Cross-Shift Information Loss and The Standardization of Logbook Handovers: A Critical Review of Communication Failure and Continuity Protocols in Continuous Processing Operations

rather than on the words used (Lilian, Liadi, Yeboah, & Apelehin, 2020). 2.2 Omission Rather Than Distortion A distinction that runs through the empirical literature and that has direct design implications concerns the failure mode. Handover can fail because information is transferred incorrectly, or because information is not transferred at all. The evidence points strongly toward the second. Studies of handover content across domains consistently find that what is transferred is generally accurate. Operators do not typically misstate the position of a valve or the value of a pressure. What is lost is information that neither party identified as requiring transfer, either because the outgoing operator regarded it as too routine to mention, or because they had ceased to notice it, or because its significance depended on a future condition neither party anticipated. This is the general pattern in which the retrospective obviousness of the omitted item is an artefact of knowing the outcome rather than a property of the situation as it presented (Woods, Dekker, Cook, Johannesen, & Sarter, 2010). The same asymmetry between what is reported and what is noticed is documented in the near-miss and hazard observation literature, where the reporting rate is governed less by the frequency of hazardous conditions than by whether the observer classified the condition as reportable at the time (Arumosoye & Obriki, 2019; Obogo, Arumosoye, & Obriki, 2021; Obogo, Uduokhai, & Ozobu, 2019; Obogo, Ozobu, Nwafor, & Adio, 2025), and the judgement and decision literature supplies the account of why the classification is systematically biased (Fobellah, 2025a). This has a specific and important consequence: the omitted information is systematically the information whose significance is not apparent at the time of handover. A temporary bypass installed six hours earlier and functioning correctly is unremarkable to the operator who installed it and is precisely the item whose omission matters. The categories most affected are therefore identifiable in advance, and comprise incomplete work and its state, defeated or bypassed protective functions, accepted deviations from normal configuration, equipment operating outside its normal envelope, and the reasoning behind decisions taken during the shift. The proposition that a condition of this kind is detectable by continuous instrumented observation where it is not detectable by recall underlies the work on continuous hazard monitoring, digital safety data streams, and predictive safety analytics, which seeks to make such conditions visible independently of whether anyone thought to mention them (Obriki, Arumosoye, & Obogo, 2023; Obriki & Arumosoye, 2023; Obriki, Arumosoye, & Ozobu, 2025). The equivalent argument in operating technology security, in corridor and asset monitoring, and in aviation hazard management is that anomaly detection must be automated precisely because the anomalies that matter are the ones no observer flagged (Adegbite, Adebayo, & Ahmed, 2023; Adebayo, Adegbite, & Ahmed, 2023; Adebayo, Adegbite, & Ahmed, 2022; Dagodzo & Ahiaeke Patrick, 2020, 2022; Adeyelu & Dagodzo, 2024b). The last of these deserves emphasis because it is the least well handled by written logs. A log records that a set point was changed; it rarely records why, and the incoming operator who does not know why cannot judge whether the reason still applies. The reasoning is the element most vulnerable to loss and the element whose loss most degrades the incoming crew's capacity to act appropriately, since it converts an inherited configuration into an arbitrary one that will be neither maintained deliberately nor reversed deliberately. This is the mechanism by which the same unsafe configuration recurs without anyone deciding to reproduce it, which the behavioural safety literature treats as a problem of the conditions that make the behaviour reasonable rather than of the individual who performs it (Obriki & Arumosoye, 2020, 2022; Obriki, Obogo, & Arumosoye, 2023). A parallel difficulty is documented in work on the drafting and translation of administrative and registry records, where the terms used in a written entry carry meaning reliably only for readers who already share the context in which the entry was made (Kpoka, 2023, 2024). The healthcare literature reaches compatible conclusions. Studies of clinical handover consistently identify omission as the dominant failure mode and identify the rationale behind clinical decisions as particularly vulnerable (Patterson, Roth, Woods, Chow, & Gomes, 2004; Raduma-Tomàs, Flin, Yule, & Williams, 2011). Patterson and Woods (2001) examine handover in space shuttle mission control and reach the same finding in a setting closer to