Australian paramedics’ perceptions of reasonable overtime: a Delphi study Authors: Matthew J Ferris, Kelly-Ann Bowles, Aislinn Lalor, Alexander P Wolkow Published: 2025 ABSTRACT AbstractBackgroundParamedics frequently incur reasonable overtime (RO) due to factors such as staffing, patient acuity and transport durations. Considering the risks associated with extended work hours, understanding paramedics’ opinions of reasonable and unreasonable overtime is vital for safe workforce management. The current study aimed to gain consensus on operational paramedics’ perception of conditions that are reasonable and unreasonable for overtime.MethodsThe Delphi methodology comprised three iterative survey rounds completed by experienced paramedics working in metropolitan areas of Australia. In Round 1, participants gave a free-text definition of RO and voted on the degree to which a predetermined list of RO-related conditions were reasonable. Free-text responses underwent a content analysis to determine additional RO-related conditions based on the definitions of RO provided by participants. These conditions were then combined with the predetermined list of RO conditions and voted on by participants in Round 2 and 3.ResultsOut of 189 initial participants, 144 completed Round 2 (76.2% response rate) and 128 completed Round 3 (88.9% response rate). Consensus (≥ 80% agreement) was achieved for 50% (n = 17) of the RO-related conditions. Nearly all paramedics agreed that overtime was reasonable for time-critical emergencies where there is an expected detrimental effect to a patient without a paramedic’s attendance (96%), whereas nearly all paramedics agreed that RO to ‘clear the board/pending queue’ was unreasonable (99%).ConclusionsParamedics reached consensus on four reasonable and thirteen unreasonable conditions for RO, resulting in the creation of the first set of consensus-based considerations for RO from the perspective of operational paramedics in Australia. These findings provide important insights into how paramedics perceive RO. Future research examining the proposed considerations for RO is needed to determine their feasibility and acceptability by ambulance services and other paramedics, and their potential effectiveness in this industry for workforce planning.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12913-025-13736-z.Keywords: Delphi study, Health and shift work, Overtime Introduction Australian paramedics complete rotating shift schedules involving long hours (e.g., 8–14 h shifts) to provide continuous emergency care to the public [1–4]. These shifts frequently extend beyond their scheduled finish time due to incidental or compulsory overtime, referred to as Reasonable Overtime (RO) [5, 6], as clinicians are required to be response-ready until the final minutes of their shift [7]. Although currently not defined, RO is often described as the time paramedics are required to work after their rostered shift finish time due to end of shift duties, including completing patient care, hospital ramping or returning the ambulance back to station or to a state of operational readiness [8, 9]. Unlike other healthcare professions, such as nursing, where personnel can handover the care of a patient to personnel on the incoming shift, paramedics are required to complete all end of shift duties prior to terminating the shift. As a result, a paramedic’s rostered shift duration can frequently extend to ≥ 13 h due to RO, which may lead to heightened fatigue levels [10, 11]. Overtime is widespread in the Australian paramedicine industry. A 2025 paramedic workforce report which included 694 Australian participants found 100% of these personnel work some form of overtime or on-call on a fortnightly basis [12]. Specifically, the report highlighted 77% of paramedics worked an extra 1–20 h per fortnight, while the remaining 23% worked 21 to >100 h a fortnight, however the report did not differentiate between additional voluntary rostered shifts and incidental/RO [12]. Research overseas suggests this is not unique to Australia, with a study in the United Kingdom (UK) showing RO occurred daily among paramedics, and 34% of these paramedics reported RO was a problem [13]. The impacts of overtime and extended shifts are well established, with research linking RO to increased fatigue, sleepiness, reduced vigilance, errors, poor work life balance and burnout [14–16]. Fatigued paramedics pose a significant workplace hazard to both themselves and the patients for whom they care for. For example, fatigued paramedics are more likely to make an error [6] and miss clinical deterioration [16]. Despite these consequences, most Australian jurisdictional ambulance services stipulate in their enterprise agreements that paramedics may be required to undertake ‘reasonable overtime’ [5, 6, 17–21], though the conditions and application of RO can vary widely [8, 9, 18–22]. To the author’s knowledge, no definition of RO has been universally adopted across these services in Australia. Additionally, although enterprise agreements are negotiated between employers and employee representatives, limited research has taken a systematic approach to understanding and clarifying paramedics’ views on the topic of RO and what they consider reasonable versus unreasonable. Understanding these perceptions is vital, as working overtime is not solely an operational issue but also impacts the psychological health of personnel [22]. Prior literature suggests that when employees are subjected to perceived unfair or unreasonable job demands it can lead to heightened stress, burnout and job dissatisfaction [23, 24]. Therefore, while understanding paramedics’ perceptions of RO may not directly change how often RO occurs, these insights tell us how RO is experienced by personnel, which in turn, can have implications for their