A complication coding conundrum

The content discussed in this episode may be specific to the guests organization or location and may not be applicable to other organizations, states, territories or countries. The material presented in this podcast is for general information purposes only and should not be considered health, legal, or financial advice.

Medical administration is a rewarding profession, but it comes with challenges. So what is it like and what should you do if confronted with a monumental administrative difficulty? This is the podcast. Takes unfortunate scenarios. and explores the essentials of clinical governance to find solutions. I'm Dr. David Rankin, a very experienced medical administrator and lead fellow with the Royal Australasian College of Medical Administrators. And this is Safeguarding. The essentials of clearing.

Today, a complicated problem. The measuring of hacks or hospital acquired complications are really quite crude measures. That's Dr. Paul Tridgel. A former clinical director of health system performance and deputy CIO with New South Wales Health. Paul runs his own consultancy company providing advice to hospitals and state governments across Australia.

And I'm challenging him with a difficult scenario. There are lots of little uh tricks around the definitions and how they're coded in different hospitals which can produce very large differences. In today's scenario, your hospital's head of cardiothoracic surgery comes to see you, concerned that the department appears to have a very high complication rate. This does not reflect well on what they believe is a high quality and very safe service.

Their main concern is the apparent rate of atrial fibrillation or arrhythmia following cardiac surgery. This is a known complication of heart surgery and usually settles over a couple of days. The department head is adamant that this is not a complication and certainly not something which the hospital should be penalised for. I started by asking Dr. Paul Tridgel what major issues this scenario raised. The measuring of hacks or hospital acquired complications are really quite crude measures.

And when you actually drill down to an individual department and try and look at something like arrhythmies and cardiac surgery, there are lots of little tricks around the definitions and how they're coded in different hospitals.

which can produce very large differences in the reported complications, which are no reflection on the quality of the service. So if we dig into this sort of atrial fibrillation one a little bit more, there'll be a group of patients who will have had atrial fibrillation in the past. And a coder when they read the record needs to decide from the documentation, is this occurring for the first time in hospital or has it occurred in the past?

So and this relies on the adequacy of the documentation done by the medical staff to say whether the patient had had it in the past. The second major issue for atrial fibrillation is that this gets coded as a hex. But most atrial fibrillation, once it receives some treatment, which might be some magnesium, a bit of ameroderone, potentially defibrillation, is the patient goes back to a sinus rhythm. Now, the medical staff are able to document this as paroxysmal AF.

And this is coded to a code of I forty eight zero, which is not a hack. So a large part of this arrhythmia's in cardiac surgery is really just how well have the medical staff been educated that they're to document the atrial fibrillation as paroxysmal atrial fibrillation when it goes back into a normal sinus rhythm. So let's walk through the coding process. How does a hospital create a set of codes against which they raise an account or determine the relative funding for that patient episode?

The coders are required to read through the whole set of documentation that relates to an admission. But they're also blinkered to only read the documentation which relates to that admission. So there may be some things which were documented in a clinic they attended prior, which the coders aren't allowed to look at. and so there may be some codes missing. Then the coders need to decide for each code has something been done about it?

And so that a patient has had atrial fibrillation in the past doesn't mean that the AF is coded. It's only if there is a management change, medication, additional investigation that's done about that. And then for each code they need to decide was that present on admission or was this new? The book for the coding standards is about two hundred and sixty pages long. So what I've described is a bit of a this is the average case rather than the specifics. So for example the diabetic code.

uh are coded whether it's the condition is considered significant or whether there was an action or not. This particularly impacts on hospital-acquired complications when you get into the issue of a diabetic who had a mild case of hypoglycemia. And you know, severe hypoglycemia can be a very significant issue, but a mild hypoglycemia with a nurse just gave a jelly bean, for most other codes, a nurse initiated medication is not coded.

But for hypoglycemia it is. And this can drive an awful lot of reported hacks where there was really no impact on patient safety or outcome. So one of the critical things is that the coders can only look at what's there. And if it's not in the documentation or if the documentation is ambiguous, the coders struggle. Yeah, it's also how the documentation is written and how things are linked together. So if we look at the cardiothoracic example again.

is a major procedure where you've opened your chest, frequently opened the leg for the graft, and it's not surprising that you've got some bleeding that needs to be managed after the surgery. Now if this is documented by the medical team as postoperative bleeding, The standards say that this is just the bleeding is just occurring after an operation, not due to the operation.

as if the medical staff document it as bleeding due to the procedure or secondary to the procedure, well then the bleeding will be coded as an additional code. Some people are at much higher risk than others.

