Monthly Archives: July 2017

Cheaper PPI options: An EMR QI exercise

Back in December 2016, I was doing some work around proton pump inhibitor (PPI) costs in Ontario and did some digging into my EMR data to see what savings could be found.

Here was my Twitter thread about the topic.

In short, I identified 414 of my patients on a PPI, 314 (76%) of which were on rabeprazole or pantoprazole, the two cheapest options. I calculated the potential monetary savings by switching every patient in my practice who is on a more expensive PPI (esomeprazole, lansoprazole, omeprazole, dexlansoprazole) to generic rabeprazole; projected tally would be $22,340 over the course of a year. Not an insignificant sum.

It’s one thing to project a cost savings. But would it actually work?

I embarked on a very simple intervention in January 2017. Using my EMR (Telus PS Suite), I created a reminder that would place the note “Consider switch to cheaper PPI” on the chart of every patient on esomeprazole, lansoprazole, omeprazole, dexlansoprazole, or pantoprazole.


That reminder would appear on their chart, and would prompt a discussion at their next appointment about a potential trial of rabeprazole.


(I am intentionally omitting the issue of de-prescribing from this analysis, as de-pescribing PPIs has been another important undertaking in my practice, yet this analysis was strictly looking at looking at cost-saving from cheaper PPIs. The above reminder did prompt many instances of de-prescribing, but these were not as easily quantifiable retrospectively. In addition, I bristle at the idea of openly quantifying and applauding the cost savings from de-prescribing, as it would rightfully raise the question in patients’ eyes whether we are de-prescribing out of concern for their health, or simply to save money. This analysis looks at cost-savings from switching from one PPI to another, the goal of which is to produce no clinical negative change).

Of the patients who were engaged in a discussion, many were found not to be appropriate for a rabeprazole trial, either because of intolerance or lack of efficacy from previous trials of rabeprazole. Other patients were not open to a trial because of lack of interest in “rocking the boat” or because of an upcoming trip and concern about impacting travel insurance.


28 patients (some private pay, some publicly-funded) who required ongoing PPI therapy (due to ongoing indication or due to previous failure with PPI wean) agreed to a rotation to rabeprazole as a trial. 7 patients described a clinical worsening of symptoms (most within the first two weeks), and requested to be put back onto their initial PPI (5 patients on lansoprazole, 1 on pantoprazole, and 1 on omeprazole).  21 patients saw no clinical change, and those 21 have been found to be stable at the 3-month mark post-rotation (11 on lansoprazole, 4 on omeprazole, 4 on esomeprazole, and 2 on pantoprazole).

The cost savings were as follows (using rabeprazole 20mg tablet, $0.24/pill):

  • Esomeprazole – $1.86 per pill ($1.62 savings/d, $591.30/yr per patient)
  • Omeprazole – $0.41 per pill ($0.14 savings/d, $51.10/yr per patient)
  • Lansoprazole – $0.50 per pill ($0.26 savings/d, $94.90/yr per patient)
  • Pantoprazole – $0.30 per pill ($0.06 savings/d, $21.90/yr per patient)

Total yearly savings from these 21 patients now stable on rabeprazole: $3657.30

A few things to take away from this:

  1. This was a ridiculously simple intervention, and took me virtually no time to create the reminder or to have those discussions with patients.
  2. This should be a wake-up call to the Ministry of Health on the importance of working with physicians on finding efficiencies in our system. Physicians know where these efficiencies are, so allow innovation and creativity, reward it appropriately, and we’ll find the money.
  3. It is critical that the first script for the rabeprazole be of short duration (I prescribed two weeks), otherwise 3 months of wasted pills for the 25% failure rate would have erased some of the cost savings.
  4. Most practices will likely see greater savings than I achieved in my practice, as I have been consciously working to prescribe exclusively rabeprazole for some time now. Even if you assume my prescribing rate to be on par with the Ontario average (2200 patients), extrapolating this out to Ontario’s population of 13.6 million, this simple intervention would amount to a provincial total of $22.6 million.
  5. Note that most of the savings were found in switching patients from esomeprazole. Please don’t use it unless you are absolutely handcuffed. It is insanely expensive compared to others in the class.

All of us have these opportunities in our practices for small interventions that we may actually find some element of satisfaction in evaluating, especially with the power of a well-functioning EMR behind us.

