While all of this may seem to be a good idea from a theoretical perspective, implementing it in practice seems challenging. We believe that this may be because we are used to treating patients with acute conditions (eg, an acutely ill patient presenting to the emergency department). Although some acute conditions have clear timelines for treatment (ie, the length of time that they must be managed in the emergency department), others are less well defined, making it difficult to establish a precise timeframe for the appropriate implementation of a specific intervention. Moreover, if the intervention does not provide a clear improvement in management in terms of patient outcomes, how do we know whether this was caused by the introduction of a new process? Our proposal uses an observational design based on a quality improvement strategy similar to the stepped wedge design, but is less rigorous than a traditional stepped-wedge cluster randomised trial \[[@CR8]\]. However, we consider that it could be an ideal approach for introducing new processes in complex areas such as the introduction of a new clinical pathway in an acute care setting, and this approach can easily be followed by a randomised trial. Moreover, our proposal, if implemented, could provide a valuable learning opportunity for health care managers and clinicians, giving them experience of successfully using a new intervention in the presence of a real-life patient problem. Limitations {#Sec2} ----------- A major challenge of using a stepped wedge cluster randomised trial design is the potential imbalance in the time elapsed between different steps. This could cause bias if some steps take more time to implement than others, and in our experience, the process of preparing educational materials for our process intervention might take more time to develop than the educational programme itself. However, we think it is important to develop educational materials, because this could be a key tool for facilitating the implementation of new processes. For instance, if the educational materials were successfully developed and disseminated (which is likely to happen in step 2), we do not anticipate that the programme should take more time to implement than the education itself. Second, in many complex intervention studies, it may not be feasible to identify a well-defined time window within which the intervention is expected to provide improvements in patient care. We are not aware of any previous studies that have used a stepped wedge design to study complex interventions, and therefore the feasibility of the approach in these circumstances is unknown. We are currently conducting a study to compare outcomes associated with the intervention with outcomes associated with treatment-as-usual in our emergency department \[[@CR4]\]. If this comparison is positive, we will not use the stepped wedge design. If the stepped wedge design should be used, the results of this comparison will likely help us decide when to implement the intervention. Finally, there is always a risk that an interrupted process in a stepped wedge trial may become a 'Hawthorne effect'. However, this is mitigated by ensuring that the introduction of the intervention and observation of the outcomes are separated in time and not carried out as a sequence. Implementation in practice {#Sec3} ========================== In this section, we describe the implementation of the intervention in our study. Our intervention is based on a checklist to help doctors order the appropriate clinical tests. The checklist was implemented in a previous study \[[@CR9]\]. The improvement team decided to use the checklist again in our study, given the lack of clear improvements in management observed in the previous study. The checklist is a two-page document for which doctors have to sign to indicate that they have read and understand the contents (Table [1](#Tab1){ref-type="table"}). The contents include the type of test required to determine the diagnosis, the expected result of the test, and a brief outline of the management of each disease or injury, using a traffic light colour scheme (red indicates that a change in management is required, yellow indicates that the test results may be suggestive of the disease, and green indicates that no test result is needed). If a patient presents with a disease or injury for which a test is indicated, the doctor should check the box to indicate whether the test has been ordered or not. If a patient presents with a disease or injury for which no test is indicated, the doctor should check the box to indicate the diagnosis and order the test if indicated. Finally, if the doctor believes that more tests are required, he or she should fill in a red box, to indicate that they feel that the patient's management needs to be changed.Table 1Stepped wedge quality improvement checklistCheck listUse in the emergency departmentThe list is divided into sections corresponding to different clinical situations (acute heart failure, acute pulmonary embolism, acute stroke, minor head injury, etc).The patient has acute heart failure and diuretic therapy is indicated.The patient has acute pulmonary embolism and an urgent D-dimer test is indicated.The patient has acute stroke and an urgent brain CT scan is indicated.A patient with an uncomplicated minor head injury without unconsciousness or impaired consciousness is admitted to the emergency department.The doctor asks the patient if they would like to be tested for HIV.The doctor asks the patient if they would like to be tested for HIV.The doctor writes a yellow note to indicate that they have performed the test for HIV, but has not done any other tests.The doctor writes a yellow note to indicate that they have performed the test for HIV, but has not done any other tests.The doctor writes a yellow note to indicate that they have performed the test for HIV and HIV antibodies, but has not done any other tests.The doctor writes a red note to indicate that the patient needs more tests to establish a diagnosis, but no further tests are needed at that time.The doctor writes a red note to indicate that the patient needs more tests to establish a diagnosis, but no further tests are needed at that time.The doctor writes a red note to indicate that the patient does not need any more tests at that time.The doctor writes a red note to indicate that the patient does not need any more tests at that time.The doctor writes a red note to indicate that the patient does not need any more tests at that time.The patient is admitted to the general medical ward.The doctor writes a red note to indicate that the patient does not need any more tests at that time.The patient is admitted to the surgical ward.The doctor writes a yellow note to indicate that they have performed the test for HIV, but has not done any other tests.The doctor writes a yellow note to indicate that they have performed the test for HIV, but has not done any other tests.The doctor writes a yellow note to indicate that they have performed the test for HIV and HIV antibodies, but has not done any other tests.The doctor writes a yellow note to indicate that they have performed the test for HIV and HIV antibodies, but has not done any other tests.The doctor writes a yellow note to indicate that they have performed the test for HIV and HIV antibodies, but has not done any other tests.The doctor writes a red note to indicate that the patient needs more tests to establish a diagnosis, but no further tests are needed at that time.The doctor writes a red note to indicate that the patient needs more tests to establish a diagnosis, but no further tests are needed at that time.The doctor writes a red note to indicate that the patient does not need any more tests at that time.The doctor writes a red note to indicate that the patient does not need any more tests at that time.The doctor writes a red note to indicate that the patient does not need any more tests at that time.The doctor writes a red note to indicate that the patient does not need any more tests at that time.The patient is admitted to the cardiology ward.The doctor writes a red note to indicate that the patient does not need any more tests at that time.The patient is admitted to the paediatric ward.The doctor writes a yellow note to indicate that they have performed the test for HIV, but has not done any other tests.The doctor writes a yellow note to indicate that they have performed the test for HIV, but has not done any other tests.The doctor writes a yellow note to indicate that they have performed the test for HIV and HIV antibodies, but has not done any other tests.The doctor writes a yellow note to indicate that they have performed the test for HIV and HIV antibodies, but has not done any other tests.The doctor writes a yellow note to indicate that they have performed the test for HIV and HIV antibodies, but has not done any other tests.The doctor writes a red note to indicate that the patient needs more tests to establish a diagnosis, but no further tests are needed at that time.The doctor writes a red note to indicate that the patient needs more tests to establish a diagnosis, but no further tests are needed at that time.The doctor writes a red note to indicate that the patient does not need any more tests at that time.The doctor writes a red note to indicate that the patient does not need any more tests at that time.The patient is transferred to the intensive care unit.The doctor writes a red note to indicate that the patient does not need any more tests at that time.The doctor writes a red note to indicate that the patient does not need any more tests at that time.The patient is transferred to the renal unit.The doctor writes a red note to indicate that the patient does not need any more tests at that time.The patient is transferred to the ophthalmology clinic.The doctor writes a yellow note to indicate that they have performed the test for HIV, but has not done any other tests.The doctor writes a yellow note to indicate that they have performed the test for HIV, but has not done any other tests.The doctor writes a yellow note to indicate that they have performed the test for HIV and HIV antibodies, but has not done any other tests.The