process operations in its technical character. The wider health services literature situates these findings within the operational and workforce conditions that produce them, and within the physical and organisational design of the settings in which care is delivered and transferred (Ilodigwe & Adesemoye, 2024b; Obi et al., 2025; Ogbete, Aminu- Ibrahim, & Ambali, 2018), and the continuity-of-care literature makes the transfer of custody between successive providers an explicit design object rather than an incidental feature of the schedule (Asiedu & Asiedu, 2024a). The four vulnerable categories map onto barrier and latent condition theory precisely. A defeated protective function transferred without its authorisation basis is the canonical latent condition (Reason, 1997; Hollnagel, 2004; Sklet, 2006; Center for Chemical Process Safety and Energy Institute, 2018; Johansen & Rausand, 2015; de Ruijter & Guldenmund, 2016; Manton, 2017), and the incident record documents the consequence repeatedly (Cullen, 1990; Health and Safety Executive, 1997; U.S. Chemical Safety and Hazard Investigation Board, 2014, 2015; Baker et al., 2007; Hopkins, 2000, 2012; Kletz, 2001; Mannan, 2012; Crowl & Louvar, 2011; Turner, 1978; Vaughan, 1996; Perrow, 1984; Rasmussen, 1997; Leveson, 2004, 2011; Dekker, 2011). 2.3 The Written and Verbal Channels Standardisation efforts in operating facilities are directed predominantly at the logbook, which is the tractable element: a template can be specified, its completion audited, and its content archived. The verbal handover is harder to specify, impossible to archive, and correspondingly neglected. The consequence is that facilities standardise the channel that carries less of the load. The preference is not peculiar to handover. Across procedural standardisation, service management and internal audit, the artefacts that receive design attention are those that leave an inspectable trace, and the activities that leave none are managed by exhortation (Eyetsemitan, Oyeleye, Ambali, & Fadayomi, 2022; Oyeleye, Eyetsemitan, Ambali, & Fadayomi, 2022; Ladapo, Dosunmu, Jooda, & Abolaji, 2023b; Obogo, Arumosoye, & Obriki, 2020a; Akomolafe & Agu, 2018). 2.3.1 Distinguishable functions The two channels perform distinguishable functions. The written log establishes a durable and auditable record, permits reference at any later point, transfers information whose significance is known in advance, and survives the departure of the person who created it. These are the properties that the records and data governance literature treats as the purpose of a controlled record: retention, traceability to a responsible author, and retrievability by someone who was not present when the entry was made (Aliliele, Mbonu, & Iwuanyanwu, 2024a; Mbonu, Iwuanyanwu, Uzoka, & Oluoha, 2019; Nnabueze et al., 2021; Annan, 2022). It cannot transfer emphasis, cannot be interrogated, and cannot adapt to what the receiving party does not understand. The verbal handover permits interrogation, conveys emphasis and uncertainty, adapts to the receiving party's existing knowledge, and permits the receiving party to signal incomprehension, which is the property that work on communication across differing frames of reference identifies as the one that distinguishes exchange from transmission (Lilian, Liadi, Yeboah, & Apelehin, 2020). It leaves no record and is subject to the recall limitations of both parties. 2.3.2 Matching traffic to channel The design implication is that each channel should carry the traffic it is suited to. Standing conditions, completed actions, readings and the state of documented work belong in the log. Uncertainty, reasoning, anticipated developments, and the emphasis distinguishing the two items that matter from the twenty that do not belong in the verbal exchange. Attempting to move the second category into the log produces logs that are long, unread and therefore worse than shorter ones. The general form of the point, that a process carrying more than it is designed to carry degrades rather than saturates, is developed in the workflow and process optimisation literature, where the identification of the binding constraint precedes any addition of steps (Eyetsemitan, Oyeleye, Ambali, & Fadayomi, 2024b; Eyetsemitan, Oyeleye, Ambali, & Fadayomi, 2024a; Tonoyan, Dada, & Ayivi-Donkor, 2024b). The engineering literature on communication channels makes the same point in its own terms, treating delivery reliability as a property of the channel and its load rather than of the message (Amayo & Popoola, 2017b; Asiedu & Quainoo, 2025. The face-to-face requirement identified in the guidance follows from this analysis rather than from convention. A handover conducted by leaving a written log for a crew arriving later has eliminated the channel that carries the more vulnerable information, and the practice is accordingly identified as a specific risk in the guidance. The value placed on co- presence at the point of transfer is visible in adjacent operational settings, from integrated network and security operations centres organised so that incoming and outgoing watches overlap physically (Ladapo, Dosunmu, Jooda, & Abolaji, 2024), to the safety leadership literature in which presence at the workface rather than review of the record is what distinguishes effective supervision (Obogo, Nwafor, & Ozobu, 2023; Obogo, Ozobu, & Uduokhai, 2019a; Arumosoye & Obriki, 2024; Obriki & Arumosoye, 2019). 