Punishing all self-referral is not the solution

If I were a respirologist, I would receive many referrals from family doctors asking me to determine whether patients with shortness of breath have asthma or chronic obstructive pulmonary disease. I would take a history from each patient and conduct a physical examination. Depending on the situation, I might occasionally order some blood tests or a chest x-ray. And then, for most such patients who were referred to me, I would conduct a test known as spirometry and interpret the results. I would then inform the patient about my diagnosis, prescribe some treatment if necessary, and write a letter to the referring physician.

Does anything sound wrong with this scenario? It shouldn’t. But when you think about how physicians are paid, a potential problem arises.

Specialists in Ontario are paid both for consults—seeing a patient at another physician’s request—and also for conducting tests. In the situation described above, a respirologist would receive $157 from OHIP for seeing the patient and dictating the letter back to the family physician, and between $29 and $38 for conducting and interpreting the spirometry test. Because the history and physical examination are limited and can even be misleading in patients with shortness of breath, guidelines recommend that physicians perform spirometry to diagnose both asthma and chronic obstructive pulmonary disease (COPD).

When the same physician asks for a test and then performs it herself, this is called self-referral. There is nothing intrinsically wrong with self-referral. In fact, self-referral is often the best way a physician can care for a particular patient. Consider this alternative to the scenario described above: the respirologist could send a letter back to the family physician recommending spirometry, and the family physician could then refer the patient back to the respirologist for the test. That would not be self-referral. But it would be inconvenient for everyone involved, especially the patient. It would also delay the diagnosis, potentially resulting in harm.

Recently the Ontario government announced that it is reducing OHIP fees by 50% in situations where self-referral has occurred. The government has good reason to be interested in this issue. There is plenty of evidence to support the notion that physicians who own diagnostic testing equipment are much more likely to order tests than physicians who do not. And although physicians don’t necessarily like to talk about this publicly, we all know that some doctors order too many tests. Some evidence suggests that about one-third of the tests we order are unnecessary. For example, most patients with low back pain don’t need an MRI. Many patients don’t need echocardiograms, spirometry or colonoscopies nearly as often as they are performed.

Some of us refer patients to other physicians for these tests. And some physicians conduct these tests ourselves. In other words, we engage in self-referral. There are many reasons for why physicians order too many tests—lack of knowledge, cultural norms, patient preference and fear of litigation are probably the most common explanations. Let’s be honest and admit that in some self-referral situations, greed is also a factor. Let’s also admit that the fees for some tests are higher than they ought to be. The combination of self-referral, overly generous fees and large for-profit independent health facilities is particularly toxic.

These are issues that the medical profession and government need to address urgently. Society has granted the medical profession the privilege of self-regulation, and it is probably fair to say we have not regulated self-referral as effectively as we should have done.

But punishing all self-referral is not the solution.

The Minister of Health and Long-Term Care, Deb Matthews, has stated that she is open to speaking with doctors about how to deal with the self-referral issue. The medical profession should welcome such an invitation. The public should get good value for what it spends on health care. And doctors should not feel like they are doing something wrong when they perform a test that is in their patient’s best interest.

The comments section is closed.

  • Joe Public says:

    Simple, just pay all medical physicians a yearly salary.

  • Steven Lewis says:

    The article and the comments reveal why we need to rethink the mode of payment, and not just fees for certain activities. It is almost impossible to have a rational and constructive conversation about who should be doing what, and for what reason, when each morsel of activity generates income. There is no way to micro-manage our way off the volume-driven treadmill that fee-for-service generates. Moral hazard is present every step of the way. It doesn’t have to be overt greet or deliberate income maximization – it is built into the DNA of how the system works, and it drives the culture of medicine. Worse, it makes governments (sometimes rightly) suspicious of doctors, doctors (sometimes rightly) suspicious of each other, and patients the unwitting accomplices to diagnostic and follow-up overkill. As Irfan Dhalla points out, up to a third of these procedures are unnecessary and some of them are harmful (exposure to radiation, the creation of anxiety, chasing false positives, etc.). Reining in self-referral is a fairly crude mechanism for addressing the more fundamental problem. The root cause is equating productivity with volumes. I see no way to overcome that problem within a fee-for-service system, and especially one like ours, with almost no oversight of or management of clinical practice variations. There is a reason why the highest performing systems in the world, from the VHA to Group Health to Kaiser, use alternate payment mechanisms. And it is instructive to note that Canadian doctors who have switched to non-FFS payment feel liberated, and I know of no example where one has switched and decided to go back.