health. Furthermore, determining what paramedics’ perceive as reasonable/unreasonable regarding overtime may help ambulance services identify and minimise factors that contribute to disengagement, fatigue and dissatisfaction among their employees, enhancing their emergency response capacity. In Australia, Work Health Safety legislation [25] highlights the need for employers to consult with employees on matters that affect their safety, such as RO, to accurately identify, assess and reduce fatigue risks in the workplace. Understanding paramedics’ views on RO presents a foundational step in developing informed, evidence-based recommendations for overtime that align with paramedics’ perceptions of RO and limit the negative impacts working this extra time has on their health and safety. Therefore, the current study aimed to gain consensus on operational paramedics’ perception of conditions that are reasonable and unreasonable for overtime. Methods Setting Within the Australian out-of-hospital workforce, there are just over 25,000 paramedics registered with the national body [26] who work across a range of settings including jurisdictional ambulance services, academia or research, special events, mining/industrial or offshore, primary healthcare, patient transport, and defence or rescue services [12]. In jurisdictional ambulance services, paramedics are typically employed on a full-time basis, working between 38 [6, 17, 18] and 40 [5, 19, 20] hours per week. Rostering arrangements vary across ambulance services, with paramedics commonly working 8-, 10-, 12- or 14-hour shifts depending on industrial agreements set out between the employer and employees. Overtime is remunerated at penalty rates, which are most often set at one and a half times the ordinary rate of pay for the first two hours of overtime, and then double time [6, 18, 19] or double the ordinary rate of pay for any additional hours of overtime [5, 17, 20]. Study design The current study used a Delphi methodology, which is a widely accepted, structured research strategy used in healthcare to develop a consensus opinion using collective intelligence from participants who have experience on a subject [27]. The study had three online survey Rounds via Qualtrics [28], where the completion time of each survey was designed to be less than 10 min and completed in participants own time. Each survey Round consisted of the same six demographic questions, followed by twelve questions in the first Round, six in the second Round and one in the final Round. Each Round was open for ten days and then closed for ten days for analysis purposes. The first survey was opened on April 1st, 2024, and the last survey closed on May 17th, 2024. A reminder email was sent to all participants who had not completed the current round on the fifth and eighth day of the survey round opening. Feedback was provided to participants between rounds via an infographic with current results (See Supplementary Material). Expert selection A strategic and all-inclusive approach was used to recruit paramedics from across Australia via social media (e.g., private Facebook groups for paramedics), snowball sampling and verbal/email invitations to the authors’ personal networks. Personal invitations were used to assist with the geographic stratification of participants across Australia to enhance the study’s generalizability. For this study, experts were defined as paramedics who (i) were registered with the Paramedicine Board of Australia, (ii) were working in a metropolitan setting in Australia, (iii) had ≥ 2 years on-road experience (inclusive of a graduate year) as a paramedic and, (iv) were working in a primarily (≥ 50% of rostered shifts) patient-facing role. Experts were paramedics who had completed the minimum higher-education qualifications required to be registered with the board, which was typically a bachelor-level qualification. The study restricted participation to metropolitan paramedics, given significant differences in overtime and other work demands between metropolitan and regional/rural paramedics [29]. Metropolitan was defined as large urban centres with populations >100,000 [30]. Only participants who partook in the Round 1 survey were invited to participate in subsequent rounds of the study via email. An email address was requested as part of the demographic questions in Round 1 and 2 so that the subsequent survey could be provided to participants who had undertaken the preceding survey. To ensure anonymity, the email address and responses were immediately separated on receipt of the survey in each round. Survey development The Round 1 survey consisted of four parts: (1) demographic details, (2) causes and prevalence of RO, (3) open definition of RO, and (4) determination of reasonable and unreasonable overtime. Demographic details collected in Part 1 included age, gender, location of employment, years of service, role in the organisation and employment type. Part 2 included questions to identify causes and prevalence of RO in participants. Part 3 was an open-ended question asking participants to provide a free-text response outlining their definition of RO. The open-ended question had no prompting, and participants were requested to use their knowledge and experience in the paramedicine setting when responding. Part 4 was a list of 20 different conditions where RO could occur, voted upon using a binary or Likert scale. This was developed from the extremely limited literature available [13, 31] and the lead author’s (MF) personal experience working as a paramedic and completing RO. This list was used to encourage participants to think about the various factors that may influence RO and what could be perceived as reasonable or unreasonable in different conditions. These conditions related to job acuity, length of shift, type of shift, whether