NDT Advance Access originally published online on August 25, 2007
Nephrology Dialysis Transplantation 2007 22(11):3110-3114; doi:10.1093/ndt/gfm516
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Intention and outcome in guideline-based nephrological practice: a suitable space for ‘clinical technology’
Department of Renal Medicine, St James's University Hospital, Leeds, UK
Correspondence and offprint requests to: E. J. Will, Honorary Renal Physician, Department of Renal Medicine, St James's University Hospital, Beckett Street, Leeds LS9 7TF, UK. Email: eric.will{at}leedsth.nhs.uk; wilesroyd{at}aol.com
Keywords: chronic renal disease; clinical practice guidelines;outcome measure; quality of care
 |
Introduction |
|---|
Top
Introduction
A need to reconcile...Common sense tells us that our intended actions are a powerful means of producing desirable outcomes in nephrology, as well as in life in general. Unfortunately, the evidence suggests that intention alone is a rather weak determinant of any outcome [1]. Indeed, sometimes we are even aware of a disbenefit from the best motivation, if we act without effective mechanisms and attention to detail [2]. Where effective means are available, the energy of intention may be considered the crucial factor [3], and that leads to the blame of clinicians for not responding to guideline advice. While the motivation and response of clinicians may well be deficient, other more systematic factors have a role to play. There are very few validated methods available in nephrology that would produce predictable results from clinical intervention, whatever the intention of the physician. The potential weakness of intention is less surprising, when we reflect that while intention may give a definite direction to our actions, to suggest the need to encourage a clinical laboratory result up or down, say, it does not itself indicate the most appropriate value at which to intervene nor does it give the scale of the treatment step required. An example could be in the management of renal anaemia. In bringing about a desirable haemoglobin (Hb) level, the Hb value at which the epoetin dose should be changed, and by what amount, is not informed by the intention to respond to guideline advice, although the direction of necessary changes will be apparent.
This is important, because guideline authors frequently leave clinicians with little but good intentions in the challenge of responding to their advice. They have been typically ambivalent about providing what we might call an intermediate ‘clinical technology’ that could translate their statements of the ideal into results in the real world of patient management. This has sometimes been rationalized as a wish to avoid ‘prescriptive’ practice, given the variation in local conditions, but there is also a remarkable lack of research material on the effective management of unselected patients in everyday situations. Guideline authors have thus been obliged to create hypothetical algorithms from the literature of treatment efficacy [4], which are usually untested in practice and may give poor or very modest results [5–7].
Some of this difficulty is being managed through methods of collaboration designed to reveal ‘best practice’ from what appears to be the complicated ‘black box’ of clinical management [8]. A range of quality improvement and assurance techniques from industry is typically converted for clinical application [9]. In nephrology, some more focused methods have been developed [10], but there persists a general concern about the translation of guideline advice into practice [11].
|
A need to reconcile individual with unit results |
|---|
A further problem is that comparative audit of Dialysis Unit performance requires outcome data to be expressed as summary statistics (mean, standard deviation (SD), etc.) [12]. Unless there is uniformity of management policy at unit level, the data cannot represent more than the hit or miss total sum of patient encounters. It is not then surprising that results are widely dispersed. To the inevitable individual patient variation, we must add the logistical difficulties of inconsistent clinical management, discontinuity of medical care, available clinical informatics, delays in prescription changes, etc. Most units operate very far from the ability to predict and to mould the results they desire to achieve in the light of guideline advice.
