Missed and Delayed Diagnoses in the Ambulatory Setting: A Study of Closed Malpractice Claims

While the report of Gandhi et al. (Ann Intern Med. 2006;145:488-496) represents an interesting and important descriptive study of a large series of malpractice cases, the conclusions which they draw with regard to "errors" should be tempered by methodologic flaws. In addition to the failure to blind reviewers to the outcome of the cases and other potentially prejudicial material in the legal files, the use of medical trainees to review cases and make gold-standard determinations of "error" and project the probable consequences of those "errors" is difficult to understand. Since malpractice actions generally involve competing expert interpretations of critical medical evidence, it would seem reasonable to use experts in that particular field to review those cases to arrive at an independent conclusion.

For example, in two of the vignettes presented in Appendix Table 1, alleged error in interpretation of an electrocardiogram (ECG) was a key component of each case. Misinterpretation of ECGs is not a rare event (1,2), even with board-certified attendings, and is likely related to degree of experience with and training in electrocardiography.

These issues beg the larger question which hampers the science of error reduction "“ how to define the term "error" in a way most health care providers can agree upon and how to apply that definition meaningfully to individual cases.

References

1. Bogun F, Anh D, Kalahasty G, Wissner E, Serhal CB, Bazzi R, Weaver WD, Schuger C. Misdiagnosis of atrial fibrillation and its clinical consequences. Am J Med 2004;117:636-642.

2. Knight BP, Pelosi F, Michaud GF, Strickberger SA, Morady F. Clinical consequences of electrocardiographic artifact mimicking ventricular tachycardia. New Engl J Med 1999;341:1270-1274.

Conflict of Interest:

None declared

To the Editor:

We applaud Gandhi et al.(1) for highlighting the problem of outpatient diagnostic errors. However, malpractice claims comprise a biased data source. Primary identification of diagnostic errors in ambulatory settings remains problematic.

Wachter highlighted differences that make diagnostic error detection in ambulatory settings more difficult than in hospitals.(2) That outpatient diagnosticians typically record their observations and conclusions incompletely and illegibly on paper charts contributes to the lack of knowledge about the extent of diagnostic errors. Paper-based outpatient records are expensive to collect and analyze for outcome studies. In outpatient settings, fewer diagnostic "gold standard" procedures occur, and irregular follow-up facilitates missing "sentinel" events. Incorrect outpatient diagnoses may go unnoticed for self-limited disorders. For more serious conditions, patients may present to serial physicians for "second opinions"; earlier physicians in the chain may never learn of a definitive diagnosis made elsewhere. As a result, available estimates of outpatient diagnostic error rates may represent lower bounds for their true frequency. Malpractice claims comprise an even smaller proportion of the total.

The suggestion of Gandhi et al. to use diagnostic decision support tools is a good one since these tools have been notoriously underutilized. Barriers to adoption of these tools have included the lack of impartial, non-proprietary, high-quality, comprehensive diagnostic knowledge bases and the time required to use such systems as "stand-alone" consultants during busy clinical practice.(3-4) Data transfer from electronic health record systems may address the problem of time-consuming data entry for diagnostic systems, but a bigger problem is that outpatient clinicians may not recognize any need for diagnostic assistance. Absent consistent feedback on the accuracy of their outpatient diagnostic decisions, it is not surprising that many clinicians seem overconfident in their diagnostic skills and fail to adopt decision support tools.(5) We concur with Gandhi et al. that use of clinical decision support systems should not depend a physician's perception that a case poses diagnostic challenges.

References

(1) Gandhi TK, Kachalia A, Thomas EJ et al. Missed and Delayed Diagnoses in the Ambulatory Setting: A Study of Closed Malpractice Claims. Ann Intern Med 2006;145:488-96.

(2) Wachter RM. Is Ambulatory Patient Safety Just Like Hospital Safety, Only without the "Stat"? Ann Intern Med 2006;145:547-9.

(3) Miller RA. Medical Diagnostic Decision Support Systems--Past, Present, and Future: A Threaded Bibliography and Brief Commentary. J Am Med Inform Assoc 1994;1:8-27.

(4) Miller RA, Giuse NB. The Medicine in Medical Informatics: Medical Knowledge Bases. Academic Medicine. 1991; 66:15-7.