entitled meal break(s) had occurred and where paramedics were in their shift rotation. The Round 1 survey was pilot tested with a small group (n = 9) of paramedics who were ineligible for the current Delphi study but had > 20 years of collective experience working as paramedics. Feedback on the readability and flow of the Round 1 survey was requested during pilot testing, and the authorship team assessed face validity. Based on the survey feedback, minor revisions were made to the wording of several RO-related conditions to enhance their clarity prior to Round 1, however no new questions were developed for the survey during pilot testing, and the initial content remained largely the same. During Rounds 1 and 2 of the Delphi, participants proposed additional considerations for RO through the free-text questions, which were then carried forward in the survey and consensus tested in subsequent rounds. This process of developing and using our own survey instrument to assess RO in the current study was selected because there is an absence of validated assessment tools for RO in the field of paramedicine. Although our use of this non-validated survey is a limitation of the study, the Delphi process was chosen to iteratively refine survey items, incorporate expert feedback and establish consensus, thereby strengthening the rigour of the survey findings, despite the lack of previous validation. An overview of the Delphi procedure in the current study is provided in Fig. 1 in the results section. Flow chart of the Delphi procedure and development of reasonable overtime considerations. RO, Reasonable Overtime Data analysis Qualitative data A content, latent analysis in NVivo [32] was performed on the free-text responses in Round 1. Specifically, the free-text responses were analysed and then grouped according to meaning, in line with the four-step analysis approach described by Bengtsson [33]. In step one, the free-text responses were decontextualised into short codes through inductive coding to increase understanding. Secondly, recontextualization occurred by comparing the inductive codes with the original data to ensure no meaning had been lost [33]. Thirdly, categorisation commenced by bringing together like codes. Finally, compilation occurred where conclusions and summations could be drawn, using the words from the text responses themselves [33]. Through this process, RO-related conditions, and durations of RO were determined. Quantitative data All quantitative data was analysed with descriptive statistics. Participants voted on RO conditions on either a binary or Likert scale, depending on when the condition was presented. In Round 1, participants were provided with a 3-point Likert scale (“never reasonable”, “somewhat reasonable” or “always reasonable”) to vote on the provided conditions for RO. In Round 2, participants were provided a 3-point Likert scale (“disagree”, “neutral”, or “agree”) to vote on the open RO definitions. Once a condition was approaching consensus in Round 2 and 3, a binary scale (i.e., “agree” or “disagree”) was presented to participants. Consensus rule In the current study, consensus was defined as ≥ 80% agreement in a response option, which aligns with Jünger and colleagues [34] systematic review that found consensus was most often defined based on the percent agreement in the cohort being between 75 and 80%. If a point of information scored between 70 and 79% agreement, it was deemed to be approaching consensus and re-circulated in the subsequent round where participants then responded via the binary scale. If consensus was not reached after two rounds, the point of information was disregarded. Additionally, if a point of information scored < 70% at any stage, it was disregarded. The study is reported as per the Conducting and REporting DElphi Studies (CREDES) guidelines [34] and received ethics approval (38517) from the Monash University Human Research Ethics Committee. Each participant gave informed consent at the start of each Round. All research was conducted in line with the Helsinki declaration. Results There were 189 participants who completed the Round 1, 144 who completed the Round 2 (response rate of 76.2%) and 128 who completed the Round 3 (response rate of 88.9%) (Table 1). Participants included both paramedics and specialist paramedics from each state and territory in Australia, aside from the Northern Territory. Demographic characteristics remained consistent across each round of the study (Table 1). Table 1Demographic characteristics of participants in rounds 1, 2 and 3Demographic characteristicRound 1Round 2Round 3Number of Participants, n189144128Age, n (%) 18–25 years21 (11.1%)15 (10.4%)9 (7.03%) 26–35 years115 (60.8%)86 (59.7%)79 (61.7%) 36–45 years33 (17.5%)28 (19.4%)25 (19.5%) ≥ 46 years20 (10.6%)15 (10.4%)15 (11.7%)Gender, n (%) Male56 (29.6%)42 (29.2%)42 (32.8%) Female127 (67.2%)96 (66.7%)82 (64.1%) Gender Diverse6 (3.1%)6 (4.2%)4 (3.13%)Primary Location of Work, n (%) Australian Capital Territory3 (1.6%)3 (2.1%)3 (2.34%) New South Wales65 (34.4%)44 (30.6%)35 (27.3%) Northern Territory0 (0%)0 (0%)0 (0%) Queensland84 (44.4%)70 (48.6%)68 (53.1%) South Australia5 (2.6%)2 (1.4%)2 (1.56%) Tasmania1 (0.5%)1 (0.7%)1 (0.7%) Victoria27 (14.3%)21 (14.6%)18 (14.1%) Western Australia4 (2.1%)3 (2.1%)1 (0.7%)Primary Role as a Paramedic, n (%) Paramedic164 (86.8%)126 (87.5%)115 (89.8%) Specialist Paramedic25 (13.2%)18 (12.5%)13 (10.2%)Years of Experience, n (%) 2–5 years74 (39.2%)51 (35.4%)41 (32.0%) 6–10 years68 (36%)55 (38.2%)50 (39.1%) 11–15 years21 (11.1%)21 (14.6%)20 (15.6%) 16–20 years14 (7.4%)9 (6.3%)9 (7.03%) ≥ 21 years12 (6.3%)8 (5.5%)8 (6.2%)Employment Type, n (%) Full Time166 (87.8%)123 (85.4%)110 (85.9%) Part Time19 (10.1%)17 (11.8%)15 (11.7%) Casual4 (2.1%)4 (2.8%)3 (2.3%) Demographic characteristics of participants in