|
Categorical and continuous variables—the dispersion of clinical data |
|---|
In examining these issues, there may be a case for distinguishing guideline advice that addresses categorical as opposed to continuous variables, for example the introduction of a drug to a particular patient group (say ACEI in diabetic nephropathy) as compared with concern for laboratory values. The former are essentially ‘yes/no’ activities, which are readily documented, whereas the latter form a continuum of values and performance. Not infrequently, a lack of understanding of the roles of guidelines, clinical indicators and performance measures creates difficulty in expressing and interpreting the outcomes in both cases, but especially for continuous data, although definitions and examples are available [13]. Because of clinical events, case-mix and the logistical issues mentioned earlier, some dispersion of continuous values, expressed typically as SD, is always to be expected in the statistical representation of clinical variables [12]. This has been little examined to date, perhaps because it is only recently, apart from the UK Renal Registry (UKRR), that large scale databases have made this apparent in the scientific literature. Guideline statements have not generally been drafted to take these dispersions into account. This is partly because the definition of indicators and performance measures is often left to other more local agencies, which leads to a variety of desirable standards. However, advice that is intended to impinge on the real world of clinical practice cannot afford to ignore the consequences of such dispersion, not least because, ultimately, underachievement reflects on the status and acceptance of the guidelines themselves.
The effect of different unit mean values on overall compliance with performance measures can be expressed conveniently by plots of unit mean or median results against percent compliance with any given limit, the Rose–Day plot [14], a routine adopted by UKRR over the past 10 years (Figure 1) [15]. This format has been applied largely to express unit haemoglobin and urea reduction ratio results but can be redrawn for blood pressure, (Serum) phosphate and cholesterol with equal benefit. An awareness and acceptance of data dispersion clarifies clinical activity, since we cannot aim literally at averages, and also emphasises the probabilistic basis and continuum of unit achievement for the different clinical indicators [16]. The challenge to unit performance becomes clearer, whether it is a matter of narrowing the outcome distributions by some means or revamping a treatment policy.
|
|
The conversion of desirable limits into practical instructions |
|---|
The interaction of guideline statement and practice is nowhere better demonstrated than in the direct conversion of a declared, desirable guideline limit for a continuous variable into an instruction for clinical management, or some form of ‘target’ for clinical activity. There is no logical reason why such conversions alone should provide effective practical instructions for producing acceptable results and several reasons why they would not. Even in the best hands, in research studies, there is a persistent gap between the desired and achieved result. Perhaps the RCT comparisons of haemoglobin goals between ‘high (normal)’ and ‘low’ values provide the best example [compiled in 17]. This has been crucial to such studies, since the poorly achieved separation of the study groups has undermined conclusions about the consequence of renal anaemia for cardiac status, in particular. This gap between expectation and achievement seems to require some specific mechanisms for correction.
|
A principle of systematic over-aspiration |
|---|
One such possible mechanism has been demonstrated for that most controllable outcome, dialysis adequacy. Further analysis of the HEMO study data showed that to achieve consistently a Kt/V of 1.2 required an attempt (intention) to treat to a value of 1.4 [18,19]. It appears that there is an imbalance in the factors that help or hinder the performance of haemodialysis, with a preponderant disadvantage to adequate delivered dialysis dose. The factors that facilitate dialysis are far outweighed by those that obstruct it. This can be seen as the clinical expression of entropy (by analogy with thermodynamics), when systems have a tendency to become spontaneously less effective (organized) rather than remaining stable. It means that a systematic over-achievement must be put in place for such systems to perform reliably. All other continuous clinical outcomes in nephrology are much less easily achieved than Kt/V, so that such an insight must be even more relevant to them. This demonstration suggests that there may be treatment principles that could be developed as a framework for clinical management, which otherwise can be characterized only piecemeal in the ‘black-box’ exercises of uncovering best practice.
|
A principle of pre-emptive intervention |
|---|
The inevitable trajectory (over time) of clinical values in each patient, which scarcely applies to Kt/V, is particularly relevant when clinical values show a tendency to move in one direction, for example blood pressure to rise, or haemoglobin to fall. In that case, clinical intervention must be made well before guideline limits are reached, since by the time treatment is changed, the values are likely to have moved beyond the desirable limit. This type of pre-emption has been a feature, probably inadvertent, of some guidelines, where authors were aware that a ‘buffer’ was required to allow for the inevitable logistical and physiological delays of intervention. It was built more deliberately into the UK NICE Renal Anaemia advisory algorithms [20]. Another example, of Se ferritin levels being constrained below 800 ng/ml in haemodialysis, by the discontinuation of iron supplements at, say, 500 ng/ml, has taken a long time to be accepted [21]. This operates also at the lower end of the desirable ferritin range [22].
