Opinion: telehealth is not quite the colt from old Regret but it sure as hell has got away
Telehealth has bolted away as fast as the colt from old Regret as a result of the Covid-19 pandemic. The federal Health Minister, Greg Hunt, says he wants telehealth to continue post pandemic, but for that to happen, the industry will need to offer more than just ‘courage in our quick impatient tread’, as the ode goes.
We will need to understand and support the government’s need to break in the colt, saddle and hold it on a tight rein, because the sustainability of Medicare depends on it.
How to do that? Well, the Man from Snowy River, though a wonderful story, can’t tame Medicare, but a basic rule of data can: collect data used to care for the patient at the point of care, not as a separate fiscal claim. In this way, administrative overheads are reduced and real data from the patient’s care are used as the evidence for claiming.
It sounds like a lovely theory, but it is one we are in a position to make a reality. SNOMED CT codes are the taming tool and using them as a condition precedent to ongoing MBS telehealth ticks a lot of boxes.
Under Medicare, Australian governments have never been able to see what they pay for. If you think about MBS item 23 for a moment, all it tells the government is that a patient saw a GP for between five and 20 minutes, nothing more. MBS item 132 tells the government a patient spent 45 minutes with a physician, but for what? A headache or a heart attack?
So when it comes to telehealth, how will the government know whether a telephone call between a patient and a doctor ever took place, or how long it really took, when the current honour system of billing is solely dependent on provider knowledge of the system and personal integrity?
In a recent post on LinkedIn that garnered a lot of attention, a user of a telehealth service described his disappointment with the short, four-question ‘chatbot type exchange’ he received, which was bulk billed to Medicare using Covid telehealth MBS item 91810, drawing $73.95 from the public purse for a consult of over 20 minutes. Problem was, this service was never provided, as the user took a screen shot at the end of the chat (there was no verbal contact at all), showing the time elapsed as 04:48.
The matter was reported, the claim withdrawn and apologies given, but the truth of the matter is that non-compliant Medicare billing costs Australian taxpayers at least $1 billion annually. Precise quantification of this problem has proven elusive, though some commentators suggest the figure is as high as 20 per cent of the scheme’s total cost.
This example is certainly at the more egregious end of the incorrect billing spectrum, bordering on criminal fraud, and the government has a constitutional obligation to do something about it. We know that jumping on this type of conduct early, rather than relying on expensive post payment audits, is the right approach, and digital health makes that possible now, with the right codes and technology.
Coding non-admitted patient encounters
It may come as a surprise but there is one thing the US health system has that Australia needs but doesn't currently have: payer visibility. US payers can see what they are paying for. In Australia, Medicare reimburses ‘clinically relevant’ services, whereas the US reimburses ‘medically necessary’ services (no, they are not exactly the same thing) and the mechanism for determining medical necessity in the US is coding, including the coding of non-admitted patient encounters.
Every single non-admitted patient encounter is coded using the US system of ICD10-CM/CPT codes, which has been adopted in many countries including Indonesia and much of the Middle East. So, there is nothing novel in this approach, with much of the world already coding non-admitted care. We are not suggesting the wholesale adoption of the US system of health data collection because it has its own deep flaws and is over burdensome, but rather, our suggestion is that we take the Covid opportunity to leverage our progressive electronic health record systems to inform reimbursement policies and deliver a more accountable healthcare fiscal policy while not creating a clinical burden in data capture.
Australia has already agreed that ICD is not the right code set for our digital health strategy of interoperability, and was an inaugural member of the organisation which owns and manages SNOMED CT. ICD is a classification designed to count activities, diseases and procedures, and is widely used for public health analysis, fiscal and service planning and reporting.
SNOMED CT is designed for clinical use directly for patient care. There are approximately 70,000 ICD codes but millions of SNOMED CT codes, so the level of specificity is much greater in SNOMED CT, which is designed to represent data directly in clinical systems retaining the precise meaning intended by clinicians. Moreover, many Australian software vendors have already integrated SNOMED CT into their clinical systems, so we are over halfway there.
In the example of the non-compliant bill, extracting relevant SNOMED CT codes from the patient’s record into the claim before it was submitted would not necessarily have prevented the egregious conduct, but it may have made the doctor think twice, because each additional cognitive step in the billing process allows for the entry of the conscience and learnt billing practices and ethics.
If in the ‘reason for attendance’ data field of the PMS all that was entered was ‘repeat script’, SNOMED CT would have returned code 182918009 (repeated prescription (situation)). This would enable the government to ask the right questions without having to issue expensive, time consuming, legal requests for clinical records when concerns arise. Why did a repeat script take over 20 minutes to issue? Additional reasons for a single attendance are of course common and easily accommodated by SNOMED CT.
