Letters

Is there a rationale for rationing chronic dialysis?

BMJ 1999; 318 doi: https://doi.org/10.1136/bmj.318.7198.1619 (Published 12 June 1999) Cite this as: BMJ 1999;318:1619

Two points need clarification

  1. Rasheed Ahmad (TheAhmads{at}aol.com), Consultant renal physician
  1. Royal Liverpool University Hospital, Liverpool L7 8XP
  2. Jersey General Hospital, St Helier, Jersey, Channel Islands
  3. Department of Medicine and Therapeutics, University of Aberdeen, Aberdeen AB25 2ZD
  4. Renal Unit, Aberdeen Royal Infirmary, Aberdeen AB9 2ZB
  5. Department of Nephrology, Lister Hospital, Stevenage SG1 4AB

    EDITOR—I agree with Chandna et al that late referral is an important factor in the survival of patients receiving dialysis.1 I am unaware, however, of the source of their information that an NHS consensus statement recommends that nephrology referrals be made at a serum creatinine concentration >135 μmol/l in women and >180 μmol/l in men. The workload generated by this practice would be so enormous that British renal units under present conditions would be unable to cope.

    I am also concerned that of the authors' 292 patients receiving dialysis, 26 (admittedly high risk) patients spent 44% of their lives in hospital. The costs of inpatient treatment for these high risk patients would be substantially higher than the average cost of £250 a day that the authors quote.

    References

    Study's objectives were not achieved

    1. T Farrant (106712.1276{at}compuserve.com), Specialist registrar in gastroenterology
    1. Royal Liverpool University Hospital, Liverpool L7 8XP
    2. Jersey General Hospital, St Helier, Jersey, Channel Islands
    3. Department of Medicine and Therapeutics, University of Aberdeen, Aberdeen AB25 2ZD
    4. Renal Unit, Aberdeen Royal Infirmary, Aberdeen AB9 2ZB
    5. Department of Nephrology, Lister Hospital, Stevenage SG1 4AB

      EDITOR—Chandna et al's study asking whether there is a rationale for rationing chronic dialysis does not fulfil its stated purpose.1 The study successfully shows the groups of patients already receiving dialysis who could be expected to do poorly, and this might be used as a rationale for withdrawing treatment in this small group. The cost savings that their unit would accrue are nevertheless small (just 3.2% of the cost of their chronic dialysis programme).

      Even in the authors' high risk group five of the 26 patients were long term survivors. This is similar to the effect of interferon treatment in hepatitis C in terms of long term response and better than results of in vitro fertilisation programmes. One could argue, therefore, that these patients should not be refused treatment either. Quality of life considerations would be important in this group since these five patients may have had an excellent quality of life, as indeed may some of the other patients.

      This study cannot be considered a basis for rationing access to renal replacement. The authors do not report the characteristics or reasons for refusal of patients referred to them for renal replacement. Their assertion that Wiltshire Health Authority's recommendations may be used as a starting point in deciding prioritisation is also not valid. They agree that these recommendations are couched in general terms but do not specify them in their paper. Furthermore, they state that patients were assessed on an individual basis, with no formal criteria being used.

      The study shows that the criteria used to decide acceptance on to the authors' renal replacement programme work for the patients who are accepted. What it does not tell us is whether the correct criteria are being used for those refused renal replacement. It would be interesting to know what happened to those refused replacement and on what grounds this decision was made.

      References

      Question is difficult to address on basis of retrospective studies

      1. Fergus J Caskey (f.caskey{at}abdn.ac.uk), Clinical research fellow,
      2. Wendy Metcalfe, Clinical research fellow,
      3. Alison M MacLeod, Reader,
      4. Izhar H Khan, Consultant physician
      1. Royal Liverpool University Hospital, Liverpool L7 8XP
      2. Jersey General Hospital, St Helier, Jersey, Channel Islands
      3. Department of Medicine and Therapeutics, University of Aberdeen, Aberdeen AB25 2ZD
      4. Renal Unit, Aberdeen Royal Infirmary, Aberdeen AB9 2ZB
      5. Department of Nephrology, Lister Hospital, Stevenage SG1 4AB

        EDITOR—Chandna et al's paper confirms the influence of comorbidity on survival of patients receiving renal replacement that has been shown in previous retrospective studies.1 The question of rationing dialysis, however, while pertinent, is difficult to address on the basis of such studies.