|
The framing of guideline statements |
|---|
As well as the intervention principles of systematic over-achievement and pre-emptive action, it is worth considering the possible practical effects of the way that guideline statements are drafted and framed. Persistent, presumably fashion-related, differences in the framing of advice are noteworthy and possibly confusing. Desirable blood pressure control for example, is always expressed as values ‘less than’ a certain limit. There is also a lack of common understanding of the terminology, where a ‘target’ is taken to imply either a treatment goal or an aiming-point or both. For example, a blood pressure goal may be taken to be ‘.... the best estimate of a safe and achievable level at which treatment should be directed ...’, which is arguably conflating an aiming-point and a goal of clinical activity [23]. This appears at first to be a rather pedantic distinction, until it is realised that it is responsible for the direct conversion of guideline limits into practical instruction, which is liable to be ineffective because of the principles we have discussed.
Furthermore, there is, in practice, not even a requirement for a literal aiming-point in most clinical intervention [24]. At the start of any treatment it may be useful to envisage a notional ‘target’ (aiming point) value to define a desirable course, but in the maintenance stage of treatment, clinical management consists typically of adjustments to encourage low values to rise and high values to fall. This depends on a judgement about the value at which intervention is necessary and prescription of a suitable dose of medication, precisely those elements that intention does not specifically address. This is not really a precise enough exercise to deserve the ‘target’ (aiming point) analogy at all, although a goal expresses usefully the direction of intent. It is this type of analysis that has allowed the development of treatment decision support based on standardized thresholds for intervention and dose ladders [25], as well as approaches that attempt to incorporate previous treatment responses [26].
There may also be a failure to express fully the implications of a desirable limit, such as the necessary conversion of ‘blood pressure should be <130 mmHg’ into the useful instruction ‘start the last hypotensive agent at 130 mmHg ...’. The uncertain behaviour of clinicians working close to desirable guideline limits during consultation has been recognized, but very little investigated [27]. For guideline instructions couched as ‘less than’ a given value, the consequences of the form of statement remain to be seen, although even with current terminology a benefit from defined low intended values (goals) is recognized in research studies [28,29]. With the known inevitable dispersion of outcome values and the need in principle to aspire beyond the declared limit, there must be uncertainty as to just how far below any such limit is being suggested as a notional aiming point. This all suggests that guideline statements can themselves have an effect on the interpretation of necessary clinical management and so create scarcely perceived confusion, even perhaps underachievement, at the practical level.
|
Conclusion |
|---|
In summary, it is likely that the gap between intention and outcome can be populated usefully with a range of clinical techniques based on principles of clinical intervention. It is possible to see the outline of a practical clinical technology that would improve clinical results, both at the level of individual patient and patient cohort. Combining unit calibration using Rose–Day plots, systematic over-aspiration, pre-emptive intervention, more carefully framed guideline statements and threshold (rather than targeting) paradigms and structured patient review might well reduce the persisting frustrations of inadequate ‘implementation’ of guidelines [30]. This approach should provide a starting point for, and complement, methods designed to expose ‘best practice’ in routine clinical management. Arguably, if guideline authors do not become more engaged with the practical consequences of their advice, other agencies are likely to fill the resulting vacuum. This is already apparent in renal anaemia management in the US, where Medicare have imposed detailed treatment conditions for re-imbursement involving both an upper Hb intervention threshold (haematocrit 36%/Hb 12 g/dl) and an epoetin dose change specification (reduce by 25%).
It is not easy to find the balance between individual and group clinical management that will give the best results, but some reconciliation appears necessary, to allow predictable outcomes and effective comparative audit. Sooner or later, it seems likely that the quite understandable, but perhaps unnecessarily narrow, overstatements of the need for individual management will come to be seen in a more general context, and for the general good. Essentially, the most advantaged individuals will be found in the best performing groups.