ACE-inhibitors/ARBs in diabetics: Evidence doesn’t support the CDA guidelines

As part of my work on a provincial committee, the issue came up recently regarding a quality metric evaluating proportion of all diabetic patients aged >55  who are being prescribed an ACE-inhibitor or ARB. When it was pointed out that the evidence supporting this specific metric is weak, the response was that this recommendation is part of the Canadian Diabetes Association Clinical Practice Guidelines, and that for a committee to go against an established guideline would be difficult.

The full CDA Clinical Practice Guidelines were last updated in 2013 (partial updates in 2015, February 2016, and November 2016), including the section specific to vascular protection in people with diabetes. Here is the recommendation around ACE-inhibitors/ARBs:


The efficacy for ACE-inhibitor/ARB use in established macrovascular disease was established back in 2000 with the HOPE trial. No argument there. Same with patients with diabetes and kidney disease (including microalbuminuria) from the RENAAL trial showing losartan improved proteinuria and renal outcomes.

I want to focus on the first half of point #2, where they recommend ACE-inhibitors for all patients 55 or older “with an additional risk factor”. From the HOPE definition, these risk factors would include hypertension, cigarette smoking, elevated total cholesterol levels, low HDL levels, or microalbuminuria.

Should ACE-inhibitors be the preferred treatment for hypertensive diabetics without cardiovascular or kidney disease? Not quite. The 2017 CHEP Guidelines reviewed the evidence and found that thiazide/thiazide-like diuretics and dihydropyridine CCBs would also be appropriate choices in this population (Section XII, Point #3). American guidelines from JNC 8 came to a similar conclusion (Recommendation #6). Here’s a systematic review from the BMJ from 2016 on the topic.

The most relevant question to me is, should patients with diabetes with no history of cardiovascular disease, no history of hypertension, and no end-organ damage, be prescribed an ACE-inhibitor? Simply on the basis of having hyperlipidemia or cigarette smoking?

The HOPE study population was a mix of patients with established cardiovascular disease and those who where 55 or older with one of the above mentioned risk factors. While a benefit was found to ACE-inhibition in the study population as a whole, here were the confidence intervals for those with and without cardiovascular disease:


As you can see, it is very difficult to make any conclusions about the “no cardiovascular disease” group. Point estimate around 0.8, but very wide confidence intervals. One could argue that the point estimate could be hypothesis-generating, but certainly no more than that in terms of guiding whether these patients should be all prescribed ACE-inhibitor/ARB therapy.

The guidelines also cite a “recent meta-analysis indicates that ACE inhibitors and ARBs reduce CVD events in normotensive individuals with and without diabetes” and “accordingly, the use of ACE inhibitors or ARBs for vascular protection with persons with diabetes ≥55 years or with any evidence of organ damage is recommended, even in the absence of hypertension”.

But hold on a second. What did this meta-analysis show? Or should I ask, what studies did this meta-analysis look at? 13 studies in total: 3 studies including only those with heart failure (SOLVD, SAVE, and TRACE), the RENAAL trial (patients with nephropathy), PROGRESS (patients with stroke or TIA), CAMELOT (those with angiographically documented CAD and elevated DBP), EUROPA (stable CAD patients), DIABHYCAR (patients with microalbuminuria), DREAM (dysglycemia, no frank diabetes), TRANSCEND (established atherosclerotic disease or end-organ damage), PROFESS (history of ischemic stroke), ACTIVE-I (history of atrial fibrillation), and the above-mentioned HOPE trial.

So the only study in this meta-analysis that included the patients from our relevant clinical question (diabetes with only hyperlipidemia or smoking as a risk factor) is the HOPE trial, which we already showed we cannot look to for a definite answer on this. None of the other trials cited by the CDA Guidelines looked at ACE-inhibitor/ARB use in our subset of patients.

Why is this relevant? In my practice alone, of 220 patients with diabetes, I have 40 patients who either smoker or have hyperlipidemia, and who have no history of cardiovascular disease, no microalbuminuria, and are not on an ACE-inhibitor/ARB for hypertension. The current guidelines suggest that all 40 of these patients should be on treatment, despite the evidence not supporting this claim. This is not a minor clinical issue. And to compound this problem, we have provincial and local organizations uneasy about contradicting these guidelines, and feeding data to physicians with the intent of altering prescribing rates.

I would suggest that the Canadian Diabetes Association Clinical Practice Guidelines Expert Committee re-evaluate their stance on the issue of ACE-inhibitor/ARB use in patients 55 and older with one additional risk factor (without macrovascular disease or end-organ damage) given the paucity of evidence supporting their current recommendation.