@LIVES 2015 year in review 2
Ten things you should consider before you believe a clinical practice guideline
http://bit.ly/1CigGER
1.
A CPG has to be reasonably current, and the definition of ‘reasonably current’ seems to change. Published in 2013 [1], the Surviving Sepsis Campaign (SSC) guidelines dealing with resuscitation algorithms, resuscitation fluids or transfusion in patients with sepsis did not consider a number of studies and meta-analyses published subsequently [2, 3]. These recommendations may need revision even before the next ‘full text’ is published––a 4-year hiatus between editions seems increasingly inadequate. Ongoing electronic updates, often referred to as living guidelines, are likely coming our way.
2.
The clinical questions have to be specific enough to deliver practical and actionable answers. To this end, we suggest they follow the PICO format for Population, Intervention, Comparator, and Outcome. The next three points deal with the choice of population, comparator, and outcome.
3.
Specific questions have to take into account the ‘age of the evidence.’ Recommendations regarding the use of beta blockers at the time and following myocardial infarction may well need to rely on data emerging from randomized studies performed in the current reperfusion era with possibly little emphasis on older randomized studies as the number of new co-interventions used in both groups might have a modifying effect [4].
4.
The interventions and comparator should be clinically relevant. A comparison of all colloids to all crystalloids in the setting of sepsis make little sense in light of the realization that not only all colloids—but also all crystalloids—are not created equal [5, 6].
5.
The outcomes assessed in judging the benefits should be patient-important outcomes as opposed to convenient surrogate or substitute endpoints. Improvement in cardiac output, urine output, central venous pressure, and/or peak flows may suggest possible benefit. Unfortunately, apparently beneficial changes in surrogate outcomes have very often been associated with lack of benefit in patient-important outcomes and—not infrequently—in harm [7]. Diagnostic technologies may provide similarly misleading results: knowledge, or the impression of knowledge, does not always translate into patient benefit (most of us remember the widespread use of right heart catheters which likely had no benefit and possibly did more harm than good) [8, 9].
6.
As undesirable effects are unavoidable, guidelines should clearly and comprehensively present both the desirable and undesirable consequences of the recommended courses of action in an easy-to-access and easy-to-understand format [10]. Those estimates should be based on systematic reviews and should be clearly linked to the recommendations which are using them as justification.
7.
These estimates of effect may be more or less trustworthy. Guidelines should make the certainty of the evidence (also known as confidence in the estimates or quality of the evidence) explicit, clearly presenting the degree of confidence (from highly confident to poorly confident) associated with the evidence [11].
8.
Guidelines should clearly distinguish between recommendations that apply to virtually all patients in all circumstances and those in which alternative courses of action are both reasonable and may—in specific circumstances—be preferable. The former, strong recommendations are usually based on at least moderate certainty of treatment effects AND a decisive balance in favor of beneficial effects of intervention. Guidelines typically offer weak recommendations when intervention effects are uncertain OR there is high certainty of the estimates, but the balance between desirable and undesirable outcomes is close. The judgments underlying decisions, especially controversial ones, should be explicitly acknowledged [12].
9.
CPGs should provide an account of how the conflict of interest (COI) was addressed. Simple disclosure is unlikely to be sufficient. Active interventions to decrease the possibility of both financial and academic COIs are being increasingly recognized as needed [13, 14].
10.
Lastly, authors of CPGs should make it clear that they do not wish to generate uniform practice for all patients and all clinicians, but rather provide background information and some wider judgments necessary to make decisions. The individual decisions should then account for individual patient’s values and preferences and idiosyncratic clinical specificities. Authors should stress the non-dogmatic character of the majority of recommendations, especially those based on low certainty of effects or those associated with a close balance between harms and benefits. However, even strong recommendations should leave room for patient-specific flexibility. After all, this is all about guidance not forced uniformity [15].
sepsis guideline 在 Surviving Sepsis Campaign: Guidance on the Guidelines 的推薦與評價
Understand the history of the Surviving Sepsis Campaign, a joint collaborative between the European Society of Intensive Care Medicine and ... ... <看更多>