rounds 1, 2 and 3 Round 1 Causes and prevalence of RO Participants reported that 23.4% of their shifts finished at their rostered end time in the fortnight immediately prior to the Round 1 survey (see Table 2). The most frequently reported RO length during this period was 60–90 min, but this ranged from as little as 5 min to more than 91 min. Participants were most frequently delayed from finishing their shift on time due to driving back to their station/branch (87.3%) or returning their ambulance to a state of operational readiness (82.0%) (see Table 2). Table 2Characteristics of reasonable overtime in participants in round 1RO characteristicn (%) Shift and RO information in the last two weeks Total number of shifts worked1046 Total number of shifts that finished on time245 (23.4%) Total number of shifts with RO801 (76.6%) Reason for RO in the last two weeks† Driving back to station/branch165 (87.3%) Returning ambulance to state of operational readiness (i.e., cleaning and restocking consumables)155 (82.0%) Completing paperwork133 (70.3%) Requests for service (i.e., calls, jobs, cases)127 (67.2%) Ramping or bed blocked (i.e., delay in offloading patient at hospital)20 (10.6%) Average (self-reported) duration of RO on each shift in the last two weeks 0–30 min4 (2.1%) 31–60 min80 (42.3%) 61–90 min84 (44.4%) ≥ 91 min21 (11.1%) Longest ever RO duration as a paramedic, on top of their rostered shift 0–3 h57 (30.2%) 3.1–6 h116 (61.4%) 6.1–9 h10 (5.3%) 9.1–12 h4 (2.1%) ≥ 12.1 h2 (1.1%)† = percentages do not equal 100% because multiple selections were allowed; RO, Reasonable Overtime Characteristics of reasonable overtime in participants in round 1 † = percentages do not equal 100% because multiple selections were allowed; RO, Reasonable Overtime Definition of RO Participants provided a range of definitions for RO in Round 1 through free-text responses. Based on these definitions, three main themes were developed through the content analysis, where participants’ responses contributed to one or more themes. These three themes included: (1) conditions that are perceived to be reasonable for overtime (n = 97), (2) conditions that are perceived not to be reasonable for overtime (n = 91), and (3) quantifying the duration of overtime that was perceived to be reasonable for overtime (n = 69). The RO-related conditions and durations identified in these three themes are described in detail below, and each were consensus-tested in the subsequent rounds. Conditions perceived to be ROMany of the RO definitions that participants provided pertained to conditions that could be considered RO. Specifically, based on the definitions provided, participants perceived overtime to be reasonable when paramedics are in the following four conditions: (1) dispatched to a time critical emergency where there is an expected (from the perspective of the participant) detrimental effect to the patient without paramedics attending (e.g., cardiac arrest, anaphylaxis, choking); (2) completing the request for service that was currently assigned; (3) dispatched on a time critical request for service in the final 30 min of the shift; (4) completing end of shift duties (i.e., paperwork, and cleaning, restocking and returning vehicles, such as the ambulance, to a state of operational readiness), and returning to station/branch. Conditions perceived to be RO Many of the RO definitions that participants provided pertained to conditions that could be considered RO. Specifically, based on the definitions provided, participants perceived overtime to be reasonable when paramedics are in the following four conditions: (1) dispatched to a time critical emergency where there is an expected (from the perspective of the participant) detrimental effect to the patient without paramedics attending (e.g., cardiac arrest, anaphylaxis, choking); (2) completing the request for service that was currently assigned; (3) dispatched on a time critical request for service in the final 30 min of the shift; (4) completing end of shift duties (i.e., paperwork, and cleaning, restocking and returning vehicles, such as the ambulance, to a state of operational readiness), and returning to station/branch. Conditions not perceived to be ROThe definitions of RO provided by participants also included suggestions of conditions that could be unreasonable for overtime. Based on these responses, participants perceived overtime to be unreasonable in the following four conditions: (1) dispatched to a low acuity request for service close to the rostered shift end time and (from the perspective of the participant) there is unlikely to be an adverse outcome to the patient if left pending (e.g., an ambulance is yet to be dispatched); (2) experiencing operational factors, such as being ramped/bed blocked at a hospital; (3) dispatched on a request for service after rostered shift end time, responding to a request for service and not being provided a ‘hot tag’ (i.e., back up crew) at the rostered shift end time, and arriving on scene after the rostered shift end time, and; (4) implicit or assumed expectation of overtime by supervisor/manager (e.g., by completing end-of-shift tasks after rostered finish time) and inappropriate (from the perspective of participants) upgrading or downgrading of requests for service by supervisor/manager. Conditions not perceived to be RO The definitions of RO provided by participants also included suggestions of conditions that could be unreasonable for overtime. Based on these responses, participants perceived overtime to be unreasonable in the following four conditions: (1) dispatched to a low acuity request for service close to the rostered shift end time and (from the perspective of the participant) there is unlikely to be an adverse outcome to the patient if left pending (e.g., an ambulance is yet to be dispatched); (2) experiencing