    • Ryan Herriot says:

      Beautifully put.

    • BigDuke6 says:

      That’s hilarious.

      First off, Kaiser’s the system fined multiple times for dumping patients in wheelchairs in back alleys, ivs still in, when they cannot pay their bills. Hardly a paragon of healthcare virtue by my measure.

      Secondly, the simple truth is Ontario with its dismal financial situation, cannot afford to put all docs on salary or even capitation without serious decreases in physician salary. And if they decrease physician salary, you will see physicians moving… FFS docs are the most productive, simply put. They earn every penny. And don’t sneak out early on Friday afternoon

      Show me the studies that show less volume=higher quality. There are few problems in medicine that require an hour appointment, and many which need only a few minutes of the doctors time.

      And as to management of clinical practice variations, it’s the patient who ultimately chooses their treatment. You have a fundamental misunderstanding of the doctor’s role. Docs explain, offer alternatives, and the patient chooses whether or not they want surgery or more intense DM2 treatment or whatever. Management has no role there, and to judge docs by average A1C of their diabetics misses the primacy of the patient in this whole affair…

      Some diabetics do not want to go on meds to get A1C down, even after labourious and clear explanations are given. Some people do not want surgery for a sarcoma, even after the most graphic descriptions of outcomes.. Some hypertensives do not want more meds. Some obese people do not want to exercise and diet. Some smokers do not want to quit. All drawn from my own practice. Those are patient variations, not practice variations, which cannot be assessed in some manager’s office by someone who’s concerned with numbers in a spreadsheet cell.

      • Ryan Herriot says:

        You can make your point without the angry tone and the ad hominem statements.

    • Mark MacLeod says:

      Stephen – thank you for the thoughtful comment.

      As Irfan and I were back and forth on yesterday, the flavour du jour seems to be that FFS is bad and that APPs or AFPs are good. Good for patients, good for the system.

      I think we need a more nuanced approach. First I would say that there are persons for who getting doctors off of FFS or reducing unnecessary work have become code phrases for paying doctors less for the work they already do or reducing service volumes, whether that work be necessary or unnecessary. If those who talk about reducing FFS are using proxy arguments I would like them to do so up front. Otherwise, the arguments are disingenuous and not helpful.

      Secondly, I don’t know where the idea of unnecessary volumes has come from and I’m open to understanding it better. Do we really know that X percent of work done is “unnecessary” and if it is, what are the reasons? Duplication? Inadequate EMR system? Patient reasons? System delays? Doctor reasons? Blaming physicians seems unhelpful where the problem is likely much more complex.

      Thirdly, experience tells us that FFS is a very good volume driver and I don’t think we are looking to reduce volumes – or are we? The biggest dividend in reducing what a physician does is not in her/his payment, but what the reduction in service means for system cost. Salary based models have had large impacts on patient throughput and I don’t think that is a good thing. Reductions of up to 30 to 50 percent have been noted, but auspiciously very few written reports – these were in large field circumstances where before and after were clearly different or productivity compared to peers was easy to compare. Those reductions were not in unnecessary work, but very necessary work that shifted elsewhere. If you pay someone a salary without accountability, metrics, followup, penalty etc, you basically tell them they are being paid not to work. Why push to do that extra case if it makes not difference in income? Why book 30 new patients in a day when booking 20 means the same outcome Asking doctors to behave as if they are not human is naive. If an alternative to FFS requires a big administration to ensure volumes, what success has been gained? I bet those who switched to a different model didn’t want to go back -not paying for billing alone would be a bonus, but I suspect that the pressure to work for an income was also gone. Nice work if you can get it.

      Thirdly, we need different models for different kinds of work. Chronic disease management may well be better served with a non FFS model with whatever accountability mechanisms are needed to guarantee access, quality etc. Episodic care like mine in orthopaedic surgery is likely best served with FFS as FFS drives volumes. All models need some form of P4P.