Just as claiming a tax deduction requires a reason, no one could argue the government should not also require a reason before distributing taxpayers' money, particularly via an additional code that is already integrated into most systems and can easily be automated. It would actually reduce the administrative workload for doctors compared to their current burden of recording adequate and contemporaneous records to meet Medicare compliance requirements.
Further, SNOMED does not require human coders nor is it designed to be used in that way, so is cheaper to implement and administer. There will of course be privacy considerations but none that we have not already overcome, noting that the only thing Medicare needs is the codes, not the complete clinical record.
Visibility over billing
Other benefits of SNOMED CT are that by giving the government what it sorely needs in terms of visibility over billing, it may be less likely to restrict telehealth services in other ways such as by allowing only one telehealth consult per patient per day or per week, which often does not align with appropriate clinical practice. It is not the number of claims that matters, but the reasons for those claims and we must accept that, in the current context of a government paying blind and non-compliant billing being a significant problem, likely exacerbated by telehealth, that government will have no option other than to impose restrictions of some sort if the industry doesn’t come up with something more appropriate.
We tested a SNOMED/MBS claim on our ECLIPSE enabled billing software and were able to successfully transmit a claim in our sandpit environment by using ‘MISC’ in the item code data field and putting the SNOMED CT codes in the service description. This means there is zero development work required to pilot this right now by those who are already SNOMED CT enabled.
Further, for the majority of providers who already take Medicare compliance seriously and bill correctly, there is nothing to fear or lose, and the benefits to the government and the Australian Digital Health Agency are obvious. Patients will be the big winners here by being at reduced risk of receiving unnecessary services and by improved public health data being reported to My Health Record.
Acknowledging and accepting there are multiple reasons for a patient to visit a doctor and symptoms are converted to a problem/diagnosis list, the SNOMED/MBS combination will enable the government to quickly identify outliers and combinations that just look odd. If every patient presenting to a particular clinic with the SNOMED code for ‘cough’ has three moles removed, an ECG, a mental health treatment plan prepared, a 40-minute consult and a brain scan ordered, then questions can be asked.
It won’t solve everything but will certainly slow down creative billers who describe the services they provide with scant attention to clinical relevance or indeed the ethics of medical billing.
Mandatory bulk billing
Solving the government's Medicare compliance challenges will take more than SNOMED CT but using it will set us on the right path. SNOMED CT is already used in My Health Record reporting, so if it is adopted as suggested, SNOMED CT representation of the reason for attendance should be required by everyone billing telehealth, not just GPs and not just doctors for that matter.
In the interests of the national digital health strategy, if introduced, SNOMED CT should apply initially to all telehealth billings no matter who the provider, including allied health, nurse practitioners, optometrists etc. Everyone should be required to add SNOMED CT codes before their claims will be paid.
Enablers will include the government scrapping forced bulk billing by GPs, which is of questionable constitutional validity anyway, and scrapping the pre-Covid telehealth items (which are already causing confusion among doctors – do I bill Covid telehealth or usual telehealth?) This approach should also mean no costs to clinicians, only to software vendors who have to set up the system. However, it is in the interests of the government to subsidise these costs, recognising it as an investment likely to reap rewards of many magnitudes more than it will cost.
A final consideration is data quality, so systems claiming to be SNOMED CT enabled should be required to be conformant to quality terminology implementation before they are accepted in the Australian marketplace, or have plug-ins that support the relevant standard. Any vendors using maps to translate from their existing code systems to SNOMED CT should be required to be quality assured such as by conformance to the recently published ISO/TS 21564:2019 Health Informatics – Terminology resource map quality measures (MapQual). This should be mandatory.
Doing nothing about compliance with a broad roll out of telehealth is not an option, and at some point in the evolution of our health system we will need to start coding clinical information about non-admitted patient care. Covid has propelled our health system into the future and presented a once in a lifetime opportunity for this necessary change to happen now. We cannot let it pass.
If Greg Hunt becomes known as The Man from SNOMED River he will be hailed as the minister who modernised Medicare responsibly. However, the political will to immediately run a pilot tying SNOMED to MBS telehealth billing will require ‘pluck that is undaunted and courage fiery hot’.
About the authors
Margaret Faux is a solicitor specialised in Medicare and health insurance law, an RN, the founder and CEO of medtech company Synapse Medical and a PhD candidate on the topic of Medicare claiming and compliance.
Heather Grain is a health informatician, clinical coder, health information manager and digital health expert. She is the current chair of the international ISO TC215 health informatics semantic content working group, a former chair of the SNOMED international education working group and a co-chair of HL7 International’s vocabulary committee. She is also a member of the Synapse Medical team.
Posted in Australian eHealth