Correspondence

Learning from Our Errors

N Engl J Med 1997; 336:876-878March 20, 1997

Article

To the Editor:

The diagnostic error we are invited to learn from in the interesting Clinical Problem-Solving article by Brown et al. (Oct. 3 issue)1 would not have occurred if the original clinicians had been familiar with the medieval logical tool known as Ockham's razor. William of Ockham (1285 to 1349), a Franciscan monk and logician, postulated that “entities must not be unnecessarily multiplied,”2 or to put it another way, it is illogical to propose two processes to explain an observation when one will do.

This idea is central to the making of sensible medical diagnoses. The most likely explanation of a clinical problem is the one that is capable of explaining all the known facts.

Brain biopsies are all very well, but brainy physicians will always have Ockham's razor in their medical bags. It is an inexpensive, noninvasive diagnostic tool that carries a favorable risk–benefit assessment. It should be given to all students in medical school and should be used at the earliest opportunity in every medical consultation.

Hugh S. Boardman, M.A., M.B.
Boardman Clarke Partners, London SW20 8BA, United Kingdom

2 References

1. 1

   Brown J, Worthington MG, Lathi ES. Learning from our errors. N Engl J Med 1996;335:1049-1053
   Full Text | Web of Science | Medline

2. 2

   Ockham. In: Drabble M, Stringer J, eds. The concise Oxford companion to English literature. Oxford, England: Oxford University Press, 1987:406.
   

To the Editor:

Dr. Arnold Rich taught that tuberculous meningitis is preceded by a caseous tubercle in the substance of the brain, superficial under the meshes of the pia-arachnoid or deeper, or in the choroid plexus — a finding that should have led to earlier suspicion and diagnosis in the case described in “Learning from Our Errors.”

Eugene Blank, M.D.
4940 S.W. Humphrey Park Rd., Portland, OR 97221-2342

To the Editor:

First of all, it is of interest to remember that tuberculoma was the leading intracranial expansive process at the beginning of this century. The cost–benefit ratio of the stereotactic brain biopsy must be reviewed. In this case and many others, the results of an open cerebral biopsy would be conclusive. Time is also important, and this patient lost precious days before a correct diagnosis was made. We do not agree that a stereotactic biopsy is an accurate approach to the diagnosis of brain tumor. An open cerebral biopsy is a safe and objective approach, whereas the stereotactic biopsy is also safe but is a secondary and inaccurate approach.

Another important observation we would like to make is that an immunohistochemical study of the cerebral fragments was not done. This is an accurate procedure that would have been useful to the diagnosis, excluding the possibility of a tumor at the beginning.

João Roberto D. Azevedo, M.D.
Mario Sergio Andrioli, M.D.
Sociedade Neurologica Integrada, 04532-040 São Paulo, S.P., Brazil

Author/Editor Response

The authors reply:

To the Editor: We thank Dr. Boardman and Dr. Blank for their interesting comments, but we fear that they contain the potential for serious clinical error. The fundamental lesson from this case is that there was no definite diagnosis until there was a reliable tissue diagnosis, which ultimately meant obtaining an unequivocal tissue diagnosis of this very large brain mass. Blank is certainly correct when he points out that tuberculous meningitis may be preceded by a tubercle in the brain or its linings. However, during the six months of his clinical illness, the patient had no signs or symptoms to suggest tuberculous meningitis; he remained afebrile and had no signs of meningeal irritation. A lumbar puncture was contraindicated because of the severe mass effect seen on the initial magnetic resonance image. Thus, Blank's comments would not have helped establish a firm diagnosis earlier.

We agree with Azevedo and Andrioli that an earlier open cerebral biopsy would have been useful but disagree with their observation that stereotactic biopsy is a secondary and inaccurate technique. For the vast majority of lesions, the biopsy will yield tissue allowing an unequivocal diagnosis to be made,1 and stereotactic biopsy has repeatedly been shown to be safe and accurate.2-5 The errors made were not in the decision leading to the choice of technique but in the interpretation of the pathological description.

Ockham's razor is indeed a useful tool, but should never be understood as a logical necessity. Like other generalizations, it is sometimes right and sometimes wrong. Ockham's razor might have averted the presumption of two diagnoses, but that would still have left the patient with a pathological report of a malignant brain tumor when he in fact had a treatable infectious disease. Patients often do have more than one disease. In this case, the lung biopsy was negative for acid-fast bacilli, and a diagnosis of active tuberculosis was made only after an open brain biopsy. Even if one believed that the patient had pulmonary tuberculosis, it would have been a serious clinical error to use Ockham's razor to assume that this huge brain lesion must therefore be a tuberculoma, particularly in this country, where brain tuberculomas represent 0.15 to 0.18 percent of intercranial lesions and tuberculosis of the central nervous system accounts for less than 0.5 percent of cases of tuberculosis.6 As this case illustrates, the responsible physician needs to make a definite diagnosis, particularly when the means of establishing the diagnosis are available and specific therapy can be lifesaving.

Jeremy Brown, M.B., B.S.
Boston Medical Center, Boston, MA 02118

Michael J. Worthington, M.D.
St. Elizabeth's Medical Center of Boston, Boston, MA 02135

6 References

1. 1

   Kepes JJ. Pitfalls and problems in the histopathologic evaluation of stereotactic needle biopsy specimens. Neurosurg Clin N Am 1994;5:19-33
   Medline

2. 2

   Bouchama A, al-Kawi MZ, Kanaan I, et al. Brain biopsy in tuberculoma: the risks and benefits. Neurosurgery 1991;28:405-409
   CrossRef | Web of Science | Medline

3. 3

   Apuzzo ML, Chandrasoma PT, Cohen D, Zee CS, Zelman V. Computed imaging stereotaxy: experience and perspective related to 500 procedures applied to brain masses. Neurosurgery 1987;20:930-937
   CrossRef | Web of Science | Medline

4. 4

   Grunert P, Ungersbock K, Bohl J, Kitz K, Hopf N. Results of 200 intracranial stereotactic biopsies. Neurosurg Rev 1994;17:59-66
   CrossRef | Web of Science | Medline

5. 5

   Thomas DG, Nouby RM. Experience in 300 cases of CT-directed stereotactic surgery for lesion biopsy and aspiration of haematoma. Br J Neurosurg 1989;3:321-325
   CrossRef | Medline

6. 6

   Flannery MT, Pattani S, Wallach PM, Warner E. Hypothalamic tuberculoma associated with secondary panhypopituitarism. Am J Med Sci 1993;306:101-103
   CrossRef | Web of Science | Medline