Paediatric Sepsis Series — Data and Metrics

Transcript

With Steve calling you a subordinate.

Hi guys. Just a quick reminder. Can you please put yourselves on mute? Thank you.

Welcome everybody. We'll just give people a few more moments to join.

We've already got a good number on the call, which is just great to see. Sarah is here with us today. Sarah was here with us. When we did this talk two weeks ago, but was also juggling a little one and so it's great to have her helping us out. formally today you can see Mel who's with me as well.

We'll just give we'll just give people a few more moments to to join us. Mel just reassure me. Are you just looking at my slide or you seeing my slide and my team screen? I can see your precious slide just Okay. Well, I think I we'll make a start then.

Welcome everybody. This is the second time that Mel and I have given this talk we this is the second of the Pediatric sepsis seminar series that we're running as part of the Queensland pediatric sepsis program.

The series was started by Paula Lister who along with me co-leads the Queensland Pediatrics Sepsis Program and Paula's seminar, which you can still watch through teams online. I'm sure introduced the recently established National clinical Care standard for sepsis and this seminar that Mel and I are doing is focused on the challenges of delivering the sepsis Care standard and more than that on how we measure and report our performance against the sepsis care standard in particular how we can make use of of the data that we have access to and how we can make that as efficient. And is complete as possible.

And my name is Adam Irwin, and I'm a senior lecturing pediatric infectious diseases here at Queensland children's hospital and the University of Queensland and I'm joined today by Melissa Pilbeam who is a critical care pharmacist and the antimicrobial stewardship pharmacist on on the program.

Once again, I just like to say it's great to see so many people joining us and we had a similar number join us a couple of weeks ago.

It's really pleasing to see the representation of the state on the on this seminar and what we what we did last week two weeks ago was we asked people to to put in the chat where they were joining us from just to see how broadly we are really reaching out to to the states.

I think the furthest we achieved two weeks ago was Norfolk Island. So if anybody can beat that then you'll you'll have done very well, I think I'm just going to plug now as well that we we really value feedback from all of you and We we have we like a bit of friendly competition in the group and Paula managed to get much more feedback than I did on my first talk.

So I'm just going to remind you that within the chat Sarah's gonna put the link to a very short evaluation survey that you can do at the end of the at the end of the session and we'd really value your feedback on what what was useful and and what was missing what what did we address well, and what didn't we address .

On that note, I'll make a start and I just like to start by acknowledging the traditional custodians of the land on which we're all meeting. Which for me here in Brisbane in southeast Queensland at the Yuggera and Turrbal people's and I'd like to pay my respects to their ancestors and their descendants who continue cultural and spiritual connections to Country and I just like to acknowledge and recognize their valuable contributions both to Australian and to Global Society.

So this is what we're going to cover in the the course of this session today will introduce the the burden of sepsis is a clinical problem. And as a result of that I'll introduce for those of you who are not familiar with us the Queensland pediatric sepsis program. I'm going to talk about the national clinical Care standard for sepsis and and what that means for us and introduce the quality statements and the indicators against which we're measured when for the clinical Care standard  and given that we need to work out how we're going to report against the standard and what and which data we're going to measure how we're going to access those data and then more broadly I'll talk about the opportunities that the clinical data that we do have access to present to us for improving the clinical management of children with sepsis.

So we know that sepsis is an enormous clinical problem globally and the latest Global burden of disease survey estimated that there are a roughly 20 million cases of sepsis in children under five each year and almost three million deaths in children under five each year and sepsis is not only a problem of low and middle income countries here in Australia.

There will be approximately 50 children who will die with sepsis each year and though that doesn't sound a very large number in the global context that's more deaths and a caused by road traffic accidents or leukemia in children in Australia each year.

8% of all intensive care units 8% of it of all pediatric incentive Care Unit admissions are associated with sepsis and more than a quarter of all death on the Pediatric ICU in Australian New Zealand is attributable to sepsis. So it's a significant problem for us in Australia, too.

Here in Queensland the latest pediatric patient safety review highlighted. Once again the problem of a failure to recognize sepsis in our serious Adverse Events reporting of of the 12 sack one events. Those are the most serious Adverse Events that occur in hospitals and Health Services in Queensland and eight of those 12 sack one events were due to a failure to recognize and initiate prompts treatments of sepsis.

