Diagnostic conversations: Clinical Decision Making in surgery – Part 1

Introduction

This paper outlines two of the four stages of a clinical decision making model developed for surgery. The model is elaborated through an e-mail discussion between two surgeons, both are Directors of Surgery in their own Institution; one (David) is an experienced general surgeon specialising in colorectal surgery, the other (Spencer) is a senior paediatric surgeon. Whilst at one level the conversation is about a particular patient, and the issues it raised for the clinicians concerned, there is another level where the conversation addresses a broader collaboration in developing and promulgating a model of Clinical Decision Making (CDM). The third author (Wendy) is not a clinician: she has worked together with both surgeons on the development of the model described in this chapter, and on interpreting and commenting on the cognitive processes they and other surgeons employ.

There are four major stages we are teaching in a CDM course based on the model:

1. The Working Diagnosis and Initial Management Plan (this can take minutes, hours, days or weeks depending on the case and its urgency).

2. Preparing for a procedure – this takes place just before the procedure is performed, sometimes even whilst the patient is being anaesthetized.

3. Monitoring progress of the management plan and procedure. This includes any need to modify the plan or even in some cases abandon the procedure, or choose an alternative (bail out).

4. Reviewing what was done and whether it turned out best for the patient, the team and the proceduralist.

In this paper we focus mainly on the first two stages. A more detailed description of the first three stages, plus some of the theory underpinning the model, was published earlier this year [1].

The conversation

From: David

Dear Wendy and Spencer,

How about this for a case that I have to review?

1. Formulating a working diagnosis and a management plan

   An 84-year-old cachectic patient presents with a near obstructing lesion of the mid-transverse colon and a large 18 mm stone in a non-inflammed gall bladder. He has type 2 diabetes, weighs 50 kg, has suffered marked weight loss and his serum albumin is 24 g/L. Investigations show no evidence of metastatic disease. A colonoscopy has failed to traverse the tumour but biopsies taken confirm adenocarcinoma. The patient is consented for a laparotomy and large bowel resection, likely to involve an extended right hemicolectomy. Fortunately, the consent actually included the possibility of a stoma (if no anastomosis) and that a decision about the gall bladder and its stone would be made during the operation.

2. Preparing for the procedure: questions and issues to consider:

   • What do you consider are the significant risks to this patient and how might this modify your choice of procedure?

   • Do you expect to perform an anastomosis?

   • Has the patient consented to a stoma and even been marked for one?

   • Will you need to mobilise the splenic flexure?

   • What do you plan to do about the gall bladder if it appears healthy other than the gall stone?

   • If you were to perform a cholecystectomy will you order an intraoperative cholangiogram?

   • How long will it take and what is a reasonable length of time for this patient and this procedure?

3. Monitoring progress during the procedure:

   The team concerned opens the abdomen through a midline laparotomy.

   Findings: There is no metastatic disease, there is some ascitic fluid. The omentum seems free other than where it is adherent to the tumour. There are some lymph nodes palpable indicative of likely lymph node metastasis. The gall bladder is uninflammed.

Decision-making:

1. Do you resect this tumour?

2. Do you do an extended right hemicolectomy?

3. Do you mobilise the splenic flexure (appears necessary if you are going to do a stapled anastomosis)?

4. Do you perform an anastomosis at all?

5. What will you do about the gall bladder?

Following the procedure:

The patient underwent a stapled extended right hemicolectomy with mobilisation of the splenic flexure. The gall bladder was removed without an operative cholangiogram. This was uncomplicated and took no more than 15 min. The whole procedure took 120 min. There was no untoward blood loss or intraoperative issues. Everything went according to plan.

Review

What can be learned about the decision making in this case?

Dear David,

This is an interesting and complex case.

I wonder whether ‘Preparing for the Procedure’ should be taken one step back, looking at his life situation, family support, life expectancy, own expectations, mental alertness (e.g., ability to look after himself, or stoma etc.) and surgical options, and the benefits and limitations of each. How this might influence the clinical decision making, etc – before moving into what his treatment is going to be. What I mean is analyse the situation from the perspective of the patient, their physiology and pathology.

