Recently, I completed nearly six years working with New Zealand District Health Boards, on behalf of the Ministry of Health, nominally to help them pursue the Shorter Stays in the Emergency Department Health Target.1 While this target is about achieving a certain Emergency Department (ED) length of stay for a high proportion of ED patients, much of the work needed was facilitating patient flow through our hospitals. I was privileged to observe, frequently work with and occasionally help, the people of our hospitals as they did this work. Consequently I managed to accumulate knowledge of what worked, in what context, and what didnt. This paper is a summary of some of this knowledge about acute patient flow in our hospitals. It attempts to construct an ideal model for improving patient flow and maximising the efficient use of the capacities of our hospitals. In reality, this ideal model doesnt exist in any of our hospitals exactly in the form I propose, but the elements of it do. I am of the view that our hospitals, and our patients, would greatly benefit if these elements were brought together and made to work well.Acute demand mitigation and alternatives to hospital careIt is important to emphasise efforts to decrease demand for acute hospital care and to provide alternative care options in the community. Our hospitals cannot continue to accommodate current rates of increase in acute demand and most would agree that care close to home is in the patients, and the systems, best interests. However, this paper is about flow of acute patients in the hospital. Acute demand mitigation and the provision of alternatives to hospital care are essential activities, but they are for discussion elsewhere.The problem our hospitals are struggling with demandDemand for hospital care is increasing. Regularly, and particularly during winter, our hospitals become overcrowded and dysfunctional as demand exceeds capacity. Consequently, acute patients overflow to less appropriate hospital wards acute medical patients are admitted to surgical wards, filling these and resulting in cancelling elective surgery. Ward rounds are prolonged as clinical teams visit patients throughout the hospital ( safari ward rounds ), decision making is delayed, patients access the next phase of care later, and hospital length of stay is prolonged. Prolonged length of stay further reduces access for new acute patients to hospital beds, making the demand and capacity mismatch worse. Exacerbating this are systems which might not facilitate early definitive decision making, timely access of acute patients to diagnostics, (eg, CT scanning), timely access to other necessary interventions such as acute surgery, nor efficient discharge of the patient when hospital care is no longer needed. Because acute patients continue to present to the ED, but access to care beyond the ED has become increasingly overwhelmed, a significant although not the only manifestation of this demand and capacity mismatch is worsening overcrowding in the ED (patients keep coming in but they cant get out). ED overcrowding is associated with a number of well-documented adverse consequences, including greater discomfort and indignity for patients, worse clinical outcomes and increased mortality.2-7Backlog and queues beget more backlog and queues a downward spiral of hospital dysfunctionPatients wait in queues for ED cubicles, doctors, CT scans, hospital beds, and so on. The formation of queues represents a backlog of work work parked in a queue, needing to be done in addition to the ongoing clinical demand of those continuing to come in. Now the capacity (cubicles, doctors, CT scans, hospital beds and so on) is attempting to process the ongoing incoming demand (which it was struggling to deal with before, hence the development of the backlog), in addition to the accumulated backlog of work. Consequently, the backlog grows. Indeed, the growth in the backlog accelerates like the compounding interest of an investment in the bank, as the capacity can deal with a smaller and smaller proportion of the total work asked of it. Furthermore, the formation of queues exacerbates delays over and above the sum of all the work accumulated in the backlog. For example, if five patients wait for beds and five beds become immediately available, the 2nd, 3rd, 4th and 5th patients will have incremental additions to their waits as the patients ahead of them are processed (like a car stationary in a queue at the lights, even though the lights have turned green it does not get through the intersection because each cars departure is incrementally delayed while waiting for the car in front to depart). The mismatch of demand and capacity creates a self perpetuating cycle of worsening mismatch with more overcrowding and a larger and larger backlog of demand. Furthermore, the consequences of this are delays to treatment, longer hospital lengths of stay, increased adverse events, safari ward rounds, and so on all of which further consume capacity and all of this combines to produce a downward spiral of overcrowding and hospital dysfunction. Consequently a bad day, sending the system down this spiral, might take several days of extra capacity (eg, special measures such as forcing discharges, cancelling planned admissions, opening extra beds, etc.) to reduce the overcrowding; a bad weekend might take a week or more and a prolonged influx of winter illness might take months to return to a state where the backlog can be cleared and the overcrowding resolved.It is a very bad thingMeanwhile, the overcrowded and dysfunctional hospital has undoubtedly harmed patients, contributed to the deaths of some, and exacerbated the stress and frustration of its staff. Furthermore, the inefficiencies associated result in reduced productivity and a costly investment is required to accommodate both the inefficiencies and to get back on top of demand. Failure to match demand and capacity prospectively is bad for patients, staff, and it is very bad for the bottom line. Running hospitals like this is burning money.There is a delusion that when there is an appropriate demand and capacity match things are quiet and even inefficientWhen things return to a state where demand and capacity are better matched, hospital staff having battled with the demands of an overcrowded system often perceive things are quiet and as a consequence, somehow less productive. With no patients in corridors, with beds available when patients are ready to go to them, with all the teams patients on the one ward, when the day allows time for lunch and a cup of coffee, there is a misperception that the human and physical resources are not being optimally utilised. However, the opposite is true. As patients get the care they need more promptly, they have better outcomes and shorter hospital stays. Clinicians, sufficiently rested, fed and watered, and with less time pressures, attend to patients more quickly and more comprehensively, and make better decisions. Wasted time is removed from throughout the patient journey, there is less re-work as things are done better the first time, and there is less demand created by the complications associated with rushed, truncated and delayed care. Having capacity which accommodates demand, flexes when there are predictable spikes in demand (eg, winter), and can recognise early and respond to unexpected spikes in demand, is good for patients, staff and the bottom line. For all of these reasons, not the least of which is the best use of our health dollars, we must keep above the spiral keep the capacity above demand more on this shortly.Capacity is not a bed count and demand is not a head countA patient is not a unit of demand. Each patient will have their own needs but similar patients might be grouped and counted to identify the accumulated demand on a given resource resuscitation bay, CT scanning, operating theatre, duty neurologist, and so on. Capacity is not a hospital bed count. Capacity, in acute care, is many things the capacity of the triage nurse to triage the growing queue of patients, the cubicle capacity of the ED, the decision making capacity among the medical staff (some of whom might offer limited decision making capacity and others who often see the patient relatively late might offer a lot), CT scanning capacity, nursing capacity on the ward, and so on. If a patient journey was value stream mapped, then deficiencies in capacity (bottlenecks in the journey) might be identified. The greatest deficiency in capacity is the tightest bottleneck the most significant capacity constraint for that patient journey. Once the demand for that phase of care is exceeded then queues, backlogs, etc, manifest. However, once that bottleneck is addressed, by better matching demand and capacity for that phase, then a new bottleneck will emerge as the greatest capacity constraint for that journey. Each new bottleneck will be a lesser constraint and as each is addressed the journey will become increasingly efficient.How much capacity?We are able to forecast demand, even acute demand, accurately. Use of predictive forecasting is well established in a number of our health systems. However, all of them should understand likely future demand, based on sophisticated prediction methodologies, and should be prepared for it. Based on predicted demand a desired capacity to address that demand can be considered. If capacity at any stage in a patient journey is sufficient to cater for the average demand, then it will be overwhelmed half the time, sending the system down the spiral of dysfunction. If capacity is set to cater for more than the highest forecast peaks of demand (greater than the 100th percentile of demand fluctuation) then the system should not go down the spiral, but more than half the time there will be unused capacity. We cannot afford this level of capacity. So, the average (50%) is too low, and 100% is insufficiently frugal. The correct capacity must be somewhere in between. There has been much discussion about 85% with this level of occupancy of hospital beds often cited as being most efficient for patient flow.8,9 Of course, what this means is that a bed is available for a patient when the patient needs it, so that queues for beds dont develop. But it is not the 85% occupancy that matters, it is the avoidance of queues. Starting a hospital day with 15% of beds free will not necessarily prevent queues if the 15% spare capacity is soon used up and not replenished. Similarly, starting a hospital day with no free beds might not result in queues for beds if the hospital has the ability to free up beds as soon as they are needed. Suffice it to say, the closer your capacity is to exactly matching your demand the more responsive your system needs to be so that capacity is rapidly mobilised to prevent queues developing. In reality, our hospitals cannot maintain more than 15% free capacity of beds, or other capacities, at all times. Consequently, we must have responsive systems which know what is happening throughout the system, recognise when demand is threatening capacity and can mobilise capacity promptly so that we stay as close to the top of the spiral as we can. In essence, the key is to avoid queues, or at least to limit their length as best we can. The methodology below uses this as a foundation.Figure 1: A comprehensive approach to improving patient flow in our hospitals the Left to right, over and under model.A comprehensive approach to improving patient flow in our hospitals the left to right, over and under modelThis model has three foundation principles:\r\n The patient journey, as experienced by the patient, is the vantage point. All staff in the system are citizens of a system and are working together to maximise the utility of the whole system and not just parts of it. Avoiding queues, or at least limiting their size, is a priority\r\nAnd three conceptual components (Figure 1):\r\n Governance The patient journey Operations\r\nThe conceptual components of the model are represented in Figure 1 and are explained in more detail below.GovernanceThe over part of the model is the overarching governance of what is done and how it is done. The components of Governance include:1. StructureThe structure is the gathering of people who both lead and manage the model. The membership is seen by staff as both representative (among the group are people who represent me) and authoritative (among the group are important people who can get things done). Already this structure has provided a clear statement for staff that they have a unified and well-led system, and that they are part of it.A significant fault in our hospitals is that a great many of our staff work in their circles with little understanding of, or commitment to, the whole system. This isnt their fault. It is usually unclear to them how they fit into the scheme of things, nor what the scheme of things is. A structure such as this can provide that clarity. Furthermore, it is clear that doctors make important decisions about the elements of patient flow when to admit, investigate, operate, discharge and so on but they are seldom aware of the bigger hospital picture and very rarely contribute to bigger picture operations. To quote one senior nurse; nurses run hospitals and doctors mess them up. A structure such as this can engage doctors in the bigger operational picture.Included in this structure are good clinical governance, operational leadership, street wisdom and project management grunt.This group would work to principles such as those above, or augmented with their own, perhaps including principles relating to cooperation, respect, and quality. They would oversee the operations component of the model and would (by virtue of their influence and authority) disseminate good practice and standardise practice, as appropriate, across the system.In the Patient Journey section to follow, a number of expected practices are alluded to and many are listed in Table 1. Examples include referral and handover practises from the ED to inpatient services, use of journey boards and daily rapid rounds and a number of others. Many of these have become a standard of care, but have been applied only patchily in our hospitals. Governance structures such as this can facilitate both the expectation and the implementation of these and other initiatives.2. MethodThe method employed by the Governance Group would include commonly employed project methodology such as those used in Lean Thinking. Of most importance is that the group is well informed with good data and good diagnostics. Ultimately, the methodology is to define actions for