2.3.3 The log as a working document A further consideration concerns the log as a working document during the shift rather than as a handover artefact. A log completed at the end of the shift is a reconstruction, subject to the recall and salience biases that reconstruction entails, and it will systematically omit items that seemed unimportant at the time and became important later. A log maintained contemporaneously is a different document with different properties, and the distinction is rarely made in standardisation efforts. The contrast between a record generated as a by-product of the work and one composed afterwards is treated directly in work on instrumented condition monitoring and on automated event logging, where the record is complete because no one is deciding what to enter (Omoegun, Sunday, Essien, & Oluokun, 2023; Obriki & Arumosoye, 2023; Obriki, Arumosoye, & Obogo, 2023; Mbonu, Iwuanyanwu, Uzoka, & Oluoha, 2019; Aliliele, Mbonu, Uzoka, & Iwuanyanwu, 2025b), and in version control practice, where the authoritative artefact carries its own change history and the identity of whoever made each change (Badmus, Jooda, Ozowara, & Anunagba, 2022a). The permit and isolation systems are the other custody transfer occurring at the same moment, and the two are ordinarily governed separately despite carrying interdependent information (Health and Safety Executive, 2005, 2006c; International Association of Oil and Gas Producers, 2018; Occupational Safety and Health Administration, 1989; Obogo, Arumosoye, & Obriki, 2020c; Obriki & Arumosoye, 2024b). The same interdependence appears at the interface between maintenance execution and the supply, readiness and regulatory-compliance systems that determine what work is open at any moment, which are likewise governed as separate custodial chains (Okonkwo, Agbabiaka, Ogunwole, Mayo, & Okeke, 2021; Okonkwo, Ogunwole, Mayo, & Okeke, 2021; Agbabiaka, Okonkwo, Ogunwole, Mayo, & Okeke, 2019; Ogunwole, Okonkwo, Agbabiaka, Mayo, & Okeke, 2021; Ogbete, Aminu-Ibrahim, & Ambali, 2019; Okoruwa, Babatope, Akokodaripon, & Akinleye, 2024). 2.4 Structured Protocols and the Mechanism Question Structured handover protocols, in which the exchange follows a defined sequence of content categories, have been widely adopted in healthcare and more variably in the process industries. The evidence that they improve information transfer is reasonably consistent. The mechanism by which they do so is not established, and the question is more than academic because different mechanisms imply different design choices. The implementation literature on structured management systems, in aviation safety management, in regulatory readiness, and in organisational change more generally, is similarly clearer about whether adoption occurred than about which component produced the effect (Adeyelu, 2019; Eze, Anene, & Akinleye, 2022b; Eyetsemitan, Ambali, Oyeleye, & Fadayomi, 2023a; Ilodigwe & Adesemoye, 2025b; Jimoh, Ahmed, & Fagbade, 2023). The content coverage hypothesis holds that the structure works by ensuring that each category is addressed, so that items which would otherwise be omitted are prompted. If this is the mechanism, protocol design should focus on category completeness, and the appropriate response to a failure is to add a category. This is the logic of checklist-based compliance benchmarking and of audit protocols constructed around coverage of specified elements, in which the completeness of the instrument is treated as the primary determinant of the outcome it measures (Adeyelu & Dagodzo, 2022; Adeyelu & Dagodzo, 2024a; Oloto & Yeboah, 2024). The attention allocation hypothesis holds that the structure works principally by requiring that the handover be conducted deliberately and by allocating time to it, and that the specific categories matter less than the fact of a defined process. If this is the mechanism, protocol design should focus on protecting the time and on the conditions of the exchange, and adding categories may be counterproductive by extending an exchange whose length is already the binding constraint. Comparable findings are reported in the training and awareness literature, where the measured effect of an intervention often tracks the deliberateness and protected duration of the activity more closely than the specific content delivered (Obriki, Obogo, & Arumosoye, 2022a; Onche, Ogbonna, & Adegbite, 