operational factors, such as being ramped/bed blocked at a hospital; (3) dispatched on a request for service after rostered shift end time, responding to a request for service and not being provided a ‘hot tag’ (i.e., back up crew) at the rostered shift end time, and arriving on scene after the rostered shift end time, and; (4) implicit or assumed expectation of overtime by supervisor/manager (e.g., by completing end-of-shift tasks after rostered finish time) and inappropriate (from the perspective of participants) upgrading or downgrading of requests for service by supervisor/manager. Duration of ROIn their definition of RO, participants provided various durations of what they perceived to be RO, ranging from between five and 120 min. However, the most frequent suggestions of RO were 15-minutes (n = 4), 30-minutes (n = 37) and 60-minutes (n = 18) of overtime. Duration of RO In their definition of RO, participants provided various durations of what they perceived to be RO, ranging from between five and 120 min. However, the most frequent suggestions of RO were 15-minutes (n = 4), 30-minutes (n = 37) and 60-minutes (n = 18) of overtime. Determination of reasonable and unreasonable overtime In addition to providing their own definition of RO, participants in Round 1 responded to the list of predetermined conditions where RO could apply. In Round 1, five conditions met consensus, while six were heading towards consensus and were therefore taken to Round 2 for further testing (see Table 3). Table 3Reasonable and unreasonable overtime-related conditions that were consensus tested in round 1ConditionMet consensus, (%)Approaching consensus, (%) Conditions considered unreasonable for overtime Attend a low priority/acuity case99†-‘Clear the board’ or ‘clear the pending queue’99-On a 14-hr shift88-Entitled meal break(s) has not been taken87-‘Surge in workload’-79Entitled meal break(s) has been taken-79Attend a mid-priority case-77On a night shift-77On the final shift of rotation or week-72 To move back to geographical area - 65 On the 3rd or 4th shift of rotation - 59 On a 12-hour shift - 56 Conditions considered somewhat reasonable for overtime Attend a possibly life-threatening case-72 On a day shift - 57 On a 10-hour shift - 56 On an afternoon shift - 52 On the 2nd shift of rotation - 52 On the 1st shift of rotation - 50 Conditions considered reasonable for overtime Attend a cardiac arrest/imminent death case84†- Life threatening cases - 57 RO, Reasonable Overtime, † = Condition was merged (in the next round) with another condition, italic font denotes a condition that did not meet consensus Reasonable and unreasonable overtime-related conditions that were consensus tested in round 1 RO, Reasonable Overtime, † = Condition was merged (in the next round) with another condition, italic font denotes a condition that did not meet consensus Round 2 In the Round 2 survey, participants completed demographic questions and then voted on conditions from the Round 1 content analysis for the first time. Additionally, in this round, participants were re-presented with, and voted on, the conditions approaching consensus from Round 1. Definition of RO Participants were provided with the overtime conditions from the content analysis and asked to vote on the 3-point Likert scale to determine if they were reasonable or unreasonable. Seven conditions related to unreasonable overtime were presented, and all of these met consensus. Four conditions on RO were presented, with one meeting consensus and one approaching consensus, whereas two did not meet consensus and were thus abandoned from further testing (Table 4). Table 4Reasonable and unreasonable overtime related conditions that were consensus tested in round 2ConditionMet consensus (%)Approaching consensus (%) Conditions considered unreasonable for overtime Expectation of daily overtime by employer97†-Inappropriate up-/down-grading of requests for service (from the paramedics’ perception)97†-Dispatched to a low acuity job close to the shift finish time where (from the paramedics’ perspective) there is probably no adverse outcome to the patient if left pending97†‡-Ramped/bed blocked at the hospital93-Dispatched on a request for service after rostered finish time92-Not being provided a hot tag/ backup crew when on scene after rostered shift finish time88-‘Surge in workload’82-Attend a mid-priority case81-Arriving on the scene after rostered shift finish time81-Over 60 min past the rostered finish time-76On a night shift-74Entitled meal break(s) has been taken72 Final shift of your rotation or week 68 Conditions considered reasonable for overtime Time-critical emergency where (from the paramedics’ perspective) there is an expected detrimental effect to the patient without paramedics’ attendance. For example, cardiac arrest, anaphylaxis or choking.96‡-Attend a possibly life-threatening case95-Over 15 min past the rostered finish time87-Dispatched only in the final 30 min of shift on time-critical requests for service-76 Returning the ambulance to a state of operational readiness - 58 Over 30 min past the rostered finish time - 48 Completing the request for service you are c urrently assigned - 33 RO, Reasonable Overtime, † = Condition was assessed via binary “agree” or “disagree” scale, ‡ = Condition was merged with another condition for the purposes of the set of considerations (see Fig. 2), italic font denotes a condition that did not meet consensus Reasonable and unreasonable overtime related conditions that were consensus tested in round 2 RO, Reasonable Overtime, † = Condition was assessed via binary “agree” or “disagree” scale, ‡ = Condition was merged with another condition for the purposes of the set of considerations (see Fig. 2), italic font denotes a condition that did not meet consensus The most prevalent durations of overtime were