The 'Good Dialysis Index'

Dear HDC members and friends

At the Annual Dialysis Conference in Seattle last week, I presented a proposal I have been thinking about for some time … I have called it the ‘Good Dialysis Index’ - or GDI.

In retrospect, perhaps the ‘G’ should have been for Geelong and it should rightly be renamed the ‘Geelong Dialysis Index’.

I’ll let that one hang …

The full slide presentation given at the ADC cannot be reproduced here, though it began with the reasons why I distrust and dislike Kt/V as the primary assessment basis for dialysis ‘adequacy’ … by the way, adequacy is also a word I abhor.

From this dislike and distrust has grown a more simplistic and … yes … a more naive way of assessing the ‘goodness’ (or otherwise) of dialysis at the clinic visit - the GDI

Most importantly, the method I proposed - ‘The Good Dialysis Index’ - attempts to be more patient focused than Kt/V. Kt/V converts you, the patient, into a mathematical model … and a bad mathematical model at that.

Personally, I do not see dialysis patients as merely a complex and inaccurate jumble of bad mathematics. That approach, to me, does not do you justice, as a human being or as an individual. You deserve better.

I do not suggest that the GDI is the ‘be-all’ and ‘end-all’. The GDI is but a step in what I hope is the right direction … at least I think it is.

It has not been tested or validated. It is just a concoction, out of my personal clinical experience, of what I think matters. The questions I have included are, to my thinking, the chief things that matter - both to my patients and to me as their renal physician - when deciding if the dialysis given and received is achieving worthwhile goals. These goal are in four major target areas … patient goals, process goals, biochemical goals and outcome goals.

The sheet is designed to form part of the patient record. It can be completed in ~2-3 minutes and can be serially rated to determine progress. There is a column in which targets to overcome any ‘scoring deficits’ can be noted … focusing attention at the next clinic visit to see if by addressing the deficit, improvement has resulted.

I hope the GDI will ultimately be tested in large patient groups to assess its worth. Negotiations are underway to try to achieve this.

Until then, it remains just one far-distant renal physicians’ dreams of a better assessment tool.

Some patients (Bill Peckham, I will use your term, ‘dialyzors’) who were at the ADC have been ‘hot to trot’ to use it. I am not against this as long as it is used with the clear understanding of both the user and his/her renal physician that it has not been validated and is only ‘experimental’.

Those interested in the GDI - and/or any potential users of it - will find it at my website:

Please visit and carefully read and understand the introduction section at the website. This is important as it couches the intent of the GDI in the appropriate terms.

A printable version of the GDI is available for you at the website.

John Agar

Interesting concept that I would like to see what your experience is with your patients. While Kt/V is not completely useless, it has been used incorrectly as the only judge of good dialysis. A simple index is a good idea if it can be shown to correlate with morbidity and mortality.

I thought your presentation was very thought provoking and well received. At least I didn’t see any Kt/V-police raising issues. I’ve long thought that Kt/V is a terrible measure of “adequacy”, a term I abhor too. We should not be looking at “good enough”, but what’s the best possible treatment for the dialyzor. I haven’t thought enough of the GDI model to say whether some categories should have more weight than another, or as one person proffered that some items should have more flexibility in how answered. But what I do know is that both qualitative and quantitative data should be considered.

The fact that phosphorus, potassium, PTH and others are not currently considered shows a complete lack of understanding the importance how these affect the dialyzor’s over all health. It wasn’t a low Kt/V which caused my heart attack after a treatment. One should be able to achieve almost normal numbers without additional meds. With nocturnal, all of my indicators are “normal”, without the use of meds. There kept in check with longer and more frequent dialysis. That, to me, is the sign of good dialysis. I know I would not have survived this long without going home. Once again a sign of good dialysis. I’m active in my pursuits and don’t crash and burn after treatments. I’m able to stay up all day and do my work. I’ve been rehabilitated to a state I can function seven days per week. It’s not just because my Kt/V is better. The renal community has fallen in the Kt/V adequacy trap like it has with Conventional Hemodialysis. The GDI may begin a new shift in thinking. We ought to give it a chance.

Dear Richard

Thanks for your support for the GDI.

It will be interesting to see if a range of dialyzors … home full nocturnal, home short daily, conventional facility etc … apply it to their own data and situation and then see what ‘scores’ emerge.

John Agar

Hello Dr. Agar,

I have been tasked with ‘tracking outcomes’ at NephroPlus, the chain of dialysis centers I joined recently. I would like to use the GDI for this purpose. I had a few question on this though:

  1. We don’t generally do all the lab tests that are part of the index every month. We do most of them every quarter and some even every six months. In India, as you are aware, most patients pay out of pocket. So, there is a lot of resistance to getting more tests done. Any suggestion you have on how we can work around this?

