020.1735113)). Unlike pediatric vaccination, there is a low pneumococcal vaccination coverage rate (VCR) in adults in Flanders (187/1000 risk patients), which is possibly a public health problem. Recently the Belgian Superior Health Council published a new scientific recommendation on pneumococcal vaccination in adults (https://www.health.belgium.be/sites/default/files/uploads/fields/fpshealth_theme_file/hgr_9562_vaccinatie_tegen_pneumokokken_vweb.pdf). It might be beneficial to use the INTEGO registry to monitor and improve the vaccination coverage rate. This could be done by including AUDIT and feedback (A\&F) loops to evaluate and improve the quality of care provided to patients. A\&F is a well-known quality intervention that according to the last Cochrane review leads to "small but potentially important improvements in professional practice" (Ivers N et al. Audit and feedback: effects on professional practice and healthcare outcomes. Cochrane Database Syst Rev 2012, 6:Cd000259). Furthermore, the data generated by this study will also support the implementation of a Publicly Funded Program for pneumococcal vaccination for older adults in Flanders and generate the needed data and insights to prepare for the launches of future pneumococcal vaccines in Belgium. Therefore, this study will investigate the effect of A\&F on pneumococcal vaccination coverage in adults at risk in general practice.

Design:

This study will be a cluster-randomized trial: primary care practices will be randomized and divided into a control and intervention group. Allocation of control and intervention groups will be done by simple balanced randomization (1:1).

Objectives:

The primary study objective will be to assess the effect of "clinical AUDIT and feedback" loops on the pneumococcal vaccination coverage rate in adults at risk in general practice. Secondary objectives will be to explore whether the increase in vaccination coverage rate (VCR) after implementation of A\&F loops is different in specific subgroups (risk groups, male/female, age, smoking status).

Intervention:

Standardized automated AUDITs are available in most EHR systems in Belgium. Feedback, on the other hand, is only available in specific registration networks. At study baseline the performance in INTEGO practices will be measured. Afterwards a standardized clinical AUDIT to identify patients that may benefit from pneumococcal vaccination will be implemented in the EHR of all INTEGO practices. The Feedback will only be implemented in the intervention group. Like this we will be able to see the effect of individualized extended feedback on pneumococcal vaccination coverage on top of a standardized automated clinical AUDIT. This is important to know because the organization of feedback takes much more effort and cannot be easily scaled up to practices outside the registration network. So when a significant effect of feedback is seen on vaccination coverage this would have an important impact on the policy of future pro-active care.

Intervention group:

the INTEGO practices that will be assigned to the intervention group will receive an extended electronic feedback report with multiple components, directly implemented in their EHR, on the pneumococcal vaccination coverage in adults at risk in their practice ('push system'). There will be a direct connection between the EHR of the practice and a SAS visual analytics tool in the Healthdata environment (single-sign-on connection), that will show the extended feedback. This report will be available at baseline and updated every two months based on the current situation. They will be able to see the performance of their GP center, benchmarked to the mean VCR of the 10% best performing centers in the INTEGO network. The pneumococcal VCR will be presented in the three risk groups as defined in the eligibility criteria. The aim of the push system is to involve participating GPs and to actively direct their attention to the task at hand.

The extended feedback report will be delivered in the form of action plans and goals consisting of multiple components:

• The performance results of the audit will be compared with the mean of the 10% best performers according to the achievable benchmarks of care (ABC) method (Weissman NW et al. Achievable benchmarks of care: the ABCs of benchmarking. J Eval Clin Pract 1999, 5:269-281).
• A low cognitive load of the feedback where the results will be presented with the help of graphs and without any in-depth elements.
• Guidance on how to implement pro-active strategies to improve the vaccination coverage rate by using the clinical AUDIT available in their EHR.
• Links to disease specific guidelines. In a second step the GPs will find a clinical AUDIT in the statistical module of their EHR. The AUDIT is basically a search for patients in the practice population that apply to well-defined criteria. Three searches will be available based on the risk groups as defined in the eligibility criteria. When the GPs perform the searches they will get a list with patients eligible for pneumococcal vaccination, which were not vaccinated against pneumococcal infections. Based on this list they can perform actions in their EHR, like putting a task in the individual files of the selected patients or sending an invitation for vaccination and a prescription to the selected patients.

Control group:

every GP center assigned to the control group will only have access to the clinical AUDIT to identify patients that may benefit from a pneumococcal vaccination. GP centers in the control group will not receive an extended feedback report at baseline and every 2 months afterwards.

Sample size:

In total, 36 GP centers will be included, or 18 centres in each group to measure an absolute increase of 100/1000 risk patients in our primary outcome (80% power, alpha of 0.05) starting from a baseline vaccination coverage rate of 200/1000.

Assignment of interventions

Sequence generation:

Allocation of control and intervention groups will be done by simple balanced randomization (1:1). Enrollment of practices based on the in- and exclusion criteria, will be performed by researcher A, who is part of the trial. This will generate a numbered list of the eligible practices. Researcher B will, independently, assign numbers (denoting a specific practice) to either the control or intervention groups by means of a computer-generated list of random numbers, i.e. simple randomization.

Concealment mechanism:

Researcher B will prepare sealed, opaque envelopes containing a paper that assigns a specific practice number to a study group, based on this randomly generated number sequence. The process of the envelope preparation up until sealing and storage in a locked compartment will be videotaped by researcher B, who will thereafter be excluded from every other aspect of the trial. Allocation papers should never be visible, only the envelopes and the numbers on them. The video will be stored on an external device, which will be put in the locked compartment.

After practice enrollment, researcher A will access the locked compartment with the envelopes and review the accompanying video to ensure proper envelope preparation. Without opening or tampering with the letters, researcher A will write the appropriate mailing address on the respective envelope, based on their numbered list of practices. The envelopes will only be opened by the practices if they have an unbroken seal. This protocol is adapted from Radford et al. (Radford JA et al. Effectiveness of low-Dye taping for the short-term treatment of plantar heel pain: a randomised trial. BMC Musculoskelet Disord. 2006;7:64).

Statistical methods:

Statistical methods for primary and secondary outcomes SAS and R will be used for analysis and for the graphs of the feedback. For this study data will only be used on an aggregated level (per GP center). No analyses will be performed on the patient level.

To evaluate the effect of the intervention on the primary and secondary outcome measures, a logistic generalized estimating equations (GEE) model will be used. This model is chosen because it can investigate the average response of an intervention on a population level. The effect of the intervention will be presented as the difference in proportions together with its 95% confidence interval.

Conclusion:

The data generated by this study will also support the implementation of a Publicly Funded Program for pneumococcal vaccination for older adults in Flanders and generate the needed data and insights to prepare for the launches of future pneumococcal vaccines in Belgium.