Gratifyingly for clinicians, these issues are the substrate of a truly clinical research. There is a need to define the optimal points of intervention and dosing, for each continuous variable, of local patient groups. The challenge of working with more specific principles of clinical intervention has the potential to energize a clinical effort that is seen as deficient in guideline application. There have been, and are, enough concerns about the guideline culture to wish to avoid any unnecessary underachievement, when the principles of a necessary clinical technology have yet to be fully recognized and systematically exploited [31–33]. Guideline authors need to consider the means, and not simply the ends, of clinical achievement. However, until the principles of clinical intervention are more widely accepted, they can retort legitimately that medical cultures and health care systems get the guidelines they deserve!
Conflict of interest statement. The author is medical adviser to Media Innovations, a University of Leeds/NHS technology transfer company.
|
References |
|---|
- Ormerod P. Why Most Things Fail ... and How to Avoid it (2005) London: Faber and Faber.
- Runciman D. The Politics of Good Intentions (2006) Princeton and Oxford: Princeton University Press.
- Berwick D. Will is the way to win patient safety. Health Serv J (2007) 117:18–19.[Medline]
- Haynes B. Can it work? Does it work? Is it worth it? The testing of health care interventions is evolving. Br Med J (1999) 319:652–653.
[Free Full Text] - National Kidney Foundation. K/DOQI Clinical Practice Guidelines for Bone Metabolism and Disease in Chronic Kidney Disease. Am J Kidney Dis (2003) 42([Suppl 3]):S1–S201.[Medline]
- Arenas MD, Alvarez-Ude F, Gil MT, et al. Application of NKF-K/DOQI Clinical Practice Guidelines for Bone Metabolism and Disease: changes of clinical practices and their effects on outcomes and quality standards in three haemodialysis units. Nephrol Dial Transplant (2006) 21:1663–1668.
[Abstract/Free Full Text] - Wald R, Tentori F, Tighiouart MS, et al. Impact of the Kidney Disease Outcomes Quality Initiative (KDOQI) Clinical Practice Guidelines for bone metabolism and disease in a large dialysis network. Am J Kidney Dis (2007) 49:257–266.[CrossRef][Medline]
- Mills PD, Weeks WB. Characteristics of successful quality improvement teams: lessons from five collaborative projects in the VHA. Jt Comm J Qual Saf (2004) 30:152–162.[Medline]
- Joshi MS, Bernard DB. Clinical performnace improvement series. Classic CQI integrated with comprehensive disease management as a model for performance improvement. Jt Comm J Qual Improv (1999) 25:383–395.[Medline]
- Matchar DB, Patwardhan MB, Samsa GP, Haley WE. Facilitated process improvement: an approach to the seamless linkage between evidence and practice in CKD. Am J Kidney Dis (2006) 47:528–538.[CrossRef][ISI][Medline]
- Grol R, Buchan H. Clinical guidelines: what can we do to increase their use? Med J Austr (2006) 185:301–302.
- Ansell D, Feest T, Tomson C, Williams AJ, Warwick G, eds. UK Renal Registry Report 2005. UK: UK Renal Registry Bristol. Available at. http://www.renalreg.com/Report%202006/Cover_Frame2.htm (Accessed 7 June 2007).
- Califf RM, Peterson ED, Gibbons RJ, et al. Integrating quality into the cycle of therapeutic development. J Am Coll Cardiol (2002) 40:1895–1901.
[Abstract/Free Full Text] - Rose G, Day S. The population mean predicts the number of deviant individuals. Br Med J (1990) 301:1031–1034.[ISI][Medline]
- Ansell D, Feest T, Williams A, Winearls C, eds. UK Renal Registry Report 2005. UK: UK Renal Registry Bristol. Available at http://www.renalreg.com/Report%202005/Cover_Frame2.htm. Urea Reduction Ratio fig 7.3:104, Haemoglobin fig 8.15:122. (Accessed 7 June 2007).