change and improvement, and should be comprehensive, prioritised, pragmatic and action orientated.10-123. PlanOut of the methodology will form a plan the actions to be done. In our hospitals there are a great many things to be done. A mistake manifest in many of them is to embark on actions without applying a good methodology within an appropriate structure (the steps above). Not everything can be done at once. If good actions are taken, but they are not the most important actions just now, and particularly if they are not driven with clear outcomes, timelines and responsibilities, then they will often bear no fruit. The analogy of a dog chasing seagulls on the beach is apt. She sees a seagull and sets off after it. Out of the corner of her eye she sees another and sets off after that then another, then another and so on. Although the dog is very busy, and exhausted at the end of the day, she catches no seagulls. Our process improvement efforts in many of our hospitals have been like a dog chasing seagulls. The double whammy of this is that the expected improvement is not achieved but, in addition, those enthused to partake (particularly clinicians) become disillusioned and walk away from future improvement efforts.The actions undertaken must be the priority actions, because they address the biggest bottlenecks in the patient journey, they are the most important from a quality or safety perspective, they offer the greatest improvement for the effort instilled, or some combination of these. The prioritisation comes out of the methodology above with the superimposition of the structure of the Governance Group. Furthermore, the Governance Group enables clear outcomes, timelines and responsibilities to be associated with each action and rigidly manages these so that the actions bring results.OperationsThe under part of the model is Operations, which underpins good patient flow. One of the most important developments in our hospitals in recent years has been more sophisticated operations, sometimes based on precedents in other industries (eg, airlines) and often including an Operations Centre and operations meetings during the day. However, despite this improved sophistication there is much still to be done. Some of our hospitals still are run (particularly after hours) by senior nurses roaming the wards with a pencil and a clipboard. Even those with Operations Centres often are using them as little more than meeting rooms with electronic clipboards. Even when the information displayed and discussed demonstrates or predicts a hospital going down the spiral of dysfunction, often there is little able to be done in response. To leave the bed meeting on Friday after noting the hospital has a capacity and demand mismatch, is forming long queues and significant backlogs of work, with no further meetings until Monday, and with little more than a plan to encourage early discharges is an indictment of how poorly we manage our hospitals.Furthermore, despite the fact doctors often are responsible for key decisions which allow patient movement to the next part of their journey, it is very unusual to see one in an Operations Centre or at a daily bed/operations meeting. I doubt there are other industries which allow such crucial players to be so distant from operations.The components of Operations might include:1. Forecasting demand in the future days, weeks and months out.\r\nPredicting demand, days, weeks, months and even years in advance is important, so that predicted variations in demand (eg: Sundays in an ED; the winter months in a medical ward; or a major event in a city), are matched by pre-emptive adjustments to capacity. Our health systems should understand likely future demand, based on sophisticated prediction methodologies, and should be prepared for it.2. Knowing demand now\r\nUnderstanding demand right now is essential for the functioning of a responsive system which can shift capacity and demand effectively. Good, comprehensive, real-time data are needed to be able to respond to fluctuations in demand. It should be reiterated that demand is not a head count and capacity is not a bed count. A patient is not a unit of work and modern systems should be able to apply an appropriate weighting to the capacity required to cater for a patients needs, (the weighting in this context is sometimes referred to as acuity). Many hospitals are doing this is some form, assisted by available tools (acuity tools) and displaying this information to provide an overview of hospital demand and capacity at a glance. Operations Centres (or Control Centres) established in some hospitals provide the command and control focus charged with scrutinising and responding to such information. Our hospitals should have authoritative control of capacity informed by comprehensive, real-time, understanding of demand.3. Matching capacity and demand\r\nThe two components above are about understanding capacity and demand in the future (forecasting), and now (daily operations). Responding to that understanding is essential. In general terms three categories of response might be considered: i. Medium to long term responses: eg, winter planning. ii. Short-term: daily operationsiii. Crisis responses: also known asgridlock or overload plans. If the first two categories have been done well then crisis responses should become less and less necessary. As discussed earlier, the crisis is when the descent down the spiral of dysfunction has begun, with queues formed and backlogs of work accumulated. With this comes overcrowding, outliers, safari ward rounds, delayed care, longer length of stay, and so on. As much as we can, we should avoid getting into this state.How do we match capacity and demand?Unless we are happy to reduce the quality of care delivered to patients then, in simple terms, the options to address capacity and demand mismatch are to:\r\n Reduce demand Smooth demand and capacity curves so they are better matched more later Free-up capacity by improving processes Purchase more capacity\r\nIn reality, the answer is all of the above. While we must do all we can to reduce demand, it is ambitious to think we will be able to reverse it. Much work has been done to free-up capacity, but there is plenty of scope to do more. However, inevitably greater capacity will need to be purchased to cope with increasing demand. What we must not do is wallow unduly in options 1, 2 and 3 in the mistaken belief that we can avoid option 4 altogether. Options 1, 2 and 3 take time and, if we allow demand to exceed capacity regularly such that we descend down the spiral of dysfunction, it will ultimately be costly for patients, staff and the bottom line.Smoothing demand and capacityEvery day, demand for acute services follows a reasonably predictable curve, starting to climb out of its early morning nadir from about 10 or 11am until mid-afternoon, remaining high until late evening and then falling to its late night/early morning nadir again. However, the capacity curve lags behind. The freeing-up of inpatient beds for example, as patients are discharged tends to occur from mid- to late-afternoon. As the demand for beds begins to peak at 11am and the availability is not for some hours later, there is a demand/capacity mismatch for these hours. Many observers are dismissive of this