So in response to this significant clinical problem, the Queensland pediatric sepsis program was established and we received ongoing funding from Queensland Health last year in 2021. And so established our five year. Roadmap and our roadmap included these goals which include not not only a desire to improve clinical outcomes from sepsis such as a reduction in hospital and ICU length of stay and an improvement in standardized practice and reduce in a reduction in the variation. in care across Queensland But it was also about improving Clinicians and families knowledge of sepsis and so we have an important focus on implementing and educating around a clinical pathway and improving training for sepsis. And we also have our educational series such as this pediatric seminar series to improve awareness for healthcare providers and indeed for consumers. the pathway of care that we have implemented and And that we train clinicians in was implemented in the emergency departments firstly in 2019 and the Ed sepsis pathway of care. I was implemented in 16 emergency departments and more than 70. Regional and remote sites around the state I'm having implemented a sepsis pathway. Another priority for us is to align with State national and international bodies. And in addition to that to support sites to report on their performance along with these National national standards such as the new sepsis clinical Care standard. And Paul a Lister introduced the sepsis Care standard in much more detail if you weeks ago, but I'll just highlight that the goal of the sepsis Care standard was to ensure is to ensure that patients presenting with signs and symptoms of sepsis. I recognized early and received coordinated best practice care. So that the risk of death or ongoing morbidity is reduced. And like other clinical care standards in Australia, the sepsis clinical care times is made up of a number of quality statements and the quality statements address the quality of care along the patient Journey from early recognition to post this charge planning and care. This sepsis clinical Care standard is not a sepsis guideline, but it's a framework Within. Within which HHS is an institutions can establish clinical care standards and guidelines against which they can be measured. and let me give you an example of the way in which at these quality statements are framed and against which Our performance is measured so the quality statement to refers to the importance of time critical management. And so institutions would be tasked with reporting against. The fact that they have evidence of a clinical pathway in place that would support early and appropriate clinical management and that they have evidence of local arrangements to support access to Diagnostics and medicine's that they have access to clinicians with sepsis experience 24 hours a day that they have training based. They have competency-based training on the sepsis pathway in place and they have an audit of performance. so for example implementing the Queensland pediatric sepsis pathway working with the program and implementing the Pediatric sepsis pathway in in an institution would count as evidence towards reporting against this quality statement. However, there are other aspects of the quality statements which are much more challenging to report against such as the proportion of patients with sepsis that are treated according to the pathway and that brings us to the the question of how we how we measure these data and how we report back against the standard and that will be the focus of much of the rest of this talk. So let me tell you a little bit about what we did with the implementation of the Pediatric accepts this pathway. For those of you who are not familiar with the sepsis pathway and the sepsis pathway includes a screening and recognition tool a treatment bundle and an acute bundle of care. And those are the key interventions that you need to deliver in the emergency setting in order to improve outcomes from sepsis. And it includes antimicrobial guidelines antimicrobial recommendations and dosing along with a patient information leaflet all of which are available through through wink. And what we did was we implemented the pathway with our partners around the state was we undertook a really substantial piece of work to try to capture capture data that allowed us to measure our performance in the impact of this sepsis pathway implementation. And so what you're looking at here in this table is in let me directly to the Yellow Boxes. The first of the yellow box is on the right hand side. The proportion of children with septic shock and who were these are children with septic shock evaluated on the pathway during the course of the implementation of which there were 146 of these and the proportion of those children that received the key three elements of the treatment bundle within 60 Minutes of sepsis recognition, which would be the recommendation and somewhat disappointingly. We found that only 32% of these really sick children received those three elements within 60 minutes. The context however, I just direct you to the fact that there's a really large perspective. Statewide implementation of a sepsis pathway in New York state which found that they achieved only a 26% and success rate in delivering the three element bundle within 60 minutes. So though we were somewhat disappointed to see this this rate of delivery of the interventions. It was comparable with other international data. Another process measure against which we were Keen to measure our performance as we delivered the the sepsis pathway was the proportion of children that received antibiotics. One of the one of those key interventions in the bundle and in septic shock the recommendation is that children receive antibiotics within one hour of recognition. And again, we found that of those 146 children with septic shock only 47% of children received antibiotics within 