      Simply saying that FFS is bad is unhelpful. All payment models have distortions and opportunities for gaming. We likely need multiple models that fit the care being delivered.

      My 2 p for a Tuesday morning

      • Steven Lewis says:

        Thanks for your insights Mark. There are a number of important questions embedded in your comments. Do we want to reduce volumes? I would say yes – total system volumes for sure, and perhaps physician encounter volumes. Southcentral Foundation in Alaska, which serves an indigenous population, has reduced not only referrals to specialists and hospitalizations, but also primary care visits by 20%. They explicitly say to their doctors, if you're working harder and your patients are devoted to you, you're missing the point. The goal is health and self-management. Do less because you're not needed. That's a major cultural transformation that conflicts with the tenor of our times, but I think it's undeniable that there is a huge amount of avoidable utilization (along with some worrisome underutilization), and if one accepts that premise, I cannot find a logic that would square what we ought to do with a FFS system. There is also an important distinction between the reasonable goal of a target income and how one earns it. I have no problem with doctors being well-paid. What I challenge is the notion that getting paid for piecework is the best way to achieve that objective. I would like the payment issue to be irrelevant to the practice of medicine – i.e., figure out a pay scale, renegotiate it annually or at regular intervals like every other occupation, and get on with pursuing excellence. I don't want doctors' incomes to be affected by exercising creativity, trying new things, spending varying amounts of time with different patients. And I certainly don't want to create a system that cries out for gaming. As you point out, there are no perfect payment systems, but there are outstanding workplace and system cultures. Interestingly, where the cultures are strongest and the achievements are greatest, no one seems to be too worried about the money. Group Health Cooperative pays middling salaries to its doctors, and they get 17 applications for every opening. So let me try out this hypotheses on you: the only reason we continue to discuss FFS as if it made sense is path dependency. It's what we have, many are comfortable with it, it's the devil we know. That doesn't make it a good fit with quality improvement, focusing on complex patients, effective patient self-management, prudential use of tests, etc. Its virtue is familiarity, plus the ability to generate more income by doing a lot of routine things efficiently. Knowing what we know and observing what we observe, I doubt that any thoughtful system design exercise would choose it as the payment mode. Where we possibly might genuinely disagree is on the matter of trust and professional obligation. You posit a worst-case scenario where doctors are paid a salary without accountability metrics, etc. Even in that unwise circumstance, I cling to the notion that very few would neglect their patients as a result. You pose the question, "Why push to do that extra case if it makes not difference in income? Why book 30 new patients in a day when booking 20 means the same outcome?" The answer is, for the same reason that good and conscientious people work hard and sometimes go the extra mile without getting paid extra. It is part of their identity and a recognition that the greatest rewards – particularly in a field like health care – come from doing good works for others. I'm not naive – of course systems are gameable, of course humans vary in their devotion to duty, no organization achieves a perfectly balanced workload, some people care more about money than others. So we have a choice: design a payment system that assumes the worst about what motivates people and create an elaborate policing system to discipline those who fall prey to its temptations, or design one that assumes they are caring professionals who want reasonable incomes and create accountability tied to real performance, not crude notions like volume. The high-performing systems seem invariably to take the latter course, and I don't think it's a mystery why they do. You also raise the question of whether FFS might be suitable for some activities and not others. In theory I would say it's less problematic where those getting FFS don't generate their own demand. But it's still a problem – the radiologist who successfully teaches GPs not to order needless CT scans will be cutting her income. If you know exactly what you want to buy, and know how not to buy what you don't want, the hazards of FFS diminish. But it still a problem. Apparently newly certified orthopedic surgeons are having trouble getting jobs. Maybe there is a role for some of them in the medical management of patients, or in secondary prevention. FFS systems tend not to be adaptable enough, fast enough to accommodate such opportunities.

      • Mark MacLeod says:

        Thanks Stephen.

        I have a sense (might be wrong) that the changes you would like are more applicable to primary care (but again I could be wrong). I’ve put the offer out before to anyone – come work with me for a week and tell me which services I provide are not necessary – and bey extension which patients are not valuable enough for the system we apparently envision.