relatively brief mismatch and reassure that beds will be coming up soon. However, during these hours queues develop and a backlog builds. Overcrowding ensues meaning queues for other parts of the journey (ED cubicles, nurses, etc) also develop, creating backlogs of work for these capacities too. While beds do ultimately become available, clearing the queue for beds, then clearing the queues that developed for other capacities and the accumulated backlog of work (while continuing to deal with the prolonged afternoon/evening peak in demand), takes a long time and great effort. Meanwhile care is delayed, length of stay in the ED is prolonged, and so on, further reducing the capacity to deal with ongoing demand and the accumulated backlog. Inefficiencies beget inefficiencies, and a relatively brief period of capacity and demand mismatch creates significant problems for a significant period of time. Ultimately, over the whole day, there is enough capacity and there isnt too much demand, but the demand and capacity curves are mismatched. If this is happening at your hospital then shifting the capacity curve to earlier in the day is an important intervention.Similarly, every week Monday brings significant demand for hospital beds (usually the highest number of acute admissions). However, the weekend days have seen the lowest numbers of discharges of the week. Consequently, we superimpose a high-demand day on a low-capacity day. Furthermore, some hospitals still bring elective admissions in on a Sunday and many have high elective day of admission rates on a Monday. These capacity/demand mismatches are self-inflicted. Consequently, queues and backlogs occur early in the week and the hospital remains over capacity well into the week. If these are occurring at your hospital, then increasing weekend discharges and smoothing elective in-flow so that it occurs when acute demand is least during the week, are important interventions.Finally, it is not uncommon for hospitals to be doing large elective volumes in the winter and to do least in the summer. The reasons for this might include elective performance deadlines or to accommodate staff holiday patterns. Of course, acute demand is maximal in winter. Superimposing high elective demand on high acute demand, with a fixed hospital capacity, is inviting descent down the spiral. If this is happening at your hospital then changing seasonal patterns of elective activity to occur when acute activity is least is an important intervention.Daily, weekly and seasonal demand and capacity mismatches are sometimes due to modifiable, self-imposed patterns of work. There is opportunity to smooth the demand and capacity curves so that mismatches are minimised.The Patient JourneyCentral to the model is the patient journey. Governing the sorts of things that happen on the patient journey, and the expectations of it, is the Governance Group. Operations ensure that these things happen as best they can. The patient journey, in this model, is central because it represents a foundation principle of the model, and because it is what governance and operations are all about.There are a number of processes in the patient journey which the Governance Group should encourage and which Operations should facilitate. Figure 2 is a diagrammatic representation of the patient journey with each of the arrows representing a process which should be applied, as appropriate, to every acute patient journey through the hospital. Many of these are now considered a standard of care. To not apply them, therefore, is a failure to provide an acceptable standard of care. Listed in Table 1 is a selection of these processes.Figure 2: The patient journey with indications of a number of processes expected to apply to each journeyTable 1: A selection of expected elements of an acute patient journey\r\n \r\n \r\n \r\n 1. Efficient referral processes for primary care to ED and acute specialities\r\n \r\n Easy and not obstructive. Standardised as appropriate. To consultant if possible. With alternatives to immediate transfer (eg, acute clinics). With alternative access to hospital if ED care not required (eg, direct to MAPU).\r\n \r\n \r\n \r\n 2. Streaming from triage\r\n \r\n Triage nurse movement of patient to appropriate part of ED or other service.\r\n \r\n \r\n \r\n 3. Nurse initiated treatments\r\n and referral\r\n \r\n For appropriate cases.\r\n \r\n \r\n \r\n 4. Front loaded decision making in ED\r\n \r\n ED Consultant involved early to facilitate decision making.\r\n \r\n \r\n \r\n 5. Explicit attention to patient flow in the ED\r\n \r\n Includes explicit responsibility for patient flow either in the form of a dedicated ED flow nurse or making this an explicit function of the senior nurse and doctor in charge of the shift (a non-clinical person facilitating flow is not the preferred model as it tends to prioritise time in ED as a driver for movement rather than clinical utility) Measurement and display of metrics regarding flow (length of stay and occupancy, for example) to facilitate understanding of clinical progress of patients.\r\n \r\n \r\n \r\n 6. Clinical agreements and clinical pathways\r\n \r\n Standardised care for appropriate patient groups. Includes agreed specialist team for acceptance. Includes agreed transfer to ward criteria (like the fractured neck of femur pathways, but applicable to many other patient groups). Includes review by inpatient registrar in the ED only when required for patient benefit and not as a routine. Includes encouragement of acceptance of referrals from ED based on negotiation, no refusal, and the concept of best fit rather than definitive diagnosis. Includes what and when imaging is done. Includes agreements regarding known patients and recent discharges.\r\n \r\n \r\n \r\n 7. Maximised responsiveness of inpatient teams\r\n \r\n Medical representatives of inpatient teams (usually registrars) who are rostered to attend ED for acute care of patients under their specialitys responsibility should be available or provide an alternative. This means being free from other duties and having a backup plan (another registrar or consultant) when unable to attend promptly (for example, when doing acute surgery).\r\n \r\n \r\n \r\n 8. Efficient nurse handovers to wards\r\n \r\n Standardised among all services, as appropriate. Ideally electronic.\r\n \r\n \r\n \r\n 9. Use of medical assessment and planning units (or similar)\r\n \r\n For the inpatient team workup of acute medical patients. Includes transfer to the MAPU as soon as appropriate. Surgical units (SAPUs) or similar units for other specialities, are not considered a standard of care but might be desirable.\r\n \r\n \r\n \r\n 10. Use of expected date of discharge, journey boards and daily rapid rounds\r\n \r\n All patients admitted acutely (with few exceptions) should have an expected date of discharge displayed, communicated to the patient and family, and logged on a patient journey board or similar. From the day of admission patients discharge planning should begin, with opportunity for all members of the multidisciplinary team to see, understand and log their involvement with the patients on the journey board. Daily rapid/board rounds including representatives of all disciplines involved in the patients care sharing a mutual update of the patients