60 minutes. For those who had sepsis with organ dysfunction, but no shock 90% of those received their antibiotics within three hours. So we were disappointed that there was limited adherence with the pathway and limited success in delivering those time critical interventions. But what we were able to do is compare our performance and with the pathway implementation in 2018 and 2019 with a pre-pathway population taken from 2015 to 2018. And in the sickest of the children evaluated with the pathway in 2018 and 29 who ended up on picu 92% of those received and the three element bundle within three hours 62% received at the three element bundle within 60 minutes. Which was an increase from the pre-pathway implementation in 2015 to 2018 where 59% received the three element bundle within three hours 32% within within an hour. So the key message of this evaluation was that delivery of the three element bundle increased in the sickest patients who were admitted to ICU with septic shock? We also found that and our appropriateness of antibiotic prescribing improved with the introduction of the pathway and the antimicroheal recommendations. But we did find that at qca where we captured data from the electronic medical record and the electronic prescribing system. We did found find that there was an association with the instruction of the pathway in April 2019. That's the red line on this graph and a subsequent increase in the use of third generation kevlosporins kef track zone or kef taxim in the emergency department. So following the introduction of the pathway the number of days on therapy per thousand attendances increase from 13 to 17 and that was associated with the introduction of the pathway. So the summary of that evaluation was that the pathway was associated with improved process measures over time particularly in the sickest patients. It was improved with it. It was associated with improved empirical antibiotic use and dosing. But it seemed to be Associated to within increase in overall antibiotic consumption. And more to the point. This was an enormous piece of work. This was a really laborious process out that was dependent on engagement with all of the emergency department sites with local Champions with emergency department nurses Pharmacists and researchers in order to capture these data perspectively, so it was a really significant piece of work. So as an example that may not be a sustainable approach to data capture for the purpose of reporting for the standard. So what we need to ask is if we're going to do this sustainably, where are we going to get our data from? How are we going to collect it and clean it and organize it? How accurate are these data? and also what we measuring what are our definitions who are the patient groups that were interested in? How do we decide what? how do we decide what the definitions of the the times are against which we measure and delivery of interventions. For example. And so with that I'm going to hand over to Mel who's going to talk us through some practical solutions to this to these problems. Thanks, Adam. Everyone, I'm Melissa in as Anna mentioned. I'm at AMS pharmacist with the Queensland pediatric acceptance program. And in my other role. I'm a pediatric Critical Care pharmacist for qch and so just to kick us off in the subject of data collection. I wanted to show you a quick quiz of a real patient admitted to the age just a few months ago. We had an eight-year-old female presenting to Ed with increasing seizures fever and lethargy. Background of spastic wood CP with an existing seizure disorder is known to be previously colonized with immersei and SPL. She had a seizures managed a blood culture taken commenced on broad spectrum empirical Ivy antibiotics and admitted to the wood but we did 24 hours had an acute deterioration received two fluid boluses and was transferred to pick us. She was incubated commenced on inotropes some hydrocortisone. And how do Ivy antibiotic therapy broadened our initial blood culture came back negative to ask despite the negative blood culture in this case classifies very clearly as septic shock organ dysfunction resulting from an infection. However, she was not coded for sepsis or septic shock for this admission. And therefore would not be identified as a sepsis case in any data analysis. How could this be? Well at no point during her admission was the word sepsis documented anywhere in her notes, like in many cases of sepsis an initial blood culture is negative. So a source cannot necessarily be confirmed. However, any other cultures are taken post the commencement of any microbials and we know that sepsis is not just caused by a bacteremia. Although this is still sometimes a common misconception. Despite coaches being negative if the patient clinically improves with sepsis management and an empirical antimicrobial regime. We may Define this as culture negative septic shock which responded to sepsis management. Well, what was documented for the code is to see in this case seizures fever aspiration food refractory hypertension human dermic instability and aloc or acute loss of consciousness what code is could determine out of these notes was aspiration in the setting of seizures next slide please Adam. So how do we perform in sepsis incidence mortality and management? And how is this trading over time to assess this? We need to be able to collect and collect these important data metrics not just the snapshots by as real-time Trends. However, the most important one for all of us to get right is the total number of cases per site across Queensland and nationally which isn't currently done comprehensively if we can do this accurately, then the rest will follow You might think sepsis mortality is already tracked quite accurately. However patients that are recorded as having died of pneumonia have they not died of organ dysfunctions secondary to an acute lung infection. And is this