Who was performing this operation and what state were they in? Perhaps we could also include other external factors such as time of day, quality of registrar/assistant, availability of theatre etc. Higher order CDM has to take into account far more than just the clinical features and X-rays, etc: This is situation analysis considering the personnel and the capabilities of the team including whether or not some or all of them might be fatigued. It considers the environment and context of care delivery.

Then we return to answer the questions you propose around the chosen operation.

Spencer

Dear Spencer,

Not sure how much more detail should be given pre-op but here is some:

Patient and physiology: He had a supportive family though was a widower. He would have been able to manage a stoma which might only have been temporary had an anastomosis not been performed until his albumin was 30 g/L or more (US studies suggest that albumin <35 g/L is associated with anastomotic leak though that is not the same as not doing an anastomosis).

Pathology: He is likely to obstruct if a procedure is not done. A bypass is another possible option, but more often employed where there are known distant mets. Resecting the primary tumour usually gives the best palliation for cases that are not curable.

Prognosis: His life expectancy if Dukes’ C with adjuvant chemo (8/25 nodes were involved – not known preop) is probably 30%–40% 5 years and less than 2 years if curative resection not achieved.

Our evidence shows that after surviving the first year, patients who have a curative resection die at the same rate as the normal population or better (they’ve been strong enough to survive surgery).

David

Thanks David,

Yes that is useful extra information, particularly for clinicians not in that specialty area.

It highlights the challenge of producing a case that all course participants, regardless of specialty or discipline, can consider at a high cognitive level. We have to provide them with much of the background that influences the clinical decisions.

Commentary

Being very experienced in this area, David moved subconsciously and probably very rapidly from the clinical data to pattern recognition, a working diagnosis and then into questions, areas of uncertainty and mental preparation for the procedure. The left side of the model in Figure 1 shows the relationship between experience, pattern recognition and the sub-conscious decision making arising from experience combined with reflection and evaluation, whilst the right side of the model outlines the multiple steps taken during analytical interpretation of the same data. David’s case notes also demonstrate the way in which an expert’s thinking can move across the cognitive continuum to analysis and anticipation of variance and/or complications [2, 3]. Uncertainty is a valuable tool in this process because it functions as a warning signal indicating the need to continually question decisions.

Figure 1

A model for clinical decision making.

In comparison, Spencer could be said to be working more on the analytical side of the model, asking for more information about the patient, more consideration of possible options and risks, and a more cautious approach to the diagnosis and management plan. This approach to the clinical data about the patient is more like that of a person with less experience with the clinical condition – a novice [1, 4].

As demonstrated in the case described, our approach to decision making emphasises that CDM is not an event. Rather it is an on-going cyclical process, constantly moving both through the cognitive continuum from subconscious to analytical decision making (west – east), and from the clinical data to the patient’s current condition (north – south).

On the other hand, in this context, the main purpose of Spencer’s question to David was to “tease out” all the information that the expert (David) would have used during CDM. Mostly, this information is considered in a subconscious way which is one reason why teaching CDM effectively proves so difficult. The decision making of expert clinicians often occurs so effortlessly that they do not even fully appreciate its complexity. This is why good clinicians do not always make good teachers. The challenge offered by Spencer was whether the scenario could adequately reflect or expose the vast amount of information that normally goes into making an apparently straightforward clinical decision.

One of the advantages of our model is that it presents a way to identify and describe differences in the CDM that occurs between a very experienced surgeon, an ‘expert decision maker’ and the novice. That difference is in the extent to which an expert is able to recognise and accept a range of ambiguous information, to anticipate a range of possible different pathways, with different implications, within the procedure, and at the same time be appropriately confident that they can monitor the procedure as it progresses and be able to adapt to what they find [5, 6].

A thorough and careful preparation as outlined by David also has the advantage of reducing the expert surgeon’s cognitive load [7] (that is information that the surgeon may be required to consciously think about) thus allowing more opportunity for recognising and anticipating cues and responding adaptively [8]. Can the clinician see the “wood for the trees”, and de-clutter or weight correctly the variety of information that is presented? The expert can do this in a way that consistently produces a decision that is aligned with the patient’s best interests. It is one thing to be an expert clinician; it is another to be able to teach it.