        I honestly don’t know how we get to the idea of reduced volumes when we understand the unmet need is still so vast. The Wait 1 project in Ontario is demonstrating the hidden burden in surgical specialties alone – hidden outside of the previous wait times metrics.

        On the idea of professionalism and trust. I can identify 3 examples of significant scale in recent past in this province where productivity dropped dramatically with a switch in payment model. I’m not disappointed in that, in fact I find it affirming that doctors are human and just as subject to forces as the rest of the population. It’s a topic for another day but it is time we allowed doctors down off the pedestal – a pedestal that the public and the profession have created. It’s a false constraint and ruinously imparts to the profession neither they or the public are capable of meeting. Some do, just as some in any profession will. But like most, given the option of working hard or working less hard for the same reward, . . . .. Thusly, let’s not make assumptions about nature or character that are neither true, nor achievable. As an aside, I would wager that physicians willingness to work hard within a system is correlated with the confidence that they have in and the contentment they find within it. I think that goes a long ways to explaining the examples often used, such as GHC, Mayo, Intermountain etc etc. Like Sidney Crosby, those individuals make a choice about the place they want to play balanced with monetary reward. Move a system away from one with inherent reward, satisfaction etc, and people with value other rewards commensurately more.

        I do agree with the comment about FFS not being necessarily a good match with other values if the appropriate checks and balances are not in place. Fair comment

        At the end, I don’t believe that we have a problem of the grass being cut too often. We do have a problem of large areas of the lawn being uncut and we are trying to do the rest with a 50 year old lawnmower.


      • Ryan Herriot says:

        Mark, what do think about the idea of moving surgeons to a form of capitation (not salary) where you are essentially paid for the difficulty of the case (i.e., how long it takes), and not per case. Or perhaps moving to how anesthetists are paid, with some for the procedure and some for time?

      • Mark MacLeod says:


        Generally I think that would be a backward step.

        What is really needed is a good relative fee schedule that is based on average time.

        First of all, I don’t buy the complexity argument. If you are appropriately trained then what you do isn’t complex. For example, if you asked me to do a delivery everyone would be at risk. Mom, baby, me. However put me in a trauma room with a hypotensive trauma patient with a broken pelvis, mangled extremities – that isn’t complex – it is what I trained to do. “Complexity” has been used as a proxy argument to justify someone being paid more than another (what I do is more complex). If a procedure takes less time than it takes me to shower and shave in the morning, it is neither complex nor difficult. Period.

        I do believe that average time should be the basis for the fee schedule – and that 8 different surgeons from 8 different specialties who are all working at the same intensity (ie one isn’t a slacker() should end up with the same remuneration. So time is the leveler and fees should be based on average time. That way the person who has figured out how to be more efficient or is simply better will see a reward compared to the slow surgeon (and they are out there) or the inefficient surgeon.

        I would strongly resist any move to go to an anaesthesia model – unless it could be constructed in a way that it would penalize those who are ineffective.

        FFS intrinsically does one thing well – it encourages volumes, more than any other system. And that isn’t always or even of necessity bad.

  • Shelagh McRae says:

    Thanks for this piece. Very relevant already -for example a new form appeared in my in-basket this week requesting my signature approving the “suggested follow up plan”.

    Our patient had had complicated surgery (our practice has had only a handful of people undergo this procedure in the last 30 years) and the surgeon was recommending 6 monthly imaging at his facility to monitor the area. The plan seemed reasonable to me but:
    1) I’m a family doctor not a specialist surgeon,
    2) I wasn’t present at the surgery (and reading an OR report is no substitute for being there to see how it went and know what might go wrong in the future),
    3) I don’t know what (if any) the current guidelines are for post op monitoring for this condition and
    4) it would take a great deal of (unpaid) study and consultation for me to even consider challenging his recommendation.

    So I trusted my colleague that this “self-referral” was in the best interests of the patient and signed the form. I do hope I don’t have to worry about being billed for the cost of the procedure if some government review decides that the test (or interval) was inappropriate. I do worry that patient care is going to be delayed and that more and more of my time and that of my staff is going to be spent processing similar forms.