not a very definition of sepsis so not necessarily Accurate data collection relies on accuracy of recognition and documentation. So this is where I would suggest that everybody starts next slide. Thanks Adam. But what if we measure is changing? In fact 2020 the Australian Commission on safety and quality and Healthcare released this epidemiology of sepsis study in Australian public hospitals. This showed an increase in the rate of sepsis for the period of 2013-14 to 2017. However attributed this rise not to more sepsis cases, but to improvements in both the recognition of sepsis and the Australian coding standards used by clinical coders. Loading practice is the Cornerstone of our data collection if you don't recognize or document it didn't happen. There is still a significant proportion of sepsis patients that are under recognized or underdocumented. So it is important to consider this when analyzing your data Trends in total cases. Next slide. Thanks. In order to be able to accurately collect data in relation to sepsis. We need to rely on our coding. What coding happens in Queensland hospitals clinical coders may use different methods dependent on the hospital and available systems and this may result in variability of data collection. I'm iron paper sites across Queensland use ICD-10 coding or the international classification of diseases. Icu's or units using metavision use and speak coding which feeds through the Australian and New Zealand Pediatric Intensive Care registry. Private sites may use a modified version of these or choose their own. Regardless of coding type however, coders are not just necessarily clinicians and must rely completely on the accuracy of your clinical documentation. next slide What does happen when the documentation is unclear or ambiguous? This has a flow on effect to everything. Our hospitals rely on from not only inaccuracy in our data collection to additional workflows reduce resources all leading to the potential for inadequate Staffing and poor patient outcomes in the long term. So while it is time-consuming clear documentation is critical. next slide clinical coding practices are updated every two to three years to improve our accuracy of data collection. Whoever quotas are still faced with significant challenges when it comes to clinical documentation. They are bowed by strict criteria before they can assign a sepsis coding sepsis could only be coded when in this documented clearly that is either confirmed or suspected and documented that that is actively managed as such. Even if sepsis is initially only suspected as a differential diagnosis, as soon as you commence active stepsis management or treatment of the patient and have documented this then it can be coded for unless it is clearly ruled out when other another diagnosis is confirmed. The patient cannot be coded for sepsis if they're simply documented as being septic somewhere in the nodes or have a vague diagnosis of Euro sepsis or chest sepsis. For example. Active use of a sepsis pathway counts as a sepsis management plan. However, it must also be documented clearly in the notes that you suspect your patient has sepsis and are actively managing managing with a pathway and the pathway needs to be completed and scanned in appropriately for coders to be able to view in retrospect. next slide our health information liaison team have kindly provided a specific example of the impact of clinical documentation on their ability to code for sepsis. Here, we have a two-year-old female presenting to Ed with fever vomiting lethargy and non-specific stomach pain. After a clinical review of your patient you decide your receptors is the likely different for diagnosis. next slide the patient has blood and urine cultures taken commenced on Ivy antibiotics and given an IV fluid bolus followed by maintenance within 24 hours both blood and urine inches become positive for E. Coli confirming an E. Coli sepsis with urinary source. If the clinical note is documented as per The Orange Box here. This is too vague for coders to identify as sepsis and this patient would be coded as a urinary tract infection the medical don't in the blue box accompanied by a clear summary note would allow for a coating of sepsis. So don't also the difference in financial remuneration between Documenting in your node as well as scanning in a completed sepsis pathway gives quotas clear ability to see if management offsets was commenced. next slide to give another example for an infant presenting with neonatal sepsis documentation of a diagnosis as only potential sepsis does not give clinical code as the ability to code for sepsis until someone documents that the diagnosis is confirmed and they are treating for it and not just a suspected differential. Otherwise coders May assign observation of newborns for suspected infectious condition if no clear follow-up documentation is provided even if you diagnosis is still not necessarily confirmed with any positive cultures. If you are continuing to choose to treat as such and the patient is responding to treatment then it is important to document this so they mainly code it for sepsis. Many of our neonatal cases are culture negative so you can document for example culture negative sepsis responding to treatment with two fluid boluses and empirical Ivy antibiotics. next slide here are some practical Solutions given by the health information. The informatics layers and team to improve ability for codeine is always important to remember that your notes are not just for other medical professionals coders are not necessarily clinicians make your notes clear and concise linking your potential diagnosis with a management plan document clearly when the diagnosis and management plan changes and avoid the use of medical. Jargon. Use of the sepsis pathway will clearly confirm if a senior medical review has confirmed or ruled out a potential sepsis case. Next slide. Thanks. If you