(5) Graber, ML Diagnostic Error in Medicine - A Case of Neglect. Joint Commission Journal on Quality and Safety, 2005 31:112-9.

Conflict of Interest:

None declared

The article by Gandhi and colleagues (1) on malpractice cases is interesting reading but fails to help the primary care physician in his quest to maximize patient safety while preventing a suit for not doing so. While the authors lament the absence of a "silver bullet," I believe there is a simple one.

I propose full disclosure of every malpractice case to every physician in his county or state on a regular basis with information from the "claim file-a repository of information accumulated by the insurer during the life of claim." This would be in the form of an executive summary with the 6-point confidence scale and an attached expert opinion about how a claim might have been prevented.

We learn from case studies. When it comes to malpractice cases they are often sealed and we learn nothing. This is why history repeats itself in the form of 24% missed or delayed diagnosis of breast cancer.

References

1.Tejal K. Gandhi, Allen Kachalia, Eric J. Thomas, Ann Louise Puopolo, Catherine Yoon, Troyen A. Brennan, and David M. Studdert Missed and Delayed Diagnoses in the Ambulatory Setting: A Study of Closed Malpractice Claims Ann Intern Med 2006; 145: 488-496

Conflict of Interest:

None declared

We agree with Dr. Clairmont's suggestion that greater transparency around errors in malpractice claims would advance patient safety. If appropriately de-identified, such information could pull the liability system toward a broader culture of openness in which mistakes are seen as valuable opportunities to improve care, not problems to be hidden. All errors, regardless of whether they prompted malpractice claims, should be construed and discussed in this light. We are disappointed by Dr. Clairmont's view that our findings will not help primary care physicians. By highlighting several points in the diagnostic processes in which breakdowns frequently occur, and proposing several relatively "low tech" prevention strategies, we believe the research provides some practical guidance for clinicians.

Dr. Marine makes several reasonable methodological criticisms of our study. It was not possible within available study resources to purge the numerous pages of documentation in the claim files of all references to the litigation outcomes, so reviewers may have been aware of them. The likely effect of this knowledge would be to make claims that attracted payments more likely to be judged errors, and vice versa. Senior residents or fellows reviewed approximately one quarter of the claims, and their detection rate did not differ from that of more senior reviewers. In previous large-scale studies of adverse events (1,2), we have found the quality of reviews by upper-level trainees and attending physicians to be similar. Finally, better agreement over what constitutes error is certainly needed (3). The World Health Organization's ongoing work in this area should be applauded (4). But more sophisticated definitions and classification tools cannot avoid the complex questions of causation and appropriateness that surround errors of omission, such as missed diagnoses; they will remain intrinsically difficult to identify reliably.

We agree with Drs. Berner, Miller, and Graber that malpractice claims are a biased source of data on medical errors, but it is important to consider what effect those biases are likely to have on etiologic analyses. Many of the specific concerns mentioned are problems for a study aimed at estimating an error rate. However, our focus was causal factors. As we note in the paper, some factors, such as fatigue, may have been routinely undercounted because claim file documentation is not well- suited to capturing them. (This is a problem for any retrospective review of records.) Consequently, the prevalence estimates for some of the causal factors are likely to be lower bounds, and the multi-factorial nature of diagnostic error depicted by our findings is probably an understatement of their true complexity.

Tejal K Gandhi, MD, MPH Brigham and Women's Hospital Boston, MA 02115

Allen Kachalia, MD JD Brigham and Women's Hospital Boston, MA 02115

David Studdert, LLB, ScD Harvard School of Public Health Boston, MA 02115

NOTES

1. Brennan TA, Leape LL, Laird NM, Hebert L, Localio AR, Lawthers AG, et al. Incidence of adverse events and negligence in hospitalized patients. Results of the Harvard Medical Practice Study I. New Engl J Med 1991;324:370-376.

2. Thomas EJ, Studdert DM, Burstin HR, Orav EJ, Zeena T, Williams EJ, et al. Incidence and risk factors for adverse events and negligent care in Utah and Colorado in 1992. Med Care 2000;38:261-271.