Recently, I completed nearly six years working with New Zealand District Health Boards, on behalf of the Ministry of Health, nominally to help them pursue the Shorter Stays in the Emergency Department Health Target.1 While this target is about achieving a certain Emergency Department (ED) length of stay for a high proportion of ED patients, much of the work needed was facilitating patient flow through our hospitals. I was privileged to observe, frequently work with and occasionally help, the people of our hospitals as they did this work. Consequently I managed to accumulate knowledge of what worked, in what context, and what didnt. This paper is a summary of some of this knowledge about acute patient flow in our hospitals. It attempts to construct an ideal model for improving patient flow and maximising the efficient use of the capacities of our hospitals. In reality, this ideal model doesnt exist in any of our hospitals exactly in the form I propose, but the elements of it do. I am of the view that our hospitals, and our patients, would greatly benefit if these elements were brought together and made to work well.Acute demand mitigation and alternatives to hospital careIt is important to emphasise efforts to decrease demand for acute hospital care and to provide alternative care options in the community. Our hospitals cannot continue to accommodate current rates of increase in acute demand and most would agree that care close to home is in the patients, and the systems, best interests. However, this paper is about flow of acute patients in the hospital. Acute demand mitigation and the provision of alternatives to hospital care are essential activities, but they are for discussion elsewhere.The problem our hospitals are struggling with demandDemand for hospital care is increasing. Regularly, and particularly during winter, our hospitals become overcrowded and dysfunctional as demand exceeds capacity. Consequently, acute patients overflow to less appropriate hospital wards acute medical patients are admitted to surgical wards, filling these and resulting in cancelling elective surgery. Ward rounds are prolonged as clinical teams visit patients throughout the hospital ( safari ward rounds ), decision making is delayed, patients access the next phase of care later, and hospital length of stay is prolonged. Prolonged length of stay further reduces access for new acute patients to hospital beds, making the demand and capacity mismatch worse. Exacerbating this are systems which might not facilitate early definitive decision making, timely access of acute patients to diagnostics, (eg, CT scanning), timely access to other necessary interventions such as acute surgery, nor efficient discharge of the patient when hospital care is no longer needed. Because acute patients continue to present to the ED, but access to care beyond the ED has become increasingly overwhelmed, a significant although not the only manifestation of this demand and capacity mismatch is worsening overcrowding in the ED (patients keep coming in but they cant get out). ED overcrowding is associated with a number of well-documented adverse consequences, including greater discomfort and indignity for patients, worse clinical outcomes and increased mortality.2-7Backlog and queues beget more backlog and queues a downward spiral of hospital dysfunctionPatients wait in queues for ED cubicles, doctors, CT scans, hospital beds, and so on. The formation of queues represents a backlog of work work parked in a queue, needing to be done in addition to the ongoing clinical demand of those continuing to come in. Now the capacity (cubicles, doctors, CT scans, hospital beds and so on) is attempting to process the ongoing incoming demand (which it was struggling to deal with before, hence the development of the backlog), in addition to the accumulated backlog of work. Consequently, the backlog grows. Indeed, the growth in the backlog accelerates like the compounding interest of an investment in the bank, as the capacity can deal with a smaller and smaller proportion of the total work asked of it. Furthermore, the formation of queues exacerbates delays over and above the sum of all the work accumulated in the backlog. For example, if five patients wait for beds and five beds become immediately available, the 2nd, 3rd, 4th and 5th patients will have incremental additions to their waits as the patients ahead of them are processed (like a car stationary in a queue at the lights, even though the lights have turned green it does not get through the intersection because each cars departure is incrementally delayed while waiting for the car in front to depart). The mismatch of demand and capacity creates a self perpetuating cycle of worsening mismatch with more overcrowding and a larger and larger backlog of demand. Furthermore, the consequences of this are delays to treatment, longer hospital lengths of stay, increased adverse events, safari ward rounds, and so on all of which further consume capacity and all of this combines to produce a downward spiral of overcrowding and hospital dysfunction. Consequently a bad day, sending the system down this spiral, might take several days of extra capacity (eg, special measures such as forcing discharges, cancelling planned admissions, opening extra beds, etc.) to reduce the overcrowding; a bad weekend might take a week or more and a prolonged influx of winter illness might take months to return to a state where the backlog can be cleared and the overcrowding resolved.It is a very bad thingMeanwhile, the overcrowded and dysfunctional hospital has undoubtedly harmed patients, contributed to the deaths of some, and exacerbated the stress and frustration of its staff. Furthermore, the inefficiencies associated result in reduced productivity and a costly investment is required to accommodate both the inefficiencies and to get back on top of demand. Failure to match demand and capacity prospectively is bad for patients, staff, and it is very bad for the bottom line. Running hospitals like this is burning money.There is a delusion that when there is an appropriate demand and capacity match things are quiet and even inefficientWhen things return to a state where demand and capacity are better matched, hospital staff having battled with the demands of an overcrowded system often perceive things are quiet and as a consequence, somehow less productive. With no patients in corridors, with beds available when patients are ready to go to them, with all the teams patients on the one ward, when the day allows time for lunch and a cup of coffee, there is a misperception that the human and physical resources are not being optimally utilised. However, the opposite is true. As patients get the care they need more promptly, they have better outcomes and shorter hospital stays. Clinicians, sufficiently rested, fed and watered, and with less time pressures, attend to patients more quickly and more comprehensively, and make better decisions. Wasted time is removed from throughout the patient journey, there is less re-work as things are done better the first time, and there is less demand created by the complications associated with rushed, truncated and delayed care. Having capacity which accommodates demand, flexes when there are predictable spikes in demand (eg, winter), and can recognise early and respond to unexpected spikes in demand, is good for patients, staff and the bottom line. For all of these reasons, not the least of which is the best use of our health dollars, we must keep above the spiral keep the capacity above demand more on this shortly.Capacity is not a bed count and demand is not a head countA patient is not a unit of demand. Each patient will have their own needs but similar patients might be grouped and counted to identify the accumulated demand on a given resource resuscitation bay, CT scanning, operating theatre, duty neurologist, and so on. Capacity is not a hospital bed count. Capacity, in acute care, is many things the capacity of the triage nurse to triage the growing queue of patients, the cubicle capacity of the ED, the decision making capacity among the medical staff (some of whom might offer limited decision making capacity and others who often see the patient relatively late might offer a lot), CT scanning capacity, nursing capacity on the ward, and so on. If a patient journey was value stream mapped, then deficiencies in capacity (bottlenecks in the journey) might be identified. The greatest deficiency in capacity is the tightest bottleneck the most significant capacity constraint for that patient journey. Once the demand for that phase of care is exceeded then queues, backlogs, etc, manifest. However, once that bottleneck is addressed, by better matching demand and capacity for that phase, then a new bottleneck will emerge as the greatest capacity constraint for that journey. Each new bottleneck will be a lesser constraint and as each is addressed the journey will become increasingly efficient.How much capacity?We are able to forecast demand, even acute demand, accurately. Use of predictive forecasting is well established in a number of our health systems. However, all of them should understand likely future demand, based on sophisticated prediction methodologies, and should be prepared for it. Based on predicted demand a desired capacity to address that demand can be considered. If capacity at any stage in a patient journey is sufficient to cater for the average demand, then it will be overwhelmed half the time, sending the system down the spiral of dysfunction. If capacity is set to cater for more than the highest forecast peaks of demand (greater than the 100th percentile of demand fluctuation) then the system should not go down the spiral, but more than half the time there will be unused capacity. We cannot afford this level of capacity. So, the average (50%) is too low, and 100% is insufficiently frugal. The correct capacity must be somewhere in between. There has been much discussion about 85% with this level of occupancy of hospital beds often cited as being most efficient for patient flow.8,9 Of course, what this means is that a bed is available for a patient when the patient needs it, so that queues for beds dont develop. But it is not the 85% occupancy that matters, it is the avoidance of queues. Starting a hospital day with 15% of beds free will not necessarily prevent queues if the 15% spare capacity is soon used up and not replenished. Similarly, starting a hospital day with no free beds might not result in queues for beds if the hospital has the ability to free up beds as soon as they are needed. Suffice it to say, the closer your capacity is to exactly matching your demand the more responsive your system needs to be so that capacity is rapidly mobilised to prevent queues developing. In reality, our hospitals cannot maintain more than 15% free capacity of beds, or other capacities, at all times. Consequently, we must have responsive systems which know what is happening throughout the system, recognise when demand is threatening capacity and can mobilise capacity promptly so that we stay as close to the top of the spiral as we can. In essence, the key is to avoid queues, or at least to limit their length as best we can. The methodology below uses this as a foundation.Figure 1: A comprehensive approach to improving patient flow in our hospitals the Left to right, over and under model.A comprehensive approach to improving patient flow in our hospitals the left to right, over and under modelThis model has three foundation principles:\r\n The patient journey, as experienced by the patient, is the vantage point. All staff in the system are citizens of a system and are working together to maximise the utility of the whole system and not just parts of it. Avoiding queues, or at least limiting their size, is a priority\r\nAnd three conceptual components (Figure 1):\r\n Governance The patient journey Operations\r\nThe conceptual components of the model are represented in Figure 1 and are explained in more detail below.GovernanceThe over part of the model is the overarching governance of what is done and how it is done. The components of Governance include:1. StructureThe structure is the gathering of people who both lead and manage the model. The membership is seen by staff as both representative (among the group are people who represent me) and authoritative (among the group are important people who can get things done). Already this structure has provided a clear statement for staff that they have a unified and well-led system, and that they are part of it.A significant fault in our hospitals is that a great many of our staff work in their circles with little understanding of, or commitment to, the whole system. This isnt their fault. It is usually unclear to them how they fit into the scheme of things, nor what the scheme of things is. A structure such as this can provide that clarity. Furthermore, it is clear that doctors make important decisions about the elements of patient flow when to admit, investigate, operate, discharge and so on but they are seldom aware of the bigger hospital picture and very rarely contribute to bigger picture operations. To quote one senior nurse; nurses run hospitals and doctors mess them up. A structure such as this can engage doctors in the bigger operational picture.Included in this structure are good clinical governance, operational leadership, street wisdom and project management grunt.This group would work to principles such as those above, or augmented with their own, perhaps including principles relating to cooperation, respect, and quality. They would oversee the operations component of the model and would (by virtue of their influence and authority) disseminate good practice and standardise practice, as appropriate, across the system.In the Patient Journey section to follow, a number of expected practices are alluded to and many are listed in Table 1. Examples include referral and handover practises from the ED to inpatient services, use of journey boards and daily rapid rounds and a number of others. Many of these have become a standard of care, but have been applied only patchily in our hospitals. Governance structures such as this can facilitate both the expectation and the implementation of these and other initiatives.2. MethodThe method employed by the Governance Group would include commonly employed project methodology such as those used in Lean Thinking. Of most importance is that the group is well informed with good data and good diagnostics. Ultimately, the methodology is to define actions