are planning on data collection at your site, please liaise with your clinical coders. The health information liaison Hub has been set up to help clinicians understand coding practices. So there are lots of resources out there for anyone that's interested. next flight So let's take a practical look at some pediatric data collection in conjunction with coding methods here. We requested data for the total number of pediatric ICU admissions across Australia and New Zealand relating to sepsis or septic shock in the last three years as per the ants big coating. And not ICD-10. The infected codes on the side here are used to close code for sepsis cases. If it was the primary reason for admission to ICU noting here that the ants pick registry is only included less than 16 year old patients. Whereas locally, we would classify less than 18 year old presentations as pediatric sepsis. Also, noting the list of infected causes for reason to deteriorate and come to pick you is not exhaustive. For example, we've seen several recent cases of flu away or covid related sepsis. So unless these are documented as pneumonitis in the ants big coding they may be excluded from the data set. Looking at this data alone. However, we can say sepsis or severe infection was responsible for 12.6% of all pediatric emissions to ICU in the last three years and that having a diagnosis of sepsis when being admitted to ICU on average doubles your mortality risk, not such a small number, but potentially could be underrepresented. next time we wanted to compare how our tertiary Queensland pick you accept the starter Compares with the ants big data overall. If we can't currently answer. What is the true incidence of pediatric sepsis in Queensland, then surely we could answer. What is the true incident of pediatric sepsis at a tertiary Pediatrics in single Center ICU? Here we collated retrospective data for total admissions to qch pick you over the last three years where sepsis or septic shock was coded as the primary reason for admission for admissions for the same episode of care were excluded. Easy, right? Well, not exactly as before mentioned retrospected about data collection retires relies on entirely on clinical documentation. So if septic shock was the primary reason for admission, this was clear and easier to collect although not necessarily as I showed you in my first case, however sepsis without shock becomes a lot murkier due to difficulties recognizing and defining in the early stages. Also, if we wanted to additionally capture the incidence of patients developing Hospital acquired sepsis during the course of their ICU admission it became evident that this proved extremely difficult to capture retrospectively. The amount of patients earning empirical antimicrobial commencement for fevers either it requirements or increased ventilatory requirements, which you could say by definition is sepsis. We're not able to be captured. As it is not routinely documented as such in the notes. Habitual documentation of the exact indication and rationale for escalation of treatment as an inpatient is quite variable. Collecting metrics such as time to antibiotics also becomes extremely difficult to collect. When and how do we Define Time Zero when they first presented unwell to ID, and how does that map when you're an impatient? Under recognition is our biggest barrier to sepsis management. So we car reliably use Time Zero as when we actively started treatment because how do we then track if this was too late? That one is that we all need to improve our recognition and documentation in order to accurately capture the true incidence of Pediatric Services in Queensland looking at this we can say that three and a half percent of our total admissions in the last three years were coded for sepsis or septic shock, but we were unable to determine our total incidence of sepsis. Voting this starters slightly different to the 8% Adam mentioned at the start of the talk and the 12.6% from the inspic data that we saw earlier. Next slide. Thank you. Digging into what data we do have who are our patients? What organisms do they have? And where are they coming from? The majority of our steps is patients are not necessarily the immunocompromised many are neonates with no comorbidities and many a culture negative. In order to track organisms responsible for sepsis admission. So however, this requires the Glorious manual collection unless it's clearly documented in medical note occasionally a patient will have a clear coding for Group B strep sepsis for example, but this is very inconsistent. From this however, we can say that over a third of our patients come from into Hospital transfers across the state approximately a third come through the doors of our Ed the remainder come from within our own Hospital Woods highlighting earlier recognition is just as tricky and important in a hospital setting as the Ed and this really shows the importance of the need for a consistent sepsis screening tool pathway across both inpatients and the Ed setting. next slide I'd like to leave you with another case example, which is another actual patient this little four month old female presenting to Ed with febrile seizures. One term Neil medical history cultures taken and commence on Ivy antibiotics for a paraphernal abscess. She was admitted she was administered an ivy fluid bolus for hemogenic instability intubated and take taken to the theater for Urgent draining of the abscess and then transferred to pick your post. Her blood culture was negative but a swab of pass from the abscess taken in theater was positive for methicillin sensitive staff Aureus early Source control and treatment obviously prevented progression to septic shock and a bacteremia. However, this is a clear case of sepsis but was not coded as such. What was the coding for paraphernal abscess and figure procedures remembering the paraphragio abscess was not one of the