  2. What is a two month Kt/V? Is it the average in the last two months?


Dear Kamal

The 1st thing to say about re the Good Dialysis Index is that although it is designed to be more inclusive of a number of the things about the dialysis patient that I think are important than, say, the current Kt/V-centric approach - and this especially means the inclusion of the actual patient within the index - it has …

(1) simply grown from my own experience

and …

(2) has not, to date, been validated - although we have been collecting some local data which suggests it is a helpful ‘holistic’ data set by which to follow the quality of the dialysis treatment in individual patients and across patient groups and dialysis modalities - yet to be published.

That said, it should also be ‘adaptable’ to differing conditions and situations. While we measure, in our program, the parameters included in the index each 6 weeks, there is no reason why this same inter-test or recording interval needs to be the frequency adopted for the measuring set by all services that may seek to use it. A quarterly set - even a 6 month set - should be just fine. Further, it should be possible to create a sub-set of more frequently measured parameters (say 2nd or 3rd monthly) with a lower total score and a ‘full set’ each 4 to 6 months.

It is also not ‘set in stone’ as the best questions to ask. I hope it might be an ‘index in progress’, with better versions, tighter questions, some of the current set - perhaps in time - excluded, with others replacing them. I hope for it to be a living, changing index - as dialysis lives and changes: not set, immutable and unchanging, despite movements in treatment goals and patient aspirations.

Our local assessments have, for example, challenged the inclusion of Q’s 7 and 8 (? too NHD biased … though my personal bias to NHD wanted them there!), and Q 20 … for a similar reason. Even with them 'gone, NHD scores well above conventional dialysis markers.

I have a nice graph of the comparative scores in our patients … were I more skilled at working out how to copy it here, I would … but, I will be showing it briefly at the ADC in Phoenix next week.

It is the GDI ‘concept’ that is key. The questions themselves can be - maybe should be - molded to suit local circumstance.

By putting forward the GDI, my main purpose was to make people think more widely than Kt/V as a measure of the ‘goodness of the process’.

Good dialysis is more than a process … good dialysis should have, as it’s overriding outcome, the patient as its’ core and as its’ ever-present centre-piece.

I do not believe Kt/V does that. I think the GDI at least tries to.

Your local team, cognisant of local conditions, measuring cycles and available data, could work to adapt the GDI such that it makes sense, for you, for Indian dialysis practices.

In short, it doesn’t have be a rigid questionnaire. It can be adapted but always, I think, with the same broad mix of patient-derived, lab-derived, process-derived and treatment-derived questions.

Have a go at adapting it. Maybe post the outcomes of that effort. Thereby, allow others to have an input into the design.

The GDI has to be, in the end, a living index. Kt/V is not. Kt/V is pure (and crazy) mathematics alone … and mathematics is not life.

As a living index, it must also mean an index that can grow, subtly change and adapt to local conditions.

Make it live.

Sure. I totally understand Dr Agar. We will have a go at it and see what we come up with.

Thanks again!


Hello Dr. Agar,

Which formula do you use to calculate the Kt/V, which is a part of the GDI? I have found many different formulae on the internet for this.


Dear Kamal

As you know, I am not a fan of Kt/V - either as a concept or as a measure of dialysis ‘adequacy’.

In Australia, few of us (if any) measure Kt/V at all. I know of no Australian dialysis unit that preferentially or routinely uses it as it’s ‘measure of adequacy’ beyond, perhaps, paying it lip-service for research and publication. Here, in Australia, we routinely use the PRU (the percentage reduction in urea) … which is the same as the URR (the urea reduction ratio), but converted to a percentage. Indeed, the PRU is the only ‘adequacy’ measure recorded by our national ANZDATA registry - though it is possible to extract, for research purposes, an equilbrated Kt/V from the national data-base. However, the measurement of Kt/V has not been formalised nor embraced in Australia. Simply put - we just don’t use it.

You will notice the the data-point used in my GDI proposal is written as "Kt/V =or>1.3 or PRU =or>70%". While I included both as an either/or option in the GDI set of 20 questions, Australian and New Zealand units don’t bother to calculate Kt/V but, rather, use only the PRU but I included the Kt/V option (largely, I suspect, potential US users) for those who can’t let go of Kt/V and wish to use it as their measuring point for answering Q9 in the original version of the GDI.

If Kt/V is required by some, then I suggest using the ready calculator for Kt/V as it has been developed by John Daugirdas. There are a number of these calculator sites and all use the same formula. Samples can be found at the HDCN site @ or at the DaVita site @

As a non-beliver on Kt/V … the whole point of proposing (something like) the GDI is that I believe that good dialysis is far, far, far more than just a mathematical formula.

I don’t really beleive it matters a jot which Kt/V equation is used: simple … standardised … one pool … two pool … equilibrated … etc etc!

In my view, all are error-ridden and, again in my view, none are valid measures of good dialysis.

That is exactly why something like the ‘Index’ is needed - if for nothing more than to re-focus our attention back onto the patient, the person, the dialyzor.

I totally understand Dr. Agar. So, we will also use the PRU in that case. Thanks!