Negligent Interpretation of Pap Smears
Published by the American Society of Cytopathology in their Journal, Pathology Patterns Reviews.
Cervical cancer continues to kill approximately 4900 women annually. This is a particularly alarming statistic to the 15,700 women who are annually diagnosed with new cases of cervical cancer.
Each year many women diagnosed with cervical cancer, or their significant survivors consult lawyers to determine whether the stage of the cervical cancer at the time of diagnosis and the related morbidity and mortality could and should have been avoided. According to Dr. Scott, examining pathology claims at one insurance carrier corresponding to the period of July 1, 1993 to August 10, 1995, 17% of the claims involved Pap smear interpretations as the central issue. The dollar value of the claims was estimated to represent 25% of all new pathology claims during the measured interval. Though an examination of the claims for any interval of any given insurance carrier is not necessarily reflective of national or regional trends, there is no question that it is the common experience of knowledgeable physicians and lawyers that Pap smear cases represent significant liability exposure for health care providers.
The number of claims made involving Pap smears represents, at best, a very small percentage of the total number of false-negative cases which have, in fact, resulted in needless delays in diagnosis and treatment and resultant excess morbidity. This is obvious when one considers that the false-negative rate of a single Pap smear is widely quoted to be 20%. . Some maintain that more recent data actually indicate the sensitivity of a single Pap smear in women undergoing routine screening may be closer to 50%.
Koss has asserted that invasive cancer is rarely preceded by truly negative serial Pap smears when available for the 2 to 3 years prior to the diagnosis of invasive cancer being made. My experience is consistent with that of Dr. Koss.
In 1987, Van der Graaf, Y., et al., reported striking findings concerning screening errors in cervical cytoloty. In a screened population of 165,185 women, 555 women demonstrated "moderate dysplasia" or a higher lesion, three years after a negative screening. An examination of the slides previously read as negative showed that the adequate smears all were misread. Fully 1/3 represented "dysplasia" or a carcinoma in situ. According to Dr. Koss, in those cases contained in the literature where prior negative smears of patient with invasive cancer are examined, cancer cells are apparent on review.
In 1961, at a time when few American women were being screened, the incidence of cervical cancer was 33 per 100,000. By 1987, when the population of women screened at some time during their lives reached 80%, the incidence of squamous cell carcinoma of the cervix dropped to 8.3 per 100,000. The rate of cervical cancer, in countries where screening does not occur, continues at the same rate as that of industrialized countries before the initiation of large-scale screening programs. It is obvious that screening programs have greatly reduced the incidence of the disease. However, the continued occurrence of cervical cancer in the screened population suggests that inadequate screening occurs or that the screening process is inherently incapable of identifying a significant percentage of women with disease prior to the occurrence of invasive cervical cancer. It is this author's position that the occurrence of 8.3 cases per 100,00 is due to inadequacies in the screening process. That inadequacies exist is well borne out by the literature. It is these inadequacies which are the subject of this paper.
The object of reducing the occurrence of false-negative Pap smears has been a major focus of legislative reform and of ongoing quality control programs throughout the country. Whether the 4900 annual excess deaths from cervical cancer are subject to further reduction has been questioned by some authors recently. Frabel and others focus primarily on the perceived plight of the cytotechnologist, cytopathologist or laboratory, or the assumed increased cost which would result from any significant improvement to the existing system. , DeMay argues that cytologists are being held to an impossible standard of perfection in malpractice litigation and that a zero error standard is both unreasonable and impossible to achieve in practice. The legal system does not demand perfection. The legal system does demand that cytotechnologists and cytopathologists perform in conformity with those principles of practice about which there is uniform agreement. From a legal perspective, a zero error standard is reasonable, whether or not it is possible to achieve in practice. All standards of practice must be at least theoretically obtainable. If it were theoretically impossible to perform in a certain manner in practice, such performance would not be a standard to which any professional would be held. We do not ignore in our culture the incidence of carelessness simply because it is supposed that human beings will, on occasion, be careless.
Though there are a variety of diagnostic errors which occur due to negligence in the processing and interpretation of Pap smears some misinterpretations are unavoidable. Assume that a gynecologist properly obtains a Pap smear that is then properly applied to a slide and fixed. Though the patient, at the time is suffering from a pre-invasive lesion of the cervix, so few abnormal cells are present in the smear that they are unlikely to be located during a careful screening of the slide. The delay in diagnosis that results from this unrevealing smear is unavoidable and undoubtedly occurs. However, in the author's experience the patient presenting with advanced cervical cancer, following serial purportedly benign annual Pap smears has been the victim of professional carelessness. Review of slides, in such cases, reveals inadequate smears which were not reported, abnormal cells that were either identified but then ignored by the physician performing the Pap smear or were misinterpreted or not located by the cytopathologist or cytotechnologist examining the smear.
The devastation that results from a needlessly advanced case of cervical cancer, a disease, which most believe is largely preventable, is tragic. It is the obligation of the lawyer representing claimants in such cases to obtain through the legal system what compensation is there available. However, the legal process also can reveal the causes of a client's injuries and properly identify those persons responsible. Ideally, the process of litigation should inform those persons responsible of their errors and encourage them and others to act more responsibly in the future.