        Chandna et al costed the potential savings from rationing chronic dialysis on the basis of comorbidity. As no source or methods were given it was not clear whether the costs were from the perspective of the institution, the healthcare system, or society. Nor was it clear whether these costs related to haemodialysis or peritoneal dialysis. To be included in any rationing debate, clinical and cost data must be robust.

        An obvious limitation of any retrospective study is that some of the information needed may not have been recorded in the case notes. This makes the assessment of functional data (Karnofsky scores) and comorbidity unreliable and difficult to reproduce. A subjective comorbidity scoring system, completed in many cases after death, is a potential source of bias, and we are currently studying prospectively the influence of comorbidity and functional status on outcomes of renal replacement in various parts of Europe.

        Chandna et al compared their comorbidity severity score with that used by Khan et al in a retrospective survival analysis2 and found considerable differences in the proportion of patients in each risk group (for example, high risk group 51% and 23% respectively). Although this may reflect higher comorbidity in Chandna et al's patients, their interpretation of the risk stratification system may have differed. Furthermore, patients receiving a transplant or recovering renal function (18%) were left in Chandna et al's survival analysis whereas Khan et al censored patients at such events. The better prognosis of this subgroup would certainly influence the overall survival rates.

        The purpose of risk stratification to study outcomes should be twofold: to enable valid comparison of outcomes between centres, and to identify groups of patients for whom efforts should be directed at improving outcomes rather than to deny such patients treatment. The risk stratification system used by Khan et al and Wright was developed to correct for case mix and allow comparison of survival between centres. 2 3 Recognising that a patient is at high risk also allows us to be more informed when discussing prognosis and outlook with patients receiving dialysis and their families.

        References

        Authors' reply

        1. Shahid M Chandna (shahid.chandna{at}lister.org.uk), Associate specialist,
        2. Roger N Greenwood, Consultant,
        3. Ken Farrington, Consultant
        1. Royal Liverpool University Hospital, Liverpool L7 8XP
        2. Jersey General Hospital, St Helier, Jersey, Channel Islands
        3. Department of Medicine and Therapeutics, University of Aberdeen, Aberdeen AB25 2ZD
        4. Renal Unit, Aberdeen Royal Infirmary, Aberdeen AB9 2ZB
        5. Department of Nephrology, Lister Hospital, Stevenage SG1 4AB

          EDITOR—Ahmad is right that there is no NHS consensus recommending nephrology referrals. This was a typographical error and should have read NIH (National Institutes of Health).1 Encouraging early referral undoubtedly means more work in nephrology outpatient departments, but this may partially be offset by reduced work and costs owing to fewer unplanned presentations for dialysis; at present these account for 44% of all patients entering the long term dialysis programme.

          Ahmad may be right in expecting higher than average inpatient costs for the 26 high risk patients. This does not, however, alter our conclusions, since many of these patients did not live long. In total they had 428 inpatient days during the first three months, and even if these are costed at an average of £350 a day, the figure in our table 3 only changes from 3.2% to 3.5%.

          We do not agree with Farrant that our study's objectives were not achieved. We identified several factors affecting survival and hospitalisation in patients receiving dialysis and, using these, defined a small high risk group with a poor one year survival. We suggested that the methods used to define this group might be tested prospectively. We did not advocate using them as a basis for formal rationing, but they might contribute to the understanding of a complex decision making process. Farrant poses the same question that we did: is a one year survival of 19% low enough to justify withholding dialysis? This is a value judgment that a study such as ours cannot make.

          We agree with Caskey et al that the question of rationing is difficult to address, but it is important to make a start. The costs we quoted were those of dialysis (both haemodialysis and peritoneal dialysis) and hospitalisation. Clearly, other costs also arise, but these are difficult to quantify retrospectively. We look forward to the results of Caskey et al's prospective work.

          The original purpose of risk stratification was to compare outcomes between centres. 2 3 We hypothesised that a more tightly defined high risk group could be identified in which renal replacement was inappropriate.4 Our analysis did not substantiate this but may have resulted in an improved definition of high risk in this setting. This new definition may indeed be better geared for counselling patients with potentially poor prognosis. Wright's and Khan et al's classification, with a two year survival of 55% in high risk group in our population, is clearly not suited for this purpose.

          References

          View Abstract

          Sign in

          Log in through your institution

          Free trial

          Register for a free trial to thebmj.com to receive unlimited access to all content on thebmj.com for 14 days.
          Sign up for a free trial

          Subscribe