also presented to participants for consideration. One met consensus (i.e., 15-minutes), one approached consensus (i.e., 60-minutes), and one did not meet consensus (i.e., 30-minutes) and was abandoned from future consensus testing. Determination of reasonable and unreasonable overtime Participants were re-presented the six RO conditions that were approaching consensus after Round 1 and asked to vote via a binary (i.e., “agree” or “disagree”) scale. Three of these conditions met consensus, the other three did not and were thus abandoned from future consensus testing. Four related conditions were merged into two to aid in the development of clear and succinct RO Considerations. Specifically, ‘Attend a cardiac arrest/imminent death case’ was merged with ‘Time-critical emergency where (from the paramedics’ perspective) there is an expected detrimental effect to the patient without paramedics’ attendance. For example, cardiac arrest, anaphylaxis or choking”. Additionally, ‘Attend a low priority/acuity case’ and ‘dispatched to low acuity request for service close to the shift finish time where (from the paramedics’ perspective) there is probably no adverse outcome to the patient if left pending’ were also merged. Round 3 In the Round 3 survey, participants completed demographic questions and were re-presented the two conditions that were approaching consensus from Round 2. Both of these conditions reached consensus in this final round (see Table 5). Table 5Reasonable and unreasonable overtime related conditions that were consensus tested in round 3ConditionMet consensus (%)Approaching consensus (%) Conditions considered unreasonable for overtime Over 60 min past the rostered finish time90- Conditions considered reasonable for overtime Dispatched only in the final 30 min of shift on time-critical requests for service91- Reasonable and unreasonable overtime related conditions that were consensus tested in round 3 RO considerations Based on the results of this Delphi study, a set of considerations were developed outlining the conditions that paramedics agreed were reasonable and unreasonable for overtime. These considerations are presented in Fig. 2. In Fig. 2, conditions above the solid line are considerations suggested as reasonable for overtime, whereas those below the line are considerations suggested as unreasonable for overtime. Each consideration in Fig. 2 is independent and unique, and therefore they cannot be combined. It is acknowledged there could be some overlapping of these conditions in the real world, however presenting the combined effects of these conditions was beyond the scope of the current study. Percentages in Fig. 2 represent the level of consensus reached by participants for each condition. Considerations that are reasonable and unreasonable for overtime based on paramedics’ perceptions Discussion The present study examined the rate of RO in Australian paramedics and used a Delphi methodology to identify conditions, such as durations of RO, that operational paramedics deem reasonable and unreasonable for overtime. The study found RO occurs on over 75% of shifts and frequently by over 60 min. Consensus was achieved on 50% (17/34) of conditions related to RO, with four conditions being deemed reasonable and thirteen unreasonable. The RO conditions ranged from operational situations, time constraints, acuity of patient presentation and whether factors such as meal breaks were met. Based on the perspectives of participants, we developed a set of considerations for RO that can be used in collaboration with ambulance services to inform future fatigue risk management systems for personnel working in out-of-hospital environments. Beyond fatigue, further investigation of these proposed considerations for RO may also have the potential to minimise the negative impacts of RO on job satisfaction, work-family conflict and health in paramedics [23–25]. Our results indicate paramedics regularly complete RO, which, based on prior research, has the potential to adversely affect their family, mental health, and career longevity [15]. Indeed, participants reported that RO occurred on 76.6% of their shifts in the fortnight preceding Round 1 of the study, and that the duration of RO was commonly >60 min (55.5%). It is documented that one of the most common shift durations completed by paramedics is 12 h [6]. Therefore, based on the amount of RO reported in our sample of paramedics, it is possible paramedics work more than 13 h on duty. While limited research in paramedics has examined the health and safety impact of any RO, let alone specific durations, important insights can be gained from other industries. For instance, a systematic review of research in nurses found those working 12-hour shifts had double the risk of workplace accidents when compared to 8 h [35]. Bae [11] further suggested that shifts >12-hours in duration resulted in poor quality of nursing care and patient safety, increased errors and near errors, poor patient satisfaction and increased patient mortality from pneumonia. Stimpfel et al. [36] also found a clear relationship between longer shift durations in nurses and higher rates of burnout, intention to leave the profession and patient dissatisfaction, and these relationships were strongest in shifts >13 h in duration. In line with this past literature highlighting the adverse impacts of extended shifts, paramedics in the current study agreed that RO beyond a 14-hour shift was unreasonable. While limited research in paramedics has specifically examined periods of work that extend beyond 14 h, in other healthcare professions, shifts of this duration can significantly increase fatigue, reduce responsiveness and vigilant attention and lead to less-satisfactory patient outcomes [11, 35, 36]. Further research should be directed to paramedics working RO