The problem with numbers, the curse of intermediate end-points

The curse of treating chronic kidney disease is that one is always treating patients to the numbers:

  • Blood pressure. I need to get my patients below 130/80
  • Cholesterol. I need to get their LDL below 100
  • Metabolic bone disease. I need to keep their PTH
    • KDOQI 
      • Stage 3: 35-70
      • Stage 4: 70-110
      • Stage 5: 150-300
    • KDIGO
      • In patients with CKD stages 3–5 not on dialysis, in whom serum PTH is progressively rising and remains persistently above the upper limit of normal for the assay despite correction of modifiable factors, treatment with calcitriol or vitamin D analogs is suggested. (hey KDIGO, thanks for the guideline)
  • Diabetes. I need to keep their Hgb A1c less than 7
  • Anemia. I need to keep their hemoglobin
But these numbers are all intermediate, and from a patient perspective, pretty abstract. Patients don’t get PTH angina. Targetting the numbers is a way to shift the odds toward better patient outcomes, to load the dice in the patient’s favor. However we cannot allow the numbers to substitute for the real goals of care. I don’t really care about your blood pressure, I just want to prevent the heart failure, dementia, kidney failure, stroke and erectile dysfunction that result from the high blood pressure. If you give me a pill that magically improves the blood pressure but doesn’t avoid those end-points I’m not interested.
But as the number game has become a larger part of medicine we are getting medications that are pursued and approved only for their ability to fix the numbers. Some have been super successful, statins have repeatedly and reliably shown their ability to reduce events in lockstep with reducing the cholesterol. Lately however, it is feeling like success of the statins to reduce LDL and also reduce cardiovascular events maybe more the exception than the rule.
The recent experience with ESAs and hemoglobin have been beat to death in the nephrology community. See this post for a deep dive. The core issue, is that low hemoglobins are bad for patients, but using ESAs to improve the hemoglobin does not mitigate the risk. And not only does it not mitigate the risk, it appears that the current agents bring with them novel arterial and venous thrombotic risks.
The experience with A1c seems to be playing out using a similar script. Glitazones were approved based on their ability to reduce blood sugars. They effectively lower blood sugar but Rosiglitazone increased the risk of cardiovascular death by 64% and was associated with increased composite outcome of stroke, heart failure and total mortality compared to pioglitazone.
And June 9th pioglitazone was pulled from the shelves in France for increased risk of bladder cancer. A position validated by the FDA on June 15th.
This comes on the heals of three studies in 2008 and 2009 that question the notion of very tight (less than 7%) hemoglobin a1c targets to improve patient outcomes.
In cardiology, following the stunning success of statins and LDL we have a string of failures, Ezetimibe (Zetia/Vytorin) for LDL and niacin/torcetripib for HDL
I often feel the only reason we still treat PTH is that no one has done the study to show that it helps and when we get around to that trial, I’m looking at you Abbott, we will find that it too, has been a waste of money and attention.

3 Replies to “The problem with numbers, the curse of intermediate end-points”

  1. Ah,yes….numbers! Everyone seems to love numbers. Numbers are so easy, so quantifiable. The more of them we measure and the more complex they are, the better the care. Oh, really?

    There's only one slight problem. Numbers are medical-centric, not patient-centric. Numbers are just that–numbers! They often tell us very little about the actual patient experience.

    To divert a bit: Consider the New York Heart Association Functional Classification for heart failure as an example. This simple 1-4 scale is based on how the patient feels, what he is able to do. It's not about ejection fraction. Numbers take a back seat. Would this scale be the sole criteria for diagnosing HF? Of course not! However, a high level of functioning translates into quality of life and a patient's ability to pursue desired life goals such as employment, hobbies, family involvement or travel.

    Maybe the kidney community should take a lesson from cardiology. Why not have a simple scale, similar to the numeric pain scale? "On a scale of 0-10, 0 being 'life sucks and I can't do much of anything I want to do' and 10 being 'I feel terrific, like I can conquer the world', where are you on the scale?" Then a follow-up question would ask, "What is an acceptable level on the scale, for you personally?" In a patient centered model of kidney care, the health team would work with the patient and care partner to address ways to reach the goal on the scale. That may be a somewhat tongue-in-cheek verbiage but the idea is valid. It's about how the patient feels and functions.

    During my late husband's 25 years on HHD, his nephrologist's first questions were invariably "How are you feeling?" and "How's the golf game?" I daresay my husband's answers to those questions told his physician as much about his current condition as the lab results.

    I agree that numbers are only intermediate measurements with the primary endpoint being a patient who can honestly say that, despite CKD/ESRD "I feel good. I can do what I want to do!"

    A number is meaningless unless it is relative to the real world. In the recent heat wave that gripped most of the country, we could sit comfortably in our air conditioned homes thinking that "yep, 115 degree heat index is hot" but going outside and being engulfed by a stifling oppressive heat makes that temperature real! Probably not the best comparison, but another example of relying too heavily on numbers only.

    In my opinion, stop looking soley at numbers and start looking–and really communicating with–the PATIENT.

  2. Interesting post and well written as usual. I fear however, that more and more of our compensation and reputation will be based on meeting these "numbers". These will be set by those who aren't taking the patient into consideration. I am concerned about how the numbers will control/influence practice in the future.

Comments are closed.