infections specifically listed our instant coaters list and it seizures can be a red herring in a pediatric sepsis presentation. If you are collecting data at your site, remember to understand your local coding practices and optimize and educate rotating stuff on clinical documentation. next slide picks without a definitive diagnostic marker capturing all true sepsis patient at cases isn't currently possible sepsis diagnosis coding is dependent on clinical documentation. However, differences including practices can also create inconsistency and data. A more objective approach to report sepsis cases is currently being investigated using clinical markers. For example, investigations collected medications given Vital Signs Etc from the electronic health record. However, it's unlikely to be perfect with a likelihood of false positives false negatives and miscases. Clinician recognition and documentation is an always will be vital to a sepsis diagnosis. Also, these electronic recognition tools would only be applicable to fully digital sites of which there is only currently 14 of 122 across Queensland. Next slide. Thanks. In summary my advice that everyone would be to start with improving the recognition and documentation of sepsis at your site. Use a sepsis pathway encourage that it is completed and scanned into the medical record and document that this is your management plan if it's commenced. Document clearly in your notes for both escalation and de-escalation of treatment or when identifying a different diagnosis for your patient. Embed sepsis Champions into specific clinical roles, so they may be sustainable over time and provide continuous education and data collection. And most importantly share your data between hhss impedes. We are already dealing with smaller numbers in comparison to adults. So pooling Statewide data not only gives us improved statistical accuracy, but it also allows us to see where resources are required and to enable oversight of sepsis type and incidents to allow accurate updating of a Statewide pathway. Science developing individual solutions for sepsis dashboards without collaboration may lead to the siloing of data. If we start with accurately recognizing documenting and collecting sepsis instance starter than other data metrics will flow on from this and hopefully future Innovations will continue to help us improve. And on that note. I'll hand you back to Adam to talk about a little bit more about some of our future directions. Thanks Adam. Thanks so much Mal. And that's great. I'm gonna do this awkward transition that I did a couple of weeks ago. That didn't seem to work all that. Well where I Come back to teams. You can at least hear me at the moment. I hope right. Okay, so I thought Mel was that was a really clear explanation of how how important it is to have mechanisms for capturing data how we need to understand our definitions Patient Group definitions the definitions of the terms. We're trying to measure And then we need very clear documentation. And I think that whether we're in an IMR site or a paper site, it's it's a critical importance isn't it that we get and clear documentation of the patients that we're managing and the interventions. So I'm just going to take a moment before I go back to the slides to see where our audience is from we've got plenty of chat in there. Haven't we? Cook down. Taurus and Cape Sunny Coast great Cairns we've got somebody from UE. That's great Donna Simmons from you. and Forsyth I don't know where foresight is what I proposed with somebody did last week was put a put a pin in a Google map and show us just how far away you are and you know, then you might get the prize. I don't know all the prizes but so to go back to to the slides and I hope I can do that. So and it smells stressed the importance of documentation. She's stressed the importance of and well one of the things that we always focus on is just how difficult is to recognize sepsis acutely to recognize sepsis in order to In order to intervene. So that's a perspective recognition of the child with sepsis. And I think what she illustrates is that it can be quite difficult even retrospectively and to recognize these children. And in order for us to report against the sepsis clinical Care standard. We really need as much information as possible and that requires clear documentation if we're going to be able to recognize this population of children and against which we can then report for the standard the sepsis clinical Care standard as we think about the requirements of the steps of clinical Care standard and there are there is overlap with the sepsis clinical Care standard with the number of other important clinical care standards the most obvious one of which is the antimicrobial stewardship clinical Care standard So within the answer microbials clinical Antimicrobial stewardship clinical Care standard. There's a requirement to report about appropriateness of antimicrobials. And the antimicrobial stewardship standard does in fact offer some resources that you can download against which you can then audit your practice in order to report for for that standard so you can we can perhaps put the link in the chat again for these resources that are available on the commission's website. But this is an example of an audit tool for capturing clinical data for the antimicrobial stewardship and clinical Care standard. so we need to be able to capture these data and there are significant opportunities as well. Of course if we can make use of the data that we have available to us the IMR in particular offers a really great opportunity to be able to use routinely collected clinical data, not only for reporting but for improving clinical care, and so One strand of our program and some of our goals relate to how we're going to use data better to improve the clinical care of of children. So that includes a translation of our paper-based sepsis pathway