beyond 60 min to better understand the impact this overtime has on their health and performance. In the current study, the most prevalent antecedents for RO were returning to the ambulance station/branch and returning the ambulance to a state of operational readiness. Interestingly, however, paramedics could not reach a consensus on whether returning to the station/branch or restoring the ambulance to a state of operational readiness were reasonable or unreasonable conditions for overtime. Naturally, ambulances must get back to their starting location and must be restocked for the next crew of paramedics, and while some paramedics could perceive this as an essential component of the job requiring overtime, others may not, which may explain why consensus was not met. Several ambulance stations/branches in New South Wales, Australia, as well as a large portion of the ambulance services in the UK, use a ‘Make-Ready’ service, whereby non-clinical employees clean and, at times, restock ambulances outside of the paramedics’ shift [37, 38]. This ‘Make-Ready’ service typically occurs in a large ‘superstation’ type setting [37, 38]. Morris et al. [37] found there was a significant improvement in the cleanliness of the ambulances that used the ‘Make-Ready’ approach compared to those who underwent standard cleaning provided by paramedics, who were typically on some form of overtime. While the cleanliness benefits of the ‘Make-Ready’ model are notable, such a model also has the potential to reduce RO, which could subsequently enhance the health and wellbeing of employees [31]. However, there is no research quantifying the effect of the ‘Make-Ready’ model on RO. Thus, ambulance services should further investigate the ‘Make-Ready’ model as one solution to reduce RO. Participants considered RO related to actual (i.e., cardiac arrest, choking and anaphylaxis) and potential life-threatening requests for service were reasonable, with both meeting a high level of consensus of 96% and 95%, respectively. The role of paramedics is evolving, however, traditionally they were known for their rapid response, critical interventions and transportation of patients experiencing medical emergencies [39, 40]. While ‘saving lives’ and paramedics have been thought of as synonymous (p.1680) [40], since the COVID-19 pandemic, recent research suggests the role of pre-hospital clinicians is changing to become part of the broader healthcare system, including non-emergency healthcare [40–42]. Therefore, the willingness of paramedics to extend their shifts for cases that use their full scope of practice could be related to their intrinsic motivation to save lives [43], compassion satisfaction [44], the ability to use their technical high-fidelity-low-frequency skills [45, 46] and/or their altruism [47]. Participants identified 13 conditions where they perceived RO to be unreasonable. It was established that ramping or bed blockages (i.e., delay in offloading a patient at hospital) was considered unreasonable for overtime by 93% of participants in Round 1. Literature suggests ramping can expose paramedics to verbal and physical abuse and contribute to poor mental health, including depression, anxiety, stress and post-traumatic stress disorder [48]. These impacts of ramping could explain paramedics’ aversion to this aspect of the job, particularly when working RO. Additionally, feelings of powerlessness and frustration can overcome paramedics whilst at a standstill, which could challenge their professional intrinsic motivators [43, 48]. Further, participants reached an 87% consensus that RO without a meal break was unreasonable. It is well-established that paramedics are already a high-risk profession for poor nutritional habits based on the mobility of the workforce and opportunistic eating [49, 50]. Additionally, without a meal break, physical fatigue, sleep disturbances and emotional exhaustion can increase, which may affect performance [51]. Although these adverse impacts are known for some of the RO-related conditions identified in the current study, such as meal breaks and ramping, many have not been investigated, and thus require further investigation. While overtime is associated with increased burnout and fatigue, it can also provide benefits for paramedics [14–16], including additional remuneration and increased flexibility in managing work-life balance [7]. Importantly, this study does not propose the removal of overtime all together, but rather the development of consensus-based parameters around its reasonable use. Moreover, although 12-hour shifts are a recognised contributor to fatigue [2], nurses working 12-hour shifts have reported less fatigue and emotional exhaustion when compared to working 8-hour shifts [52]. While there is limited research comparing the impact of different shift lengths on fatigue and wellbeing in Australian paramedics, it is possible that, like nurses, paramedics experience similar work-life balance benefits of consolidated work hours that provide extended periods off duty. This is important when considering the applicability of the RO considerations identified in the current study, especially those related to durations of overtime, and highlights the need to examine the impact of these RO considerations on paramedics’ health and wellbeing. Building on the findings of the current study, an important next step will involve convening a multidisciplinary team comprising industrial relations experts, fatigue specialists, policymakers, risk analysts and ambulance service executives, who will translate the identified RO considerations into actionable recommendations for practice and policy in paramedicine. This collaborative approach will be critical to ensure any proposed changes to RO are evidence-based and operationally feasible and are underpinned with a mutual focus on