- Will E J. Aiming at averages. J Roy Soc Med (2001) 94:617–619.
[Free Full Text] - National Kidney Foundation. KDOQI Clinical Practice Guidelines and Clinical Practice Recommendations for Anemia in Chronic Kidney Disease. Am J Kidney Dis (2006) 47([Suppl 3]):S50.
- Depner T, Beck G, Daugirdas J, et al. Lessons from the hemodialysis (HEMO) study: an improved measure of the actual hemodialysis dose. Am J Kidney Dis (1999) 33:42–49.
- National Kidney Foundation. KDOQI Clinical Practice Guidelines and Clinical Practice Recommendations for 2006 Updates: hemodialysis adequacy, peritoneal dialysis adequacy and vascular access. Am J Kidney Dis (2006) 48:S31.[CrossRef]
- Anaemia management in Chronic Kidney Disease 2006. http://guidance.nice.org.uk/cg39/?c=91526 (Accessed 7 June 2007).
- Richardson D, Barlett C, Will EJ. Optimizing erythropoietin therapy in hemodialysis patients. Am J Kidney Dis (2001) 38:109–117.[ISI][Medline]
- Richardson D, Bartlett C, Jolly H, Will EJ. Intravenous iron for CAPD populations: proactive or reactive strategies. Nephrol Dial Transplant (2001) 16:115–119.
[Abstract/Free Full Text] - Singer GM, Izhar M, Black HR. Guidelines for hypertension: are quality-assurance measures on target? Hypertension (2004) 43:198–202.
[Abstract/Free Full Text] - Will EJ. Targets and Targeting. Am J Kidney Dis (2001) 38:411–414.[ISI][Medline]
- Will EJ, Richardson D, Tolman C, Bartlett C. Development and exploitation of a clinical decision support system for renal anaemia. Nephrol Dial Transplant (2007) 22([Suppl 4]):iv31–iv36.
[Abstract/Free Full Text] - Martin Guerro JD, Olivas ES, Valls GC, et al. Use of neural networks for dosage individualisation of erythropoietin in patients with secondary anemia to chronic renal failure. Comput Biol Med (2003) 33:361–373.[CrossRef][ISI][Medline]
- Ferrari P, Hess L, Pechere-Bertschi A, et al. Reasons for not intensifying antihypertensive treatment (RIAT): a primary care antihypertensive intervention study. J Hypertens (2004) 22:1221–1229.[CrossRef][ISI][Medline]
- Sarnak MJ, Greene T, Wang X, et al. The effect of lower target blood pressure on the progression of kidney disease: long term follow up of the modification of diet in renal disease study. Ann Intern Med (2005) 142:342–351.
[Abstract/Free Full Text] - Clinical Practice Guidelines and Clinical Practice Recommendations for Diabetes and Chronic Kidney Disease. Am J Kidney Dis (2007) 9([Suppl 2]):S1–S180.
- Plantinga LC, Fink NE, Jaar BG, et al. Frequency of sit-down patient care rounds, attainment of clinical performance targets, hospitalization, and mortality in hemodialysis patients. J Am Soc Nephrol (2004) 15:3144–3153.
[Abstract/Free Full Text] - Locatelli F, Andrulli S, Del Vecchio L. Difficulties of implementing clinical guidelines in medical practice. Nephrol Dial Transplant (2000) 15:1284–1287.
[Free Full Text] - Uhlig K, Balk EM, Lau J, Levey AS. Clinical Practice Guidelines in nephrology – for worse or for better. Nephrol Dial Transplant (2006) 21:1145–1153.
[Free Full Text] - Narins R, Bennett WM. Patient Care Guidelines: problems and solutions. Clin Am J Soc Nephrol (2007) 2:1–2.[CrossRef]
Accepted in revised form: 4. 7.07
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||