for change and improvement, and should be comprehensive, prioritised, pragmatic and action orientated.10-123. PlanOut of the methodology will form a plan the actions to be done. In our hospitals there are a great many things to be done. A mistake manifest in many of them is to embark on actions without applying a good methodology within an appropriate structure (the steps above). Not everything can be done at once. If good actions are taken, but they are not the most important actions just now, and particularly if they are not driven with clear outcomes, timelines and responsibilities, then they will often bear no fruit. The analogy of a dog chasing seagulls on the beach is apt. She sees a seagull and sets off after it. Out of the corner of her eye she sees another and sets off after that then another, then another and so on. Although the dog is very busy, and exhausted at the end of the day, she catches no seagulls. Our process improvement efforts in many of our hospitals have been like a dog chasing seagulls. The double whammy of this is that the expected improvement is not achieved but, in addition, those enthused to partake (particularly clinicians) become disillusioned and walk away from future improvement efforts.The actions undertaken must be the priority actions, because they address the biggest bottlenecks in the patient journey, they are the most important from a quality or safety perspective, they offer the greatest improvement for the effort instilled, or some combination of these. The prioritisation comes out of the methodology above with the superimposition of the structure of the Governance Group. Furthermore, the Governance Group enables clear outcomes, timelines and responsibilities to be associated with each action and rigidly manages these so that the actions bring results.OperationsThe under part of the model is Operations, which underpins good patient flow. One of the most important developments in our hospitals in recent years has been more sophisticated operations, sometimes based on precedents in other industries (eg, airlines) and often including an Operations Centre and operations meetings during the day. However, despite this improved sophistication there is much still to be done. Some of our hospitals still are run (particularly after hours) by senior nurses roaming the wards with a pencil and a clipboard. Even those with Operations Centres often are using them as little more than meeting rooms with electronic clipboards. Even when the information displayed and discussed demonstrates or predicts a hospital going down the spiral of dysfunction, often there is little able to be done in response. To leave the bed meeting on Friday after noting the hospital has a capacity and demand mismatch, is forming long queues and significant backlogs of work, with no further meetings until Monday, and with little more than a plan to encourage early discharges is an indictment of how poorly we manage our hospitals.Furthermore, despite the fact doctors often are responsible for key decisions which allow patient movement to the next part of their journey, it is very unusual to see one in an Operations Centre or at a daily bed/operations meeting. I doubt there are other industries which allow such crucial players to be so distant from operations.The components of Operations might include:1. Forecasting demand in the future days, weeks and months out.\r\nPredicting demand, days, weeks, months and even years in advance is important, so that predicted variations in demand (eg: Sundays in an ED; the winter months in a medical ward; or a major event in a city), are matched by pre-emptive adjustments to capacity. Our health systems should understand likely future demand, based on sophisticated prediction methodologies, and should be prepared for it.2. Knowing demand now\r\nUnderstanding demand right now is essential for the functioning of a responsive system which can shift capacity and demand effectively. Good, comprehensive, real-time data are needed to be able to respond to fluctuations in demand. It should be reiterated that demand is not a head count and capacity is not a bed count. A patient is not a unit of work and modern systems should be able to apply an appropriate weighting to the capacity required to cater for a patients needs, (the weighting in this context is sometimes referred to as acuity). Many hospitals are doing this is some form, assisted by available tools (acuity tools) and displaying this information to provide an overview of hospital demand and capacity at a glance. Operations Centres (or Control Centres) established in some hospitals provide the command and control focus charged with scrutinising and responding to such information. Our hospitals should have authoritative control of capacity informed by comprehensive, real-time, understanding of demand.3. Matching capacity and demand\r\nThe two components above are about understanding capacity and demand in the future (forecasting), and now (daily operations). Responding to that understanding is essential. In general terms three categories of response might be considered: i. Medium to long term responses: eg, winter planning. ii. Short-term: daily operationsiii. Crisis responses: also known asgridlock or overload plans. If the first two categories have been done well then crisis responses should become less and less necessary. As discussed earlier, the crisis is when the descent down the spiral of dysfunction has begun, with queues formed and backlogs of work accumulated. With this comes overcrowding, outliers, safari ward rounds, delayed care, longer length of stay, and so on. As much as we can, we should avoid getting into this state.How do we match capacity and demand?Unless we are happy to reduce the quality of care delivered to patients then, in simple terms, the options to address capacity and demand mismatch are to:\r\n Reduce demand Smooth demand and capacity curves so they are better matched more later Free-up capacity by improving processes Purchase more capacity\r\nIn reality, the answer is all of the above. While we must do all we can to reduce demand, it is ambitious to think we will be able to reverse it. Much work has been done to free-up capacity, but there is plenty of scope to do more. However, inevitably greater capacity will need to be purchased to cope with increasing demand. What we must not do is wallow unduly in options 1, 2 and 3 in the mistaken belief that we can avoid option 4 altogether. Options 1, 2 and 3 take time and, if we allow demand to exceed capacity regularly such that we descend down the spiral of dysfunction, it will ultimately be costly for patients, staff and the bottom line.Smoothing demand and capacityEvery day, demand for acute services follows a reasonably predictable curve, starting to climb out of its early morning nadir from about 10 or 11am until mid-afternoon, remaining high until late evening and then falling to its late night/early morning nadir again. However, the capacity curve lags behind. The freeing-up of inpatient beds for example, as patients are discharged tends to occur from mid- to late-afternoon. As the demand for beds begins to peak at 11am and the availability is not for some hours later, there is a demand/capacity mismatch for these hours. Many observers are dismissive of this