into a digital pathway and digital decision support tools. It includes the development of dashboards and dashboards could measure exactly those. Those things that we need to report back to the sepsis clinical care standards. So the establishment of clear definitions that are then documented in the form of a dashboard which allows hhss to measure their performance and Benchmark their performance against others would facilitate the reporting to the sepsis clinical Care standard. If we could agree on standardized definitions, we could establish a sepsis registry in a sepsis registry something that many of us feel would be a really valuable tool to improve the care of children with sepsis. And if we're going to do this we need to be able to capture these data sustainably. And if we have a sustainable method for capturing. Large clinical data sets then that introduces the possibility of the use of more Innovative applications of data such as artificial intelligence and machine learning to direct to direct. Sepsis care and of course, none of this happens unless you're adequately funded. So we're working with Partners to try to identify opportunities for funding. So what I'm going to talk about now is our progress towards a digital sepsis Pathway to move from the paper pathway towards a sustainable solution to to a digital pathway and and alongside that I'll talk about what we've been doing to build a Statewide sepsis dashboard. I'll make some reference to the approach to a sepsis registry and then the work that we're doing towards developing digital sepsis prediction tools using artificial intelligence. So this is work that has been done by. And one of our team a clinical implementation called briskio along with his Partners at ehealth Queensland and which was looking to translate the the paper-based pathway screening and recognition tool. Into a care pathway within the IMR to direct clinicians to the most optimal empirical treatment. And this work has been ongoing for a couple of years now and it's made substantial progress and overcome significant and challenges and barriers but we still are working to overcome some of those remaining barriers and that requires a partnership with Cerna and a new build of a care pathway and the implementation timeline for which is unfortunately not as clear as we'd hoped. So it's it's unlikely to be delivered in the next 12 months, but we're working hard to towards that. and the other aspect of our work which aims to use the data within the IMR better for for clinical care and for reporting is the development of a of a dashboard and we've done significant work with our partners at chq in the bi team. Particularly Catherine Arnett and Reed mile seed and who've done significant work drawing data out of the electronic medical record. And applying algorithms to identify particular patient groups and identify those process measures that I am described in our emergency department pathway implementation earlier in the talk and to do that automatically to be able to present data for individual institutions that they could then compare against other institutions. And so far we've established a pilot at qch that we'd like to be able to to establish the clinical care and we have been approached by a number of other. Hhs's and to use this work and to use the coding that's gone into this work for their own dashboards, and we're keen to support that. So this is an example of the dashboard that has been developed by Bruce and the team at chq. And you can see that what we're trying to do is first of all identify the right patient groups. So over on the left hand side, you have different patient cohorts suspected sepsis sepsis with organ dysfunction sepsis with shock and the IMR and uses the clinical data that's available to to put patients into each of those individual cohorts and from that then within those patient groups, we can look at our performance against different process and outcome measures such as utilization of the pathway and compliance with the bundle and I would just caution that these are these the early days when we've been trying to establish our definitions, so I wouldn't be too worried about the low numbers that you're seeing here and the bundle compliance. And and along with those measures of process measures and outcome measures. We also were Keen to be able to incorporate. Data from the prescribing system much like I did for the Emergency Department evaluation so that we can see our Trends in antimicrobial use over time as well. So this is an antimicrobial stewardship dashboard within the sepsis dashboard. Excuse me. now what we once we've established a system for capturing data within the electronic medical record and we we then would like to be able to move to it to a system. Whereby those data that are coming out of the electronic medical record can go through an analytical algorithm. To try to predict the onset of sepsis. We all know how difficult it can be to recognize sepsis in children. That's the that's the premise upon which this whole talk is based and and what the variables that we use to try to predict sepsis. Are all captured routinely within the electronic medical record and if we could establish substantial data sets against it within which we could derive and validate our predictive models using artificial intelligence. And it may be that we could open up the opportunity for AI driven sepsis prediction. And this is becoming a reality in a number of different parts of the world. this very recently published evaluation from the States from Boston was a multi-sight evaluation of a system called Truse a learning based early warning system for sepsis, and it was a very impressive piece of work evaluating hundreds of thousands of clinical episodes, but some of some of their interesting outputs were that the machine learning approach really had to had to combine with clinicians they talked about Human machine teaming