  • Will power says:

    Radiologists should not be able to bill fo fluoro used by others. What a rip off. The hospital owns the equip and pays the tech. Radiation safety is the responsibility of all.

  • RAD says:

    As a radiologist, I’ve long understood the attitudes of physicians who interpret basic x rays and make clinical decisions, and sit back and think “Now those radiologists are going to read it after the patient is got and get PAID”. I’ve heard it over and over.

    I will admit I particularly dislike reading fracture clinic films, it’s pretty mind numbing. However, having said that, I also firmly believe that if you are going to argue that perhaps the reimbursement should be taken away from the radiologist, then the clinician must sit down, dictate the report, sign the name and become responsible for the image interpretation as a legal document.

    The same with ER docs that want to use US – they should SAVE the images, and provide an interpretation of them- I’ve had some frightening converstations regarding ER docs interpretation of US – including bladder carcinoma which was really the prostate, and free fluid which was really just content within a normal boweI loop.

    I also read piles of ER films that ER docs, orthopods etc have read and interpreted on their own, and they make WRONG interpretations. Missing a subtle fracture, missing a cancer (it’s happened more than you think!!), calling a pneumonia when there isn’t one. Now, these may have little immediate life threatening consequences, but when I’m doing this level of interpretation, I do see myself as a quality control check, ensuring that the right diagnosis was given to the patient. With new time implementation for turn over of studies, the delays have greatly decreased at my institution, such that we can have real time report generation, and we always welcome calls for immediate interpretation in more advanced cases. But remember as well, that we just don’t read plain films in fracture clinics….so I think the Brownian motion comment regarding radiologists demonstrates a lack of understanding of what we do.

    • Mark MacLeod says:

      Oh it’s not just fracture clinic films – it’s the tibia fracture that I’ve seen, seen the films, fixed and so on. The “total hip has been implanted” report and so on. I don’t know exactly how to but the role needs to be more clearly defined, duplication of needless activity circumvented etc. It’s a challenge.

      Similarly I really need interpretation of many things and that is where the real system value lies. I’m the first to ask for it and rely on it. So I understand the complexity.

  • Mark MacLeod says:

    The emphasis has gone to “doctors being greedy by self referral” and we have lost focus on the need for clinical pathways (not guidelines) for investigations and treatment. Such pathways would mean MRI for low back pain and nonmechanical knee pain could not be ordered. The counterpart to this would be to look frequently at the negative findings rate to be sure that today’s low back pain patient didn’t all of a sudden have “leg pain” as a way to circumvent the guideline – wide variations in negative cath rates in the country suggest something is up in terms of how cardiac catheterization is used.

    In the same line would be echos in preop patients, arthroscopy for knee pain in patients greater than 50 without mechanical symptoms . . . . the list is likely endless. We also need to understand clearly what treatments are not effective and should no longer be part of the care pathway.

    But back to the self referral piece – why not fund the episode of care in such a way that the testing part cannot be used as part of the income stream for the physician? Pay a reasonable T fee that covered the operational cost only and pay no professional fee for the interpretation of the test.

    It’s complicated, some groups make all of their income from test interpretation but we have to come to grips with how they are important or not in various parts of medicine. I don’t agree that we should be paying for interpretations of x-rays where I as the orthopod have already looked at the x-ray, made a judgement, implemented an action, long before the radiologist even looks at the film. Its a simple thing to change torte law to make me responsible only for the orthopedic interests – because someone will trot out the “but Dr. MacLeod might not recognize . . . . . . in the x- system”. I simply don’t beleive we should be paying for brownian movement in the medical system – we need a different mechanism.

    Self referral should continue but the funding should be such that the self referral generates no additional income. It’s a doable fix.


Irfan Dhalla


Irfan is a Staff Physician in the of Department of Medicine at St. Michael’s Hospital and Vice President, Physician Quality and Director, Care Experience Institute at Unity Health Toronto. Irfan also continues to practice general internal medicine at St. Michael’s Hospital.

Republish this article

Republish this article on your website under the creative commons licence.

Learn more