relatively brief mismatch and reassure that beds will be coming up soon. However, during these hours queues develop and a backlog builds. Overcrowding ensues meaning queues for other parts of the journey (ED cubicles, nurses, etc) also develop, creating backlogs of work for these capacities too. While beds do ultimately become available, clearing the queue for beds, then clearing the queues that developed for other capacities and the accumulated backlog of work (while continuing to deal with the prolonged afternoon/evening peak in demand), takes a long time and great effort. Meanwhile care is delayed, length of stay in the ED is prolonged, and so on, further reducing the capacity to deal with ongoing demand and the accumulated backlog. Inefficiencies beget inefficiencies, and a relatively brief period of capacity and demand mismatch creates significant problems for a significant period of time. Ultimately, over the whole day, there is enough capacity and there isnt too much demand, but the demand and capacity curves are mismatched. If this is happening at your hospital then shifting the capacity curve to earlier in the day is an important intervention.Similarly, every week Monday brings significant demand for hospital beds (usually the highest number of acute admissions). However, the weekend days have seen the lowest numbers of discharges of the week. Consequently, we superimpose a high-demand day on a low-capacity day. Furthermore, some hospitals still bring elective admissions in on a Sunday and many have high elective day of admission rates on a Monday. These capacity/demand mismatches are self-inflicted. Consequently, queues and backlogs occur early in the week and the hospital remains over capacity well into the week. If these are occurring at your hospital, then increasing weekend discharges and smoothing elective in-flow so that it occurs when acute demand is least during the week, are important interventions.Finally, it is not uncommon for hospitals to be doing large elective volumes in the winter and to do least in the summer. The reasons for this might include elective performance deadlines or to accommodate staff holiday patterns. Of course, acute demand is maximal in winter. Superimposing high elective demand on high acute demand, with a fixed hospital capacity, is inviting descent down the spiral. If this is happening at your hospital then changing seasonal patterns of elective activity to occur when acute activity is least is an important intervention.Daily, weekly and seasonal demand and capacity mismatches are sometimes due to modifiable, self-imposed patterns of work. There is opportunity to smooth the demand and capacity curves so that mismatches are minimised.The Patient JourneyCentral to the model is the patient journey. Governing the sorts of things that happen on the patient journey, and the expectations of it, is the Governance Group. Operations ensure that these things happen as best they can. The patient journey, in this model, is central because it represents a foundation principle of the model, and because it is what governance and operations are all about.There are a number of processes in the patient journey which the Governance Group should encourage and which Operations should facilitate. Figure 2 is a diagrammatic representation of the patient journey with each of the arrows representing a process which should be applied, as appropriate, to every acute patient journey through the hospital. Many of these are now considered a standard of care. To not apply them, therefore, is a failure to provide an acceptable standard of care. Listed in Table 1 is a selection of these processes.Figure 2: The patient journey with indications of a number of processes expected to apply to each journeyTable 1: A selection of expected elements of an acute patient journey\r\n \r\n \r\n \r\n 1. Efficient referral processes for primary care to ED and acute specialities\r\n \r\n Easy and not obstructive. Standardised as appropriate. To consultant if possible. With alternatives to immediate transfer (eg, acute clinics). With alternative access to hospital if ED care not required (eg, direct to MAPU).\r\n \r\n \r\n \r\n 2. Streaming from triage\r\n \r\n Triage nurse movement of patient to appropriate part of ED or other service.\r\n \r\n \r\n \r\n 3. Nurse initiated treatments\r\n and referral\r\n \r\n For appropriate cases.\r\n \r\n \r\n \r\n 4. Front loaded decision making in ED\r\n \r\n ED Consultant involved early to facilitate decision making.\r\n \r\n \r\n \r\n 5. Explicit attention to patient flow in the ED\r\n \r\n Includes explicit responsibility for patient flow either in the form of a dedicated ED flow nurse or making this an explicit function of the senior nurse and doctor in charge of the shift (a non-clinical person facilitating flow is not the preferred model as it tends to prioritise time in ED as a driver for movement rather than clinical utility) Measurement and display of metrics regarding flow (length of stay and occupancy, for example) to facilitate understanding of clinical progress of patients.\r\n \r\n \r\n \r\n 6. Clinical agreements and clinical pathways\r\n \r\n Standardised care for appropriate patient groups. Includes agreed specialist team for acceptance. Includes agreed transfer to ward criteria (like the fractured neck of femur pathways, but applicable to many other patient groups). Includes review by inpatient registrar in the ED only when required for patient benefit and not as a routine. Includes encouragement of acceptance of referrals from ED based on negotiation, no refusal, and the concept of best fit rather than definitive diagnosis. Includes what and when imaging is done. Includes agreements regarding known patients and recent discharges.\r\n \r\n \r\n \r\n 7. Maximised responsiveness of inpatient teams\r\n \r\n Medical representatives of inpatient teams (usually registrars) who are rostered to attend ED for acute care of patients under their specialitys responsibility should be available or provide an alternative. This means being free from other duties and having a backup plan (another registrar or consultant) when unable to attend promptly (for example, when doing acute surgery).\r\n \r\n \r\n \r\n 8. Efficient nurse handovers to wards\r\n \r\n Standardised among all services, as appropriate. Ideally electronic.\r\n \r\n \r\n \r\n 9. Use of medical assessment and planning units (or similar)\r\n \r\n For the inpatient team workup of acute medical patients. Includes transfer to the MAPU as soon as appropriate. Surgical units (SAPUs) or similar units for other specialities, are not considered a standard of care but might be desirable.\r\n \r\n \r\n \r\n 10. Use of expected date of discharge, journey boards and daily rapid rounds\r\n \r\n All patients admitted acutely (with few exceptions) should have an expected date of discharge displayed, communicated to the patient and family, and logged on a patient journey board or similar. From the day of admission patients discharge planning should begin, with opportunity for all members of the multidisciplinary team to see, understand and log their involvement with the patients on the journey board. Daily rapid/board rounds including representatives of all disciplines involved in the patients care sharing a mutual update of the patients