The cytopathologist can have a far-reaching role in reducing morbidity due to cervical cancer. This role is not limited to interpreting properly Pap smears submitted for review. The cytopathologist also has a responsibility to inform those submitting smears of the adequacy of the smears and to submit reports in a comprehensive and comprehensible form, together with recommendations for future management (diagnostic evaluation).
The following case, recently concluded, embraces virtually every breach of principles of practice for which a cytopathologist and laboratory can be held liable in the processing of Pap smears.
In October 1991, a 25-year-old patient completed her eighth office visit with a gynecologist since October of 1988. During this course she was attended by three different offices of gynecologists (each in a different office). Neither of her last two gynecologists requested records from the immediately preceding gynecologist. HGSIL was reported on a Pap smear in October of 1988 and colposcopy was recommended. The patient's second gynecologist, though aware that a prior Pap smear was abnormal and had resulted in a previous recommendation for colposcopy, instead performed cryotherapy. He did not perform a colposcopy until February 1990. In spite of an aceto-white lesion of the cervix being present at the time of that colposcopy, no biopsy was taken. Soon after, gynecologist "2" retired.
In June 1990, Gynecologist "3" received a history of abnormal Pap smears and a colposcopy and regrettably assumed, without asking, that the colposcopy which had been related to him historically had been accompanied by a biopsy and appropriate follow-up (in fact, this had not occurred).
In October 1991, Gynecologist "3" submitted a Pap smear to a laboratory that had never previously reviewed cytology or tissue of this patient. The Pap smear was submitted with the requisition shown in Figure 1. The gynecologist did not fill out this requisition and the clerical personnel of the gynecologist's office who completed the requisition did not secure accurate history from the patient's chart. Therefore, the clerical personnel erroneously recorded that there had been a previous Pap smear in June of 1990, which was negative. In fact, there was a Pap smear in June of 1990, which was interpreted as "ASCUS seen in a less than optimal specimen due to the absence of endocervical component." The clerical personnel also failed to record upon the requisition that there were obvious historical risk factors for cervical cancer.
The Pap smear specimen was routinely screened by a cytotechnologist, who interpreted the Pap smear as ASCUS. This technologist, unaware, because of the deficient requisition, of the patient's high-risk status, exercised discretion and did not request rescreening by the cytotechnologist supervisor or a pathologist. Instead he issued a report containing no recommendation as to follow-up.
Though Gynecologist "3" received the report indicating ASCUS and recognized the need for a colposcopy, colposcopy was not performed. In December 1992, the laboratory received from Gynecologist "3" a further Pap smear specimen of this patient accompanied by the requisition illustrated in Figure 2 . The form of requisition illustrated in Figure 1 had, by this time, been modified by the service laboratory . The earlier requisition (Figure 1) was interpreted by the laboratory involved as negative for risk factors only if the requester checked "no" for "risk factors present for cervical cancer." If that area of the form was blank, it was the policy of the laboratory (honored chiefly by its breach) to regard the requisition as deficient and lacking needed historical information. The laboratory supervisor became concerned that so many requisitions were left blank under "patient history" and that it would appear that they were rendering interpretations on many slides without adequate history. The requisition was therefore changed so that, if left blank, the form was interpreted as affirmatively declaring that the history was negative. This remarkable logic obviated the need to contact the gynecologist's office to secure a history when the question was left blank.
Does the blank square before risk factors for cervical cancer seen in Figure 2 relieve the lab of responsibility to make further inquiry of the gynecologist's office concerning history. As defined in the Bethesda System, a cervical smear is a medical consultation and, as such, knowledge of clinical findings is essential to the proper screening and interpretation of cytologic material.
How vigorously must a laboratory pursue clinical data? Certainly a laboratory has a responsibility to secure historical information from its own prior patient records. The implementation of CLIA 1988 requires cytologic slides must be retained for no less than five years from the date of examination. All non-negative specimens should be maintained in one file within the service laboratory allowing retrieval of slides for review or study at any time. A cytotechnologist or pathologist when interpreting a specimen must be aware of the particular patient's record within the service laboratory, which should be accessible either by a fully automated record-keeping system or by hardcopy search of files at the time a specimen is being interpreted.
Therefore, the interpreter of a Pap smear is not relieved of a responsibility for acting in ignorance of important historical or other data readily ascertained from the patient's record within the service laboratory, even if misleading information is contained within the submitted requisition. Furthermore, a laboratory shares with its client gynecologists responsibility for establishing effective communication. The gynecologist must provide detailed and comprehensive clinical data. However, the laboratory cannot avoid legal responsibility for acting upon inadequate history. This was particularly true, in the illustrated case, where the administrator of the laboratory was forced to concede, upon questioning, foreknowledge of casually prepared and inaccurately encoded requisitions.
