You are an obstetrician seeing a 31-year-old pregnant woman who had an unprovoked deep venous thrombosis of the leg 5 years ago that was treated with warfarin for 6 months without complication. She is no longer using antithrombotic medication and is otherwise healthy. Given a possible increased risk of thrombosis with pregnancy, you are considering discussing the possibility of low-molecular-weight heparin (LMWH) prophylaxis for the rest of the pregnancy.
To inform your discussion, you search first for an evidence-based recommendation and find the following recommendation from a practice guideline1: “For pregnant women at moderate to high risk of recurrent venous thromboembolism (VTE) (single unprovoked VTE, pregnancy- or estrogen-related VTE, or multiple prior unprovoked VTE not receiving long-term anticoagulation), we suggest antepartum prophylaxis with prophylactic- or intermediate-dose LMWH rather than clinical vigilance or routine care (weak recommendation, based on low confidence in effect estimates).”
The statement “weak recommendation, based on low confidence in effect estimates” leaves you uncomfortable. You decide to read further to understand the recommendation and its rationale.
In general, patient management recommendations are developed in the context of clinical practice guidelines (see Chapter 5, Finding Current Best Evidence). However, you also may find guidance originating from a decision analysis. Similar criteria of credibility apply to both approaches.2-5
Practice guidelines are statements that include recommendations intended to optimize patient care. They are, ideally, informed by a systematic review of evidence and an assessment of the benefits and harms of alternative care options.2 To make a recommendation, guideline panelists must define clinical questions, select the relevant outcome variables, retrieve and synthesize all of the relevant evidence, rate the confidence in the effect estimates, and, relying on a systematic approach but ultimately also on consensus, move from evidence to recommendations.6 To fully inform their audience, guideline panels should provide not only their recommendations but also the key information on which their recommendations are based (see Chapter 28.1, Assessing the Strength of Recommendations: The GRADE Approach).
Decision analysis is a formal method that integrates the evidence regarding the beneficial and harmful effects of treatment options with the values or preferences associated with those effects. Clinical decision analyses are built as structured approaches (decision trees), and authors will usually include 1 or more diagrams showing the structure of the decision trees used for the analysis.
Figure 26-1 shows a simplified decision tree for the scenario of the pregnant woman considering thromboprophylaxis. The patient has 2 options: to use or not use prophylaxis with LMWH. The decision is represented by a square, termed “decision node.” The lines that emanate from the decision node represent the clinical strategies under consideration.
Diagram of a Simplified Decision Tree
Abbreviation: LMWH, low-molecular-weight ...