clinical excellence and patient safety. Subsequent investigations could also use these findings to inform research in regional, rural and remote settings where paramedics experience similarly fatiguing work environments [53], but in vastly different operational settings. Limitations We used social media as the main recruitment method in the present study, which is a method that has several benefits but also some limitations. For instance, people who engage with social media are typically of a younger age [54], which may have limited the representativeness of our sample. However, the age distribution in our sample was generally consistent with national data, although paramedics aged 26–35 years were over-represented (our sample = 60.8% vs. national data = 40.9%) and those aged >47 years were under-represented (our sample = 10.6% vs. national data = 27.5%) [26]. Additionally, not all paramedics may have access to social media, which may have introduced a potential sampling bias. We attempted to combat this limitation by also recruiting paramedics via snowballing and personal contacts. Although these recruitment methods are also open to forms of bias, the use of personal contacts helped increase the overall geographical representativeness of the sample. Indeed, the geographical location of our participants suggested the smaller states were represented well, whereas Queensland (our sample = 44.4% vs. national data = 26.0%) and New South Wales (our sample = 34.4% vs. national data = 25.7%) were over-represented and Victoria (our sample = 14.3% vs. national data = 28.5%) was under-represented against national data [26]. However, as data was not collected on how participants became aware of the study, the effectiveness of the different recruitment methods is unknown. Furthermore, the findings may be influenced by self-selection bias, whereby individuals with particularly strong views or substantial experience with RO may have been more likely to participate. A further limitation of the study was that the initial list of 20 RO conditions in Round 1 was primarily constructed from the first author’s experience of when RO could be applied. Prior to presenting this list to participants, it was reviewed by the wider interdisciplinary authorship team. This approach helped ensure a holistic view of RO was captured in the list of RO conditions and limited the possibility of implicit bias from the first author influencing the list presented to participants. Despite these efforts to limit bias, it is possible the initial list of RO-related conditions in Round 1 could have swayed participants’ views in future rounds, thus these findings should be interpreted with caution and confirmed with further research. Finally, it is possible recall bias and social-desirability bias could be present in the current study. These biases could have impacted participants’ self-reported durations of RO. Further research which uses verifiable measures of RO, such as company records from ambulance services, is therefore needed. Conclusion This study gained consensus on operational paramedics’ perceptions of conditions that are reasonable and unreasonable for overtime. Based on our findings, paramedics agreed it was reasonable to complete RO when responding to potentially life-threatening cases, attending time critical cases where there is a high likelihood of patient deterioration without paramedics’ attendance, dispatched in the final 30-mins of shift on time-critical cases, and when overtime is no more than 15 min. Limited research has systematically consulted with paramedics to understand their views and experiences on matters that can impact fatigue, such as overtime. Therefore, the current study provides an important step in establishing paramedics’ perceptions of RO, laying the groundwork for future research that can use these insights in collaboration with ambulance services and other key stakeholders to develop, test and evaluate overtime requirements that limit the adverse impact of RO on paramedics’ health and safety. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 2 Supplementary Material 3 Supplementary Material 4 Acknowledgements The authors are grateful to all participants, without whom this study could not have been possible. Abbreviations Coronavirus disease 2019 Conducing and Reporting Delphi Studies Reasonable Overtime United Kingdom Author contributions MF – Conceptualization, Methodology, Software, Validation, Formal Analysis, Investigation, Data Curation, Writing – Original Draft, Writing – Review & Editing, Visualization, Project Administration. KB - Conceptualization, Methodology, Software, Validation, Writing – Review & Editing, Supervision. AL – Conceptualization, Methodology, Writing – Review & Editing, Supervision. AW – Conceptualization, Methodology, Data Curation, Writing – Original Draft Writing – Review & Editing, Visualisation, Supervision. Funding The author(s) received no financial support for the research, authorship, and/or publication. Data availability The dataset generated during the current study is available from the corresponding author on reasonable request. Declarations Ethics approval and consent to participate The study received ethics approval (38517) from the Monash University Human Research Ethics Committee. All participants provided informed consent to participate through the first online survey. All research was conducted in line with the Helsinki declaration. Consent for publication Not applicable. Disclosures The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of the article. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Material 1 Supplementary Material 2 Supplementary Material 3 Supplementary Material 4 Data Availability Statement The dataset generated during the current study is available from the corresponding author on reasonable request.