Background

Randomized controlled trials (RCTs) are considered to be the gold standard for assessing the effects of a treatment. However, RCTs are costly and usually involve a relatively brief treatment period with limited follow-up. A treatment response restricted to this brief “in-trial” period can potentially underestimate the long-term benefits of treatment and also may fail to detect delayed hazards.

Post-trial follow-up (PTFU) is defined here as extended follow-up which starts after the end of the scheduled period of the original trial. Longer term follow-up of trial participants is important as persistent effects may be detected years later after treatment cessation [1] or even enhanced benefits observed decades later—a so-called “legacy-effect” [2]. Furthermore, delayed hazards may only emerge several years after exposure to certain treatments. Therefore, PTFU may add significant scientific value to the evaluation of many healthcare interventions.

There is a wide literature describing the importance of completeness of follow-up during the in-trial period of a RCT, without which the unbiased ascertainment of outcomes may be compromised and statistical power considerably reduced [3]. Many strategies to enhance follow-up during RCTs have been investigated and this remains an area of much ongoing research [4]. Without high quality in-trial follow-up, the value of post-trial follow-up will be extremely limited.

By contrast, little research has been done to evaluate methods for PTFU. Face-to-face follow-up is widely used during the initial "in-trial" period, but is costly if employed longer term. Telephone-based approaches are more practical, with the ability to contact many participants coordinated by a central trial office, and postal follow-up has been shown to be effective [1]. Web-based techniques may become more widespread as technological advances develop [5].

The use of routine health records can provide detailed information relatively inexpensively [6], but the availability of such data and rules governing access to it varies across countries. In the UK, Health Episode Statistics (HES) are held by the Health and Social Care Information Centre (HSCIC) and can be used as a streamlined method to follow-up trial participants. These routinely collected electronic health records include diagnostic codes (ICD-10) for hospital admissions and can be supplemented with mortality records and cancer registry data.

Methods/design

Eligibility criteria

Study designs

All published, health-related RCTs which have recruited more than 1000 participants and implemented PTFU are to be included in this systematic review. The RCT must have reached its scheduled end before PTFU commenced. Only studies published between 2006 and 2016 will be included.

Health-related interventions will include medical (licensed or unlicensed drugs), surgical, or psychological treatments. There will be no time limit of post-trial follow-up (Table 1).

Table 1 Selection criteria of published articles eligible for systematic review
Full size table

Participants

Trials including participants aged over 18 years old are eligible.

Interventions

Methods and incentives (monetary or by other means) used for post-trial follow-up including direct “face-to-face” follow-up and indirect follow-up, eg, medical record review, telephone and postal follow-up, and electronic follow-up including access to electronic health records will be included.

Comparators

Methodology used to follow up participants’ post trial will be compared qualitatively in a table format.

Outcome measures

Included studies must have published the total number of participants followed-up compared to the total number alive at the end of the in-trial period to calculate retention rates. Where available, secondary outcome measures of cost, incentives used for follow-up, and cost-effectiveness will be recorded and assessed. If there are missing data, an attempt to contact the study authors will be made. Further exploratory comparisons will be made depending on the information available (for example, describing the use of different approaches according to context, such as regional variations or comparisons of industry-funded trials versus those funded through other sources).

Language

Only studies published in English will be included.

Search methods

Electronic searches

The electronic search strategy includes the last 10 years of published articles using broad search criteria (Appendix). Searches for eligible studies will take place in a structured, step-wise process. A screening log will be kept. Results of searches from each electronic database and registries will be logged. The following electronic databases will be searched:

  • Cochrane methodology group register

  • Cochrane Central Register of Controlled Trials (CENTRAL)

  • MEDLINE®

  • EMBASE

Other sources of searches will include the following trials registry:

Screening for eligible studies

One reviewer will compile the titles and abstracts of all citations retrieved from the electronic database searches and order these by record number in Endnote® reference management software. Duplicates will be removed using the “deduplication tool” [7]. The screening process will involve two reviewers. The first 10% of abstracts will be screened by both reviewers independently. Concordance of 95% between both reviewers’ decisions on screening will be sought. If concordance is not reached at this point, discrepancies will be discussed and reviewed (including consultation with a third reviewer if necessary), and a further 10% of abstracts will be reviewed (Fig. 1). Once concordance has been reached, the remaining records for screening will be shared equally between the two reviewers and abstracts will be checked for eligibility. All records that are considered to be eligible will be confirmed by both reviewers. Full-text papers will be requested for all potential eligible papers.

Fig. 1
figure 1

Process of screening abstracts and checking for concordance between reviewers

Full size image

Data collection and analysis

Data extraction and management

Two reviewers will follow a similar step-wise process for data extraction (Fig. 2). A data extraction form will be used, and data extracted from all eligible studies will be compared qualitatively. All data regarding the intervention, the participants (demographics), attrition, retention, incentives used, and if specified, costs of PTFU will be extracted. If required data items are not available in the published article, the study’s corresponding authors will be contacted. If no response is received after two further attempts or from an alternative contact, the study will be excluded from the analysis but recorded on the PRISMA diagram and in an appendix.

Fig. 2
figure 2

Process of extracting data and checking for concordance between reviewers

Full size image

Assessing the quality of the post-trial follow-up methodology

In order to investigate whether the quality of the post-trial follow-up methodology might be predicted by the quality of the methods used for the original trial, risk of bias will be assessed in those trials chosen for data extraction using the Cochrane Risk of Bias tool. The tool will be applied to the methods used in the main trial, (not the PTFU) focusing on incomplete data; outcome reporting; for-profit bias and other bias sources. Two reviewers will independently assess the risk of bias, and disagreements will be resolved by a third reviewer. The assessment of bias results will be taken into account as part of the assessment of quality of the PTFU methods used.

Presenting and reporting of results

The Preferred Reporting Items for Systematic Review Protocols (PRISMA-P) [8] will be followed, including a PRISMA diagram to illustrate the process of selecting eligible studies (Fig. 1). Using the PRISMA guidelines (Additional file 1), the results of this review will be presented and the outcomes tabulated with respect to the different methodologies used in a qualitative and comparative style.

Interpretation of findings

The findings of this review will be discussed and potential limitations considered.

Discussion

Large randomized trials are essential for determining the magnitude of the effects of an intervention. Post-trial follow-up of large RCTs is important, not only for defining the effect of an intervention long-term but also for ascertaining the safety profile and potential hazards which might not be apparent during the relatively brief in-trial period. However, randomized trials can be very expensive, and funding is limited, hence streamlined and effective methodology for PTFU is desirable. This systematic review aims to inform the design of post-trial follow-up for a wide range of randomized trials.