2024). The available evidence does not discriminate cleanly between these, and both are probably operating. The practical implication is a caution against the characteristic organisational response to a handover failure, which is to add an item to the template. A template that has accumulated items through successive incidents becomes long, is completed by rote, and the completion by rote defeats both hypothesised mechanisms simultaneously. The same degradation pathway is documented in procedural and documentation systems elsewhere, where instruments that grew by accretion came to be completed for the record rather than for the purpose, and where implementation fidelity accordingly diverged from nominal compliance (Eyetsemitan, Oyeleye, Ambali, & Fadayomi, 2022; Oyeleye, Eyetsemitan, Ambali, & Fadayomi, 2022; Ambali, Eyetsemitan, Oyeleye, & Fadayomi, 2021; Yeboah & Oloto, 2023; Oloto & Yeboah, 2025). The shared mental model literature supplies the theoretical framing for what an effective handover achieves (Cannon-Bowers & Salas, 2001), and the situation awareness framework identifies what the incoming operator must reconstruct, namely perception of the current elements, comprehension of their significance, and projection of their development (Endsley, 1995). Handover that transfers the first level and not the second and third leaves the incoming crew with data and without understanding, which is the condition that permits an inherited configuration to persist unexamined. The non-technical skills framework treats communication as an assessable competence alongside situation awareness and decision making, which supplies a route to evaluating handover performance at the level of the individual rather than the protocol (Flin, O'Connor, & Crichton, 2008). Work on coordination in distributed project teams, on multi-site implementation, on human oversight of partially automated processes, and on the coordination of autonomous agents addresses the same reconstruction problem in settings where the transferring parties are separated by geography, organisation or kind (Amayo, Owulade, & Isi, 2023c, 2024b; Omo Enabulele et al., 2025; Ladapo, Dosunmu, Jooda, & Abolaji, 2022; Akanbi & Sunday, 2024; Liadi, 2024c). 2.5 Fatigue and the Timing Constraint Handover occurs at the end of a working period, which is to say at the point of lowest performance for the outgoing crew, and at the beginning of one, when the incoming crew has not yet established its operating picture. The scheduling is unavoidable and its consequences are underappreciated. Where the work is also conducted under thermal or environmental load, the decrement is compounded, and the error-causation literature treats the resulting condition as a property of the situation rather than of the person occupying it (Arumosoye & Obriki, 2018; Obriki & Arumosoye, 2020), and the wider physiological literature demonstrates that stress states of this kind are measurable by indicator rather than inferred from self-report (Oluwadele et al., 2023). The shift work literature establishes performance decrement as a function of time on task, of time of day, and of position within a shift cycle, with night shifts and the later shifts of a consecutive run showing the largest effects (Folkard & Tucker, 2003). The night-to-day handover therefore combines an outgoing crew at or near its performance minimum with a volume of overnight information to transfer, and the morning handover combines an outgoing crew that has worked through the circadian trough with an incoming crew arriving into the busiest part of the operating day. Two design responses follow. The first is that handover quality expectations should not be uniform across the cycle, and that the handovers occurring at the least favourable points warrant additional support, whether through longer allocated time, supervisory attendance, or a more complete written record prepared earlier in the shift. The principle that capability must be planned against the conditions expected at the moment of demand, rather than against average conditions, is the organising assumption of the emergency response readiness literature (Arumosoye & Obriki, 2023; Obriki, Obogo, & Arumosoye, 2022b; Aminu-Ibrahim, Ogbete, & Iwuanyanwu, 2025b). The second is that the practice of completing the log at the end of the shift is least defensible precisely where it matters most, since it places the reconstruction task at the point of lowest capacity. The guidance on managing shift work addresses the general question of fatigue management and is the appropriate reference for the scheduling dimension (Health and Safety Executive, 2006). What it does not address, and what the handover literature has not examined systematically, is the interaction between fatigue and handover quality specifically, which would be a straightforward study to conduct in an operating facility and which appears not to have been