So undertaking a hip replacement on a person who's fallen and broken their hip and they're ninety nine. is very, very much more likely to raise a number of complications or hacks than a person that comes in for a colonoscopy. How do you try and adjust for the risk that the patient poses in looking at complications?

Oh, it's clearly age and a very big factor is emergency status. If you look at a a patient with an aortic aneurysm, if they're presenting as an emergency patient, they've probably got about a thirty percent chance of dying. As an elective it's probably well under five percent. So if someone has a major trauma with an open wound

the chance of an infection getting in that versus clean surgery are just poles apart. If someone has a hernia operation and they've got strangulated bowel that's potentially caused an infection in the abdomen, That's a completely different risk profile to someone having an elective procedure, just putting it a bit of processes to stop the hernia occurring again.

So there one needs to look at age but also emergency status and there may be a lot of other factors which impact if someone is immunosuppressed, if someone has diabetes, if someone has obesity or other comorbidities, can all have significant flow on implications to the chance of that person having a complication. So it's not a trivial process. There are the hack system which applies penalties.

does use a few characteristics to say whether a person was at higher risk and they've got a couple of categories and that reduces the financial penalty which is imposed. But it doesn't change that the hack is counted and it's just this straight count of hacks, which is how a lot of hospitals are portrayed and judged on some of the statistics that are published.

So who invented these hacks? How are they defined? There's sixteen of them? Who determined that those sixteen are the ones that we should be collecting rather than the thousands of other hospital acquired diagnoses? Well these were defined by the Australian Commission on Safety and Quality and Healthcare. There are other organizations internationally which also have indicators of quality and safety. For example, ARK in the States. One of the major sort of limitations is that the

said that they need to derive these indicators from the data that's available from the inpatient data set. So they've said that they need to use this information rather than a lot of other information. A lot of the clinical colleges and specialties have what are called clinical registries.

and these collect far more detailed data uh about the patients, about the medications they're on, about the type of complication that occurs. And those registries can provide far more valid markers around complication rates and comparative practice than you can generate just from the admitted patient data sets.

So the approach is then, well let's take this condition like delirium. And with delirium, if you've got an elderly patient and you're doing major surgery on it, if you go and look at a lot of those patients, you will find a lot more delirium than is being reported. And this, you know, the criteria for coding a code is just a yes or no. So it's was it there and was something done about it?

So if there was if the patient was mildly confused but they actually did something about it, if they specialed a nurse or they did some in investigation, well then it will be coded. Now there's things that you can do, make sure that the patient blood salts are at the right level and that you can try and and do to make sure that the to reduce the incidence of delirium, but they don't do major changes.

That raises another real interesting challenge is it depends whether you investigate, whether you diagnose. So for DVT. If a hospital were to scan everybody's calf after major surgery, they would identify a reasonable rate of DVTs. But most of those DVTs are completely harmless and dissolve or resolve on their own and wouldn't be identified or coded if the diagnostic test hadn't been run. So how can we make sure that the identification of the complication is consistent between hospitals?

Yeah, well something like a D V T is also interesting because you're only measuring things which are investigated and potentially treated within hospital as it's quite to be expected that some of those DVTs will become evident when the patient is in the community and will be treated with a GP. And they won't be count as complications at all. One way of of looking at quality and safety is to look at process markers and what

being done. So for example you've got for hand hygiene, you've got people observing samples of the clinical staff to see how they're they're doing. Similarly for DVT, you've got a risk tool which is frequently applied to patients. And then you've got prophylaxis which is done either through subcoheparin or mechanical means to try and prevent a DVT.

probably what would be a more meaningful measure for D V T is to actually know are those screening tests for D V T risk being done and are the prophylactic measures that are known being implemented rather than to what degree was there a an ultrasound done that w actually investigated the DVT or not, which is the marker for the coders adding the code is that they see something done about it.

If someone has a UTI in hospital, the question is was it present on admission or not? But also was something done about it? Now the clinical staff may not decide to treat it, and so it doesn't get coded as a hack.

and those that do treat it or just put it in the discharge summary noting the results for the G P to follow up, that's not a hack because the clinical staff didn't commence Aaron Powell Having said that, it's a courageous clinician that doesn't treat a DVT that the radiologist has identified with.

