GPs use 2% of the Medicare budget and see around 85% of the Australian population every year. So why has general practice been relentlessly targeted for optimising, refining, reforming and restructuring?
In a recent Health Services Daily opinion piece, Ian Manovel clearly articulated the case to increase technological investment in primary care.
I expect Ian, as head of Industry (Health) at Fujitsu ANZ, is highly expert in the capacity of technology to address a range of contemporary problems, especially in business and economics.
I’m not so sure he understands general practice.
As a qualitative researcher, I am well versed in reflexivity, the importance of describing my context to show the reader where my biases may lie. So here goes.
I am a GP, the lowest paid doctor in the country in a dying industry. I have taught distance education since courses arrived in manuals with cassette tapes. I’ve worked with the Royal Flying Doctors Service giving advice over radio, and I have been providing clinical supervision and healthcare over technology-enabled platforms since I graduated in 1991. I am not a technophobe.
I also have a Masters degree in value-based healthcare from the University of Texas, so yes, I do understand the economic arguments around rationing, and the measurement of quality.
I have a number of career options that would pay me more than the $50 per hour I earn seeing adolescents who have deep mental health needs who are sleeping in their cars. Instead, I am in general practice, the current scapegoat for all our healthcare dilemmas.
Mr Manovel works for Fujitsu. Fujitsu Australia and New Zealand is “a leading service provider of business, information technology and communications solutions” with a net worth of around $22 billion. He has experience as a hospital and community pharmacist and describes himself as a leader who understands “the economics of success and identifying innovative, digital solutions to achieve it”.
Now we have that out of the way, let’s deconstruct the arguments for technological revolution in primary care. He has summarised them so well, I thought I would address a counterargument, addressing the issues point by point.
In doing so, I’ll use a typical case in general practice. Just to be clear, these days, multimorbidity is the norm because the “simple” issues have been harvested by other services for greater cost, “freeing up” GPs to do the harder work at lower cost.
Poverty is common in our practices, especially for women. Complexity is the norm. Jane is a 50-year-old patient with diabetes, heart disease, obesity, treatment-resistant depression and COPD. She is a survivor of early childhood trauma, so consultations can be challenging as she has difficulties with trust. She has had a number of domestic violence partnerships, and has four children, one of whom is seriously ill with stage 4 breast cancer, so she is caring for the grandchildren. She lives on a disability support pension with unstable housing. She cannot afford any other services than me.
“Virtual connected care can put patients first”
Yes, in fact there have been innovations to address the health needs of this wide brown land for over a century, starting of course with the telephone. Pedal radios changed the health landscape for outback care in the 1920s, and it has been GPs who have introduced some extraordinary innovations. The one I most remember is Geoff Mitchell introducing iPads to palliative patients in western Queensland so they could check in with their teams virtually, and die at home. Just enough technology to get the job done.
So yes, virtual care can put patients first.
“Healthcare demand is infinite whereas healthcare supply is finite” so expenditure needs to be rationed using activity-based funding
Mr Manovel argues that fee-for-service funding is a “blunt instrument” for rationing funding. One could argue that salaries are also “blunt instruments”, based on the number of people at a certain level of skill we reckon we need, which works for other professionals providing services (teachers, nurses, road workers, economists), but I digress.
He argues that a better system is activity-based funding. Activity-based funding in hospitals relies on coders, who trawl through the clinical record and extract relevant information about the episode of care: the diagnosis, features of the patient, usage of services and so on. Each code is assigned a financial cost, and the cumulative cost determines the payment for a care episode.
All very worthy, of course.
Mr Manovel argues that “fee for service” funding is “a blunt instrument” that is no longer “fit for purpose”. Of course, I would argue that we don’t really have “fee for service” in general practice. A consultation item number involves a fee for multiple services, all integrated [LS1] into one essentially time-dependent item number.
As a doctor, I can’t help seeing healthcare interventions the way I see prescribing. I blame Dr Michael Balint, who first introduced me to the idea of the “drug, doctor.” Dr Balint argues that we prescribe ourselves, and choose the dose and frequency of intervention. We also have our own benefits, contraindications and side effects. New healthcare interventions can be seen the same way.
Let’s think about prescribing activity-based funding in general practice.
First, it requires a coding infrastructure.
Let’s think about Jane. There is no way a hospital coder will be able to deconstruct the 60-minute episode of care I provided yesterday on the basis of my current notes. In 60 minutes, I debriefed her recent blood tests, and adjusted her medication for diabetes. I monitored her blood pressure. I reminded her of the need to start mammograms, particularly given her daughter’s condition. I provided education around the NDIS, which might be helpful given her multiple needs. I wrote a letter to housing to assist in moving her up the queue now she is living with three children a few days a week in a two-bedroom flat. I listened. A lot.
Were any of those activities “unnecessary”? No. Is there evidence for all of them? Yes. I’m not sure that a coder would be able to provide a valid and reliable estimate of the value of that consultation, but let’s have a look at the benefit-to-cost ratio of having one.
This consultation cost the taxpayer around $140, if I include the incentive rise. If I added in a coder? I would have to pay the coder, and I suspect even in highly experienced hands, the coding would take at least 20 minutes, at about $50 per hour. It would also involve greater investment from me in my note-taking, as taking notes clinically, and taking notes suitable for coding are different. So, let’s add another 10 minutes to each consultation for that. I suspect the coder would have questions.