conducted. The shift work and fatigue literature is well developed and its findings bear directly on when handover quality can be expected to hold (Åkerstedt, 2003; Costa, 2010; Wagstaff & Sigstad Lie, 2011; Folkard & Tucker, 2003; Driskell & Salas, 1996; Staal, 2004; Baddeley, 1972). Rasmussen skill, rule and knowledge distinction explains why the reconstruction task at end of shift is the one most degraded by fatigue (Rasmussen, 1983, 1986; Reason, 1990, 1997; Woods, Dekker, Cook, Johannesen, & Sarter, 2010). Vigilance and sustained attention research explains why the end-of-shift reconstruction is the least reliable moment to perform it (Mackworth, 1948; Warm, Parasuraman, & Matthews, 2008; Parasuraman & Manzey, 2010; Wickens & Hollands, 2000; Folkard, Lombardi, & Tucker, 2005; Sallinen & Kecklund, 2010; Knauth & Hornberger, 2003). The design response available in other domains, where the reliability of human observation at the margin is treated as a fixed constraint rather than a variable to be improved, has been to move detection and early warning onto instrumented and analytical systems that do not degrade with time on task (Adelanwa, Basnet, & Anene, 2023b; Komi & Adamolekun, 2021; Komi & Adeniji, 2023; Sunday, Omoegun, Essien, & Oluokun, 2020; Sunday & Omoegun, 2022; Atta, Tofah, & Adenuga, 2021; Atta, Tofah, Kankam, & Adenuga, 2020; Atta et al., 2025; Asiedu & Quainoo, 2023b; Dosunmu & Ogundele, 2020, 2022, 2025c; Adegbite, Adebayo, & Ahmed, 2024b). 3. Discussion and Design Implications 3.1 Design Implications for Logbook Standardisation Seven implications follow from the review and are stated as design positions. First, specify the categories that must be transferred rather than the format in which they are recorded. A specification of content survives changes of medium; a template does not. The distinction between a requirement and its current implementation is drawn in the same terms in work on translating regulatory and operational requirements into executable workflows, and in the frameworks developed for specifying compliance obligations independently of the systems that discharge them (Eyetsemitan, Ambali, Oyeleye, & Fadayomi, 2025; Eyetsemitan, Ambali, Oyeleye, & Fadayomi, 2021; Adeyelu, 2021). Second, give explicit categorical treatment to the four vulnerable classes: incomplete work and its state, defeated or bypassed protective functions, accepted deviations from normal configuration, and the reasoning behind decisions. These are the categories the evidence identifies as omitted and they are not the categories templates typically emphasise. Making a specific class of condition the object of a named requirement, rather than leaving it to general diligence, is the mechanism by which the institutionalisation literature accounts for practices that persist beyond the tenure of the people who introduced them (Obriki & Arumosoye, 2021; Obriki, Arumosoye, & Ozobu, 2025; Obogo, Ozobu, Sanusi, & Ajirotutu, 2025). Third, require contemporaneous logging rather than end-of-shift reconstruction, and design the log so that contemporaneous entry is practical. This is largely a matter of accessibility and of not requiring the entry to be composed. The design pattern is familiar from instrumented monitoring and automated event capture, where the record is a by-product of the activity and the marginal cost of an entry approaches zero, and from capacity and resource planning systems that depend on data being written as events occur rather than assembled afterwards (Omoegun, Sunday, Essien, & Oluokun, 2023; Obriki & Arumosoye, 2023; Obriki, Arumosoye, & Obogo, 2023; Mbonu, Iwuanyanwu, Uzoka, & Oluoha, 2019; Edivri & Oteri, 2022). Fourth, protect the verbal exchange and require it to be face to face. Where the roster produces handovers without overlap, the roster is the defect. Treating a governance arrangement as the object of correction, rather than the behaviour it constrains, is the position taken in the literature on coordination across organisational boundaries and on governance alignment in jointly operated ventures (Obogo, Obriki, & Arumosoye, 2022; Obogo, Ozobu, & Uduokhai, 2019b; Amayo, Owulade, & Isi, 2023a, 2024b; Eyetsemitan, Ambali, Oyeleye, & Fadayomi, 2020; Isiekwu, Oluwo, & Dada, 2021). Fifth, resist template growth. Establish a review that removes items as well as adding them, and treat the length of the exchange as a design parameter with a limit. The corresponding practice in process improvement is the periodic removal of steps that no longer discharge the purpose for which they were introduced, and in the decision literature the recognition that instruments accumulate items in response to salient failures rather than to measured contribution (Eyetsemitan, Oyeleye, Ambali, & Fadayomi, 2024b; Oyeleye, Eyetsemitan, Ambali, & Fadayomi, 2022; Fobellah, 2025a). Sixth, require the incoming operator to state back the non-routine items rather than to