So we've looked at a number of the challenges around the coding and classification and counting of complications. How do we use the hack rate to try and improve a hospital's performance if there are so many issues, I guess, with the counting of them? Yeah, because of the financial penalties that have been introduced, a very large number of hospitals have staff which are now querying these hacks. and then going back to the definitions to see whether they were really a hack or not.

and you end up with like how you opened this session with the clinicians saying, you know, that's not really a hack. It doesn't represent our complications and the coders and sometimes the quality people sitting on the other side and say, you know, well that's the definition. So, you know, there are some benefits in that it produces a focus around some of the things that can be done around quality and safety and certainly something like pressure injuries.

We want to make sure that there are adequate nursing resources and adequate beds and protectors available to minimise pressure injuries occurring. Maybe doesn't drive the the best thinking around quality and safety to make sure that the especialist staff that we have are really focusing on the biggest quality and safety issues that something can be done about. I think it's sometimes easy to criticise the data and miss the opportunity.

Patients in Australian hospitals shouldn't develop urine retract infections. Ideally they shouldn't develop delirium while they're in hospital, particularly if it's avoidable. They shouldn't develop malnutrition, because people haven't screened or been aware of their risks. So How do we make sure medical staff actually concentrate on the preventable complications and do what they can to reduce their numbers?

Well, certainly I've participated in education sessions with junior medical staff and specialists at some major hospitals. where we talk about delirium and the factors which can be managed to prevent delirium and similarly, you know, urinary tract infections. And there are some programs with some very specific

quality indicators like infections around central line insertion. This was research done in the States which identify that really having a dedicated team for inserting central lines to make sure the most hygienic processes and using doing them in theatres and not in just any ward environment. and then measuring the number of infected lions per thousand patient line days. It provides a great way where a hospital can change its procedures and practice.

and it can use available data to monitor its outcomes and benchmark that performance against potentially other hospitals internationally. I think benchmarking's an incredibly important component of that. You develop a tolerance within your individual hospital often for s some complications, but when you benchmark with others you realise that there are opportunities to improve. Your observations about central line infections, I think's a really pertinent example of the ability to reduce rates.

And over the last ten years, Australia's seen a dramatic reduction in central line infections by implementing good standards of practice. Another issue sort of linked with say infection and infection risk. which I worked on a few years ago, r related to the use of blood and blood products. When you administer blood, it does disturb your immune system significantly and can lead to higher infection rates.

And so appropriate issue there is to make sure that you're only administering blood when it's really needed. and I did data linkage with the Clinical Excellence Commission looking at the levels of hemoglobin and the use of blood and we saw that some patients had hemoglobins which were close to normal at sort of ninety or a hundred and yet were being transfused one or two units without having a clear clinical indication for that.

and by doing that linkage and the benchmarking and taking that back to the clinical staff involved across the state through the Clinical Excellence Commission. We reduced the use of blood by about twenty seven percent, which was quite a remarkable achievement. Extraordinary achievement given that blood is a scarce resource and expensive.

I guess one of the themes that we need to emphasize is that you need to work with the clinicians and find out what the meaningful complications for them are and what they're prepared to work on. And then commit the resources to reduce those complications, implement standard processes, and then celebrate when the complication rates come down. Paul, looking at All of the...

issues, I guess, and challenges in setting definitions for complications. What are some of the key lessons we can learn about improving quality and safety? And Counting complications. If if you want to implement change in a hospital, look at your established processes which can achieve change.

Which is your sort of process assessment, looking at your indicators, looking at your interventions and getting some feedback. As you've mentioned that involves engagement of all of the key staff involved in the process. as y you need in indicators that are meaningful and an underlying analysis and understanding of the problems and the opportunities to get meaningful change.

having an indicator that isn't able to be used in a process improvement methodology is will frequently just look at the noise of that indicator rather than getting the change on the ground. I guess in winding this up, do you have any key messages that you'd like to make in relation to quality assurance, coding and complications? Don't jump to conclusions when you see a high level reported in something like a hack.

without delving into the details of the patient mix, how the data is collected, and where you do see some variation look at what other data you can bring to that analysis.

And then go back to the clinical records and review a sample of them to get an even deeper understanding. There's an awful lot of high level analysis and correlations that are done and not enough Taking that back to look at the processes that occurred in detail on the ground to inform what an appropriate response if a response is required. administrator now working as a private consultant. And that brings us to the end of this episode of Safeguarding Healthcare. I'm Dr. David Rankin.

Follow Safeguarding Healthcare free on Apple Podcasts, Google Podcasts. Spotify or wherever you listen to podcasts. This podcast is produced by Sound Cartel for the Royal Australasian College of Medical Administration.