**5. Discussion**

Overall, the literature reviewed suggests that aviation professionals are concerned about the current state and application of MELs. Notably, all positions highlighted the importance of MELs and point to the utility of the MEL as a balancing factor between safety and operations where aircraft can be dispatched with inoperative equipment as long as safety is not compromised. However, it has been postulated that a holistic approach is required to streamline the development of a framework/methodology to support the development, maintenance and monitoring of MELs [27,32]. Regulatory authorities and aircraft manufacturers are expected to offer to operators more detailed MEL-related guidance and specific tools along with requirements for respective training programmes.

Given the increasing complexity of aircraft systems, coupled with issues identified in the literature such as different MEL standards worldwide and cases of MEL mismanagement and misapplication [26,29,33], it is important to reinstate the need for standardisation and reinforcement. The issues identified in the literature and revealed through our review of the accidents and incidents above could pose a serious problem and are still prevalent despite MEL was introduced in the 1960s. The urge for MEL standardisation followed by targeted interventions to ensure its consistent and substantive application has become quite undeniable according to aviation researchers and professionals such as Pierobon [27], Hertzler [29] and Yodice [42]. Although literature and previous studies do not argue that the MEL framework should be entirely reformed, its standardisation is expected to ameliorate current issues and support proper and justified customisation of MEL to the operational profile and needs of airlines, minimisation of ambiguity in its implementation and enrichment of respective training. Furthermore, harmonisation of MEL development will allow valid comparisons of practices followed across regions and operators and offer to airlines and NAAs a common reference baseline for knowledge exchange as well as possibilities for continuous review, update and improvement of shared MEL-related processes.

Furthermore, it seems that there is an assumption that the factors/parameters considered during the MMEL development are directly applicable to the MEL, while this may be valid to some extent as MMELs are developed for operators to use as a guide for their MEL development, it is important to note that the MMEL alone might not be entirely suitable and adequate for every operator. MMEL professionals attempt to anticipate the worst possible effect of systems failure, but they may not anticipate all probable scenarios and failure modes that can emerge during operations and stem from the complexity of aircraft systems and their interactions with humans and the environment [28].

The above was also confirmed during our analysis of MMEL/MEL related events under the category UNAMEL where professionals sometimes did not consider the history of failure of an equipment/system during the justification phase. Although it can be argued that the events under this category were random, their occurrence highlights the need for operators to customise their MELs to their environment and type of operations that can affect system/componen<sup>t</sup> performance rather than just duplicating the master MMEL document which is based on different datasets of failures and performance. In addition, despite most of the events analysed in our study were not fatal or catastrophic, the outcome severity of any future event cannot be guaranteed, especially when flight crews are unaware or unfamiliar with the problem and cannot exert full control over the unfolding situation [45]. The Spanair's crash studied by Grüninger and Norgren [28] was linked to an unanticipated MEL failure.

Moreover, it is interesting that, even under the current regulatory mandates around MEL, the importance of the latter might not have been understood completely across the aviation industry as indicated by the high frequency of non-compliant cases. For operators with approved MELs, it was observed that, in several instances, aircraft were dispatched with known inoperative equipment or defects even though the operators had MELs in place (category UNJMEL). In addition, all the events where operators dispatched aircraft without an approved MEL (category OPWMEL) regarded the US region. Most of the operators falling under the latter category were FAR 91 operators or regarded operations conducted under FAR 91. This confirms Hertzler's [29,30] call to operators to apply for MELs under FAR 135 because FAR 91 operators are the most neglected in terms of MEL oversight. The latter enjoy some leniency and do not utilise the MEL concept compared to FAR 135 operators where compliance with MEL and applicable MEL intervals are mandatory as mentioned in Section 3.2 above.

Another issue identified during the review of MEL events was the late rectification of MEL items (category LMEL). Indeed, EASA [4] and Airbus [3] stress the importance of repairing or replacing an inoperative item at the earliest opportunity and not at the most convenient time for an airline. However, although someone could argue a possible relationship of these cases with human error (e.g., lapses or slips) and non-compliance, these events can also be attributed to a lack of understanding of operators about the intended objectives and philosophy of MEL. The latter, instead of being approached as a constraint to operations, should be viewed as a risk managemen<sup>t</sup> tool that can help in evaluating operational risks and specifying procedures in maintaining safety margins. Nonetheless, we did not identify literature suggesting any direct links between the MEL and the risk managemen<sup>t</sup> framework of companies.

Furthermore, the cases associated with misinterpretations of MELs (category UNSCMEL) accord with the findings of FAA cited in Pope [39] and Munro and Kanki [26]. As stressed in the literature reviewed, the clarity of MELs and their related regulations along with MEL designated roles within airlines would facilitate the MEL review and development process and improve the reliability of MEL application [27,29,42]. Additionally, air operators need to train their pilots, engineers and aircraft dispatchers on MEL-related operational and maintenance requirements. Based on the nature of events under the specific category, it can be argued that adequate training could have led to anticipation of scenarios within the operator's operational environment and could have played a positive role. Furthermore, those currently involved in the MEL process might have little or no experience in airworthiness managemen<sup>t</sup> or competencies and skills in MMEL/MEL. Being type-rated on an aircraft does not necessarily mean that an engineer or pilot is able to fully understand the parameters/factors surrounding the development and application of the MEL and interpret it correctly. Such a situation might lead to adverse events like the ones studied by Grüninger and Norgren [28] and Pierobon [27].

Regarding the ten cases indirectly related to MEL (category DHIS), Airbus [8] highlighted that a logbook entry is the starting point for assessing MEL-related defects/deficiencies. Perhaps, in conjunction with the remarks stated above about proper training, engineers and pilots might not have understood the criticality of registering technical works and problems in logbooks. Undocumented maintenance, unrecorded/unreported defects and improper handover will reduce the information available to pilots, maintenance staff and engineers in making informed decisions about the status and serviceability of aircraft.

Finally, the traditional approaches highlighted in Section 2.1 above have been criticised because they do not consider visibly and methodologically the human interactions with systems [6,46] which are inextricable parts of aircraft operations and are closely related to the development and application of MEL. Due to the interconnectedness of elements and processes that increase the complexity of modern systems, there is a need for more holistic and nonlinear frameworks to system safety analysis. Recent socio-technical systems engineering approaches, which are built upon systems theory, consider the interactions and interdependencies between human and technology [6,47] and have introduced tools and techniques to tackle the limitations of traditional approaches. For example, Leveson [6] has proposed the Systems Theoretic Process Analysis (STPA) technique, Hughes et. al. [46] recommends the Systems Scenarios Tool (SST), while Mumford [48] introduced the Effective Technical and Human Implementation of Computer based System (ETHICS) tool. Although each approach is accompanied by limitations in its endeavour to understand and deal with complexity, these tools sugges<sup>t</sup> a more structured path to socio-technical systems modelling and offer a dynamic approach to systems safety engineering. While such techniques are relatively new compared to FMEA and FTA and perhaps more resource-demanding in their application, they are promising in overcoming the limitations highlighted in Tables 2 and 3 above and, apart from the proximal technical and human components of aircraft operation systems, could also account for various complex roles aviation stakeholders hold in the MEL development process and consider contextual parameters (e.g., specific NAAs policies and strategies, cultural and societal factors).