3. Thomas EJ, Petersen LA. Measuring errors and adverse events in health care. J Gen Intern Med. 2003;18:61-67.

4. World Health Organization. More than words: A patients safety taxonomy. http://www.who.int/patientsafety/taxonomy/en/ (Accessed December 4, 2006)

Conflict of Interest:

None declared

Reasonable waiting time for both operating room procedures and ambulatory referrals is a subject of increasing interest in political and social settings (1,2). In orthopedic surgery the delay to treatment seems related to a reduced recovery of function (3). However still few experimental works are able to put in evidence the better strategies to obtain timely procedures for all waiting patients in welfare countries (4). As asserted by Ghandi et al., delayed diagnoses in the ambulatory setting are an important patient safety problem and furthermore this type of error remains largely unstudied. From several years, we are moving towards the use of explicit clinical priority access criteria named homogeneous waiting groups (HWG or RAO), briefly presented some years ago (5). Our guidelines with the priority categories and the related waiting times are similar to that exposed by New Zealand Ministry of Health on his web site (www.electiveservices.govt.nz). By means of a lasting alliance between hospital specialists and general practitioners (GP), we have got some preliminary but interesting results putting in evidence the ability of GP to attribute major priority to patients affected by significant pathologies. In particular, gastroenterologists (GE) analysed with explicit criteria that ability. In the light of guidelines, GE evaluated if appropriateness of referrals and/or the presence of significant endoscopic diseases (SED) are correlated to the priority assigned by GP to patients booked for gastrointestinal procedures (gastroscopy and colonscopy). Our more recent evaluation included 600 patients successively booked and undergone a gastroscopy or a colonscopy. Explicit criteria of appropriateness were referred to ASGE guidelines (6) and were matched to each patient before the procedure began. GP routinely used a rank of waiting time for different level of clinical priority (RAO). Results confirm that there is a significant correlation between proportion of SED and the attribution of priority to referral (47 RAO vs. 84 noRAO in patients with SED and 80 RAO vs. 389 noRAO in patients without SED; X2 = 21.74, 1 df, p « .001) and there is a significant correlation between appropriateness of referral and the presence of SED (126 appropriate vs. 5 no appr. in patients with SED and 302 appr. vs. 167 no appr. in patients without SED; X2 = 50.61, 1 df, p « .001). Further research is required on evaluate the relationship of referral appropriateness and clinical priority ranking to predict the presence of significant gastrointestinal diseases. Actually we are exploring all those variables in each referral (above all clinical priority ranking) with the hope to keep up the high value of professional involvement and continuous clinical monitoring to improve the outcome of our diagnostic path. Our preliminary experience highlights the high value of alliance between GP and specialists as a empirical answer to get timely and appropriate referrals and hence to avoid delayed diagnoses in ambulatory setting. REFERENCES

(1) Coulter A., Ham C.: The global challenge of health care rationing. Buckingham, OUP, 2000.

(2) Appleby J., Boyle S., Devlin N. et al.: Sustaining reductions in waiting times: Identifying successful strategies. London, King's Fund, 2005.

(3) Hajat S., Fitzpatrick R., Morris R. et al.: Does waiting for total hip replacement matter? Prospective cohort study. J Health Serv Res Policy, 7 (1): 19-25 (Jan), 2002.

(4) WCWL (Western Canada Waiting List) Project: Moving forward. Final Report. University of Calgary, Alberta, February 28, 2005.

(5) Mariotti G., Sommadossi R., Langiano T., Raggi R., Letter, BMJ, 318: 1698-9, 1999.

(6) ASGE, Gastrointestinal Endosc, 52: 831-7, 2000.

Conflict of Interest:

None declared

I was surprised to see that workload and fatigue combined to give a mere 8% rating as causative factors in missed diagnoses in ambulatory care settings while cognitive factors like judgment, vigilance, and lack of knowledge accounted for a 99% rating.

I bring this up because when primary care doctors are overworked, they are more likely to experience impaired judgment or see the ill effects of their lack of knowledge than when their workload is manageable. Doctors can always consult with a colleague when in doubt or go to a good medical text when they have sufficient time and are not fatigued,

Workload is underestimated as a factor in missed diagnoses, particularly for primary care doctors who are pressured to see large numbers of patients in order to generate an income that allows them to survive. It would be helpful to see a similar study done but this time asking the physicians involved to determine what they considered the causative factors in missed diagnoses to be. I suspect that workload would be given significantly greater importance.

It is also interesting to mention that qualities like judgment, vigilance, workload, and fatigue are very difficult to measure on written or computerized exams.

Conflict of Interest:

None declared