and partnership between clinicians and the Machine learning algorithms. And I think this is a really critical aspect of of machine learning there has to be a process by which the clinicians interpret the output from these machine learning algorithms in order to influence their clinical decision making so rather than viewing the system as a surrogate for clinical judgment clinicians perceive themselves as partnering with the technology even without a deep understanding of machine learning clinicians build trust through experience expert and endorsement and validation and systems designed to accommodate clinicians autonomy. We know that so far these machine learning algorithms are not sufficiently discriminatory and to make decisions on their own but in partnership with clinicians, they may improve the number of good decisions made. There are significant efforts on going within Queensland to to deliver a similar data-driven sepsis prediction. This has been LED firstly by a group in Townsville. And now out of Metro South including something called the Queensland Health sepsis AI working group. And in children, we in the Queensland pediatric sepsis program partner with an arc funded Center for information resilience and which brings together substantial data expertise with our clinical Partners around around Queensland in order to develop in order to develop data-driven prediction models for sepsis in children, and we're going to focus on both ICU and a data intensive environment and data poor environments like the inpatient Wards Ed and even even outpatients. And and that partnership with Center information resilience has also place the Queensland pediatric sepsis program the center of the network and a national network with Partners from new experts in New South Wales in learning Health Systems and along with Partners in in Victoria and Western Australia as well and together. We're looking to develop a standardized harmonized Pediatric Services data set against which we might be able to develop these machine learning algorithms and then cross validate in in each of the different states. So it's a really exciting time to be doing this work in Queensland at the moment. So the key messages then from the talk is is of course. That sepsis is a leading cause of preventable harming children the national sepsis clinical Care standard was was launched in June and the purpose of this that the clinical Care standard is of course to improve the care of children with sepsis to standard hours standardize are approach and to facilitate the mechanisms by which hhs's measure their care of children with sepsis and Benchmark that care and report it back. The implementation of the sepsis pathway facilitates monitoring and Reporting against the sepsis clinical Care standard and indeed the implementation of the pathway achieve some of the quality standards of the path of the standard. And and here in Queensland, the significant work was ongoing to better use data in the IMR to predict sepsis onset and improve outcomes. To establish those World characterized sepsis cohorts to overcome the challenges that Mel was illustrating about the difficulty of recognizing patients who do indeed have sepsis and for monitoring and Reporting against the sepsis Care standard. There are lots of useful links here and I'm just going to point you as well to the Pediatric sepsis website, whether I'll be lots of tools. This is a reminder again for you to feed back to us if you would and I think by now Sarah probably has put the link to the evaluation survey. We really value your feedback so that we can adapt our content appropriately. I'm going to flag that the next of the Pediatric seminar series will be presented by. I think Alana English one of our Advanced social workers and Megan O'Keefe and they'll be describing our approach to family support in in sepsis. And those those seminars will take place on the 6th and 7th 7th of September leading up to World sepsis day, which is on the 13th of September. And again, I think the registration link will be in the chat. So I'd encourage you to register that and that will be the family support work that's being done by Queensland Pediatric Services Program is really world-leading. There are absolutely critical part of the program, and I'd really encourage you to listen to that. And I just flagged that of course we were very lucky and privileged to be awarded the global sepsis Alliance award in back in 2020. But in fact we were also finalists for the Queensland Health awards for excellence tomorrow. So we're waiting to hear whether we might be awarded for our program in the Queensland Health Awards tomorrow, and I finally just like to acknowledge all of these people and many many more indeed in particular are consumers and I've only highlighted two here Mary Steele and Amy Wilkinson who are on our clinical Advisory Group, but they're a great many more who make a really substantial contribution to the program. I'm going to Stop there and ask if there are any questions. And if there are questions, we've got some snapshots. Excellent. Would anybody like to ask any questions? Are there any questions in the chat Mel. Did you spot any questions in the chat? No, but I think I'm not sure if forces or cooktown is winning. And the most remote place online today? We may have to ask Jana or Tracy to tell us who they think wins. I think we're equidistant. Very diplomatic. Janet Forsyth is a much smaller town than cooktown though. It's much more remote. Okay you in? What size is a single nurse post we get after? Yes once a fortnight wear it? Yeah Yuri, but you've got the Savannah Lander. Yes, we have this event. Oh Tracy, we're delighted that you joined us today, and I'd like to extend that to everybody else as well. Unless there are any other questions then I think will wrap up. Thanks everybody. Thank you. Thank you everyone. Thanks. Well done, Tracy. Thanks, Jenna. Ciao.