Whilst defining handover and its purpose is challenging given its varied nature, another difficulty is measuring its effectiveness and the effect of changes made. Literature on handover is limited considering its importance to patient safety. Despite limited literature the methodologies used are vast and include surveys [12], simulated with direct observation [13], and monitoring of data collection tools [14].

Surveys can be useful in exploring some aspects of handover such as physician knowledge, behaviours, and attitudes, overall satisfaction with the process, and can delve into the cultural norms. The use of surveys to determine the accuracy of data transfer is, however, limited by recall bias.

Bhabra et al. use a novel method to test handover methods by simulating handovers [13]. The authors of this paper created a number of fictional patients with set data points required for handover. They then had junior doctors simulate handing over these patients in an environment similar to their usual environment. The handover was audio taped and directly scored by two observers present in the room. This allowed observers to analyse the entire handover process and use the audio recordings to gain consensus amongst observers [13]. It is unknown whether the simulated nature of this experiment, and the presence of observers, could have altered the handover given by the junior doctors under study thereby introducing a level of observer bias. Caution must therefore be exercised in extrapolating these findings to actual practice.

Another methodology has been to collect and analyse handover documents used in actual practice [15]. This can allow a direct measure of the data transferred in written format without directly influencing the handover process. This method is limited however as it cannot capture the verbal aspects of handover and may underestimate the amount of data transferred as some of this may be done verbally.

The way handover should be delivered or facilitated continues to be a source of debate. One consensus however seems to be that verbal handover alone is often inadequate and carries a higher risk of data loss. In an experimental comparison of handover methods Bhabra et al. concluded that verbal only handover was associated with 66 % data loss after the first handover [13]. Note-taking improves handover with only 8 % of data lost when notes were taken, if pre-printed computerised notes are used data loss reduces to 0 % [13].

Ferran et al. looked at the effect of using standardised handover proformas in an audit against the Royal College of Surgeons of England guidelines [6] for acceptable minimal data sets to be handed over. The authors found that when doctors handed over using their own handover sheets only 72.6 % of data was handed over [14]. In 8 % of cases date of birth was not handed over and in 2 % unique hospital identifier was not handed over. With the introduction of a standardised handover proforma, which prompted the use of pre-printed labels, the authors saw a significant improvement in data handed over from 72.6 to 93.2 %. They concluded that standardised proformas were a practical way to accumulate handover data as a doctor progressed though a busy shift without the need for regular stops at a computer to input data [14]. This method may have continued relevance in an environment where costs are a priority and fully integrated patient record software is not available.

Studies have looked at using computers to assist in handover. The introduction of electronic patient records has allowed the development of handover software that can integrate with hospital systems to download patient identifiers, vital signs, laboratory results, and progress notes, into a template for handover [12]. The advantage of such a system is that it allows data to be standardised and pulls the data from already existing records thus preventing the need for duplication of data entry. In this study the authors found that doctors reported fewer patients missed on rounds, more time spent seeing patients, and better quality of handover [12].

The continued reduction in doctors working hours worldwide means that handover is not only here to stay but is likely to increase in frequency. Handover remains a weak link in patient safety and there are several established limitations; training, definition of function, measurement and delivery. While the literature is growing in this area handover research remains limited.

Training in handover techniques and the risks to patients of poor handover is likely to be key to the battle to improve patient handover. It is enlightening that medical schools do not agree that handover training is an issue for undergraduate education [8]. With doctors having to perform handovers as soon as they graduate it is difficult to understand the argument that safe and effective handover training should not be part of the undergraduate curriculum. While the debate goes on as to the timing of the delivery of this training some researchers and educators are already focusing their efforts on developing handover training that is pedagogically sound.

Darbyshire et al. reported on their design of a handover training session, based around Gagne’s nine events of instruction [16], which are mapped to the three key aspects of handover [11] (Table 17.2). The session featured group discussions, role play, handouts, and a video on handover and had good feedback from medical students taking part in the sessions [16]. This model for a handover training session appears robust and may be a good framework on which other institutions can build training sessions suited to their local circumstances.