How much time was spent on each activity? How depressed was she? Is there a loading for being a carer with all its complexity? Do I code for one episode of care? What does that include? What was the listening about? What sort of listening? Is listening even a code? How much time would that take? Thirty minutes a day? Thirty minutes a session?
So, one side effect is a substantial increase in cost.
Another side effect is reduction of availability of my time. I already spend 20% of my time doing unpaid paperwork. If the coding load goes onto me, it reduces my availability. How many patients would be delayed or refused because I am busy resolving codes?
If coding is subsumed into the cost of the consultation, it increases total cost to the community. You could, of course, argue that Jane is an outlier, and therefore the model of activity-based funding is contraindicated, and most coding would be much “easier”.
You could, but it would be a disingenuous argument. Over half of our patients have multimorbidity and the average number of problems addressed in a consultation is at least two. Over two-thirds involve problems in multiple disease areas. Even in academic studies, at least a third of the activity was not captured by coding. We don’t know how many of our patients see us for psychological problems, because we only count those where we use mental health item numbers. Even then, it’s 13% of consultations. However, it’s probably closer to 66%.
So how do we reliably code Jane’s consultation? And even if we code it reliably, is coding a valid way of capturing value? Coding implies clusters of need. All patients with certain types of depression, all procedures of a certain type, all patients with an individual characteristic (eg poverty, age, disability) and all activities can be valued using an average costing tool.
Except there is a problem: Jane does not have diabetes PLUS depression PLUS poverty etc. Jane has an integrated mix of all of these things.
There is no evidence to suggest that adding Jane’s bits of experience together is indicative of her needs or the value of her care. To give you an example, given everything she is living with, do I expect the same outcomes for her diabetes than I would expect of a rich businessman with one disease and endless resources? If not, how much less should I expect? Or should I hold the GP working in western Sydney as accountable for maintaining an HbA1c in his patient population as I would for the GP in a boutique clinic in a well-heeled suburb on the North Shore?
Going back to Dr Balint, what would be the financial side effects of trying to introduce outcome-based funding if it involves applying economic modelling to all the variations of health and healthcare delivered by GPs? Would the GPs in disadvantaged communities with more complex patients and less capacity to bill privately be more disadvantaged? Would we remove even more GPs from western Sydney as practice becomes even less viable.
Even in “simple” consultations, coding is complex. One study coding childhood UTI, a relatively simple consultation scenario for patients with little multimorbidity, ended up being able to code 80% of consultations with 460 codes. Can you even imagine what the coding structure might look like for the whole GP cohort?
And even if we did manage to achieve some sort of coding nirvana, how do we define an “intervention”? Is a reminder about a mammogram an intervention? Is it different to explaining a mammogram, or persuading a person to have a mammogram or organising a mammogram?
What about measuring outcomes. Instead of relying on outdated fee for service, let’s look at paying for performance, which, it is implied, will nudge GPs to “achieve” better outcomes at a lower cost. Let’s say we follow the idea of patient-reported outcomes. One of the difficulties is that GP complexity makes it challenging to define sub-populations. Which outcome measurements are appropriate for Jane? Are they cumulative? Based on diagnosis? Based on socioeconomic disadvantage? If we use outcome tools for single diseases, are they even valid for someone like Jane? How would we know?
Many proponents of value-based healthcare would say that we should individualise outcomes. This would mean measuring “what matters to patients” rather than generalisable, standardised measure. But if we look at patient satisfaction mechanisms, we end up with individual outcome measurements with no population statistics for comparison. If everyone defines their own outcomes, we have non-comparable data. So how do we determine the value?
Fee for service doesn’t increase safety or quality of care
Who knows? We know that general practice is the safest, most efficient, and most effective part of the health system. Introducing activity-based funding requires a massive investment in infrastructure, complex data systems, and personnel. This will obviously benefit [LS2] those providing the infrastructure, complex data systems, and personnel.
Will it benefit patients? We don’t know. What we do know is that it drives GPs away from practice. That is a side-effect that is important and measurable.
Mr Manovel mentions the “well-known UK NHS Quality and Outcomes Framework” which he believes is “a stellar example of how paying GPs to document care outcomes has led to better funded GPs delivering better patient outcomes, resulting in fewer attendances at hospital”.
Maybe. Or maybe the hospitals just stop accepting as many patients, “dumping” them back in primary care. What we do know is that this shift has side-effects: plummeting GP morale, massive losses of GP workforce, and a system that is collapsing. For me, this is like providing world-class colonoscopies without counting the people dying from bowel cancer waiting to access care.
We GPs use 2% of the Medicare budget. Every year, we see around 85% of the Australian population. In all this talk of optimising our efficiency, we need to ask why our industry has been so relentlessly targeted for optimising, refining, reforming and restructuring. I do wonder if it’s a misread of our current effectiveness or efficiency, or its simply that we are a lucrative market for those who benefit from change.
Associate Professor Louise Stone is a working GP, and lectures in the social foundations of medicine in the ANU Medical School. Find her on Twitter @GPswampwarrior.