Ironically, the laboratory in question had history in its own files concerning the involved patient as serial slides of the particular patient had been reviewed by this laboratory. Though computers in the laboratory were used for billing, as a matter of policy, they were not used to identify patient records. Even though this patient's previous Pap smear had in fact been interpreted as showing ASCUS, or higher abnormality, her slides were each read routinely as if her history had been negative. The error of the gynecologist's clerical personnel in reporting an inaccurate history was thereby compounded. When questioned by deposition, neither the gynecology personnel involved, nor the laboratory personnel involved, were sufficiently aware of the practices of each other's organizations to recognize how easily miscues could occur.
The Pap smear specimen of December 1992 was interpreted as satisfactory and within normal limits. The specimen was not rescreened. Again the accompanying requisition contained inaccurate information. The previous Pap smear of October 1991 had in fact not been normal, but had shown ASCUS. In fact, the patient had historically been subject to cryosurgery and had other risk factors for cervical cancer. A simple review of the service laboratory records for this patient would have revealed that there was a previous abnormal smear and, given the policies of the service laboratory, the specimen would have automatically been rescreened. This did not occur.
Figures 3, 4, 5 and 6 represent photomicrographs of the slide which corresponds to the requisition shown in Figure 2. Figure 3 presents a panoramic view photographed at low power. Near center is a flat cluster of cells demonstrating HGSIL. Figure 4 is a higher magnification of the flat cluster of cells marked in Figure 3. Figure 5 demonstrates another cluster at high magnification of HGSIL cells with overlapped nuclei and indistinct borders (syncytium) surrounded by LGSIL cells. Figure 6 is an HGSIL cell with large dark irregular nucleus and high N/C compared with adjacent normals.
The photomicrographs shown and discussed here are but a few of those taken of the Pap smear specimen of December 1992. Each of these photomicrographs when shown during deposition to the cytotechnologist (now a defendant in the case) resulted in her substantial agreement with the interpretations described above. Though similar abnormal cells were numerous in the specimen, the defendant cytotechnologist asserted she obviously had not seen these cells.
There was some question arising out of the testimony in the case as to whether the cytotechnologist's problem had in fact been a failure to locate the cells or failure to properly interpret them. Approximately a year and a half after the interpretation of the slide in question, this cytotechnologist was discharged by her laboratory. This occurred because she interpreted as negative a slide of another patient that was rescreened as showing a HGSIL suggesting invasive cancer in a patient who was later demonstrated by biopsy to have invasive cancer. The cytotechnologist's discharge resulted from her failure, when confronted with the discrepancy, to accept that her reading of the slide as negative was inappropriate.
In this case the service laboratory had accumulated, as required by CLIA, statistical information concerning the performance of the defendant cytotechnologist and others compared to the laboratory as a whole. Over the course of 2-1/2 years of employment, the defendant cytotechnologist consistently found squamous intraepithelial lesions at a rate far below the laboratory average by a difference sometimes as great as 5 or 6 to 1. When questioned, the physician director of the laboratory, who had not recently reviewed the data, testified with a clear memory that the defendant cytotechnologist's performance had not only been satisfactory but that the director was unaware of a single occasion when the Defendant cytotechnologist had under-read a slide. In view of the statistical analysis it is clear that the director's recollection was faulty, but it is also perhaps not a coincidence that, in this laboratory, pathologists rescreened few of defendant technologist's slides sometimes averaging less than 4% per month.
Ironically, the Defendant cytotechnologist could have preserved her position by simply agreeing she made a mistake. By this laboratory's policies, no retraining or testing was required for misreading a single slide. A laboratory cannot wait for a pattern to develop where a cytotechnologist has missed a high-grade lesion. The pattern already evident statistically raised question concerning the ability of the cytotechnologist to identify and classify neoplasia. This should have early led to a review of her negative slides and discovery of her inadequacies. A disaster would have been avoided.
The patient died in October 1997 survived by three minor children, one disabled, all orphaned by her death. This patient's grueling struggle with invasive cervical cancer is a tale not often told but occurring with regrettable frequency throughout this country. The diagnosis of the patient's cervical cancer was made in March 1994. The patient was constantly under the care of a gynecologist who she was visiting on a regular and frequent basis, had been subjected to seven Pap smears between October 1988 and December 1992, and had received a colposcopy (1990) for a visible lesion and no biopsy. It is obvious from the facts related that the management by the patient's gynecologists was far less than required by accepted standards of practice. However, it is important to note that in December of 1992 the patient stood a substantial chance of a better outcome if the Pap smear had been properly interpreted.
The mistakes made by the service laboratory in the reported case are illustrative of other cases in this author's experience. Administrative negligence allowed a cyhtotechnologist experiencing declining skill to deteriorate to a point where she became increasingly dangerous. The laboratory trivialized the function of its cytotechnologists and cytopathologists and consultants by minimizing the importance of clinical data.
All laboratories will benefit from a re-examination of their policies to assure they:
1. Guarantee communication of vital historical and clinical data between the clinician and the service laboratory
2. Issue no report on a Pap smear specimen if necessary historical data has not confidently been obtained
3. Issue no report on a Pap smear specimen without a concurrent review of the patient record within the service laboratory and a comparison of slides, if indicated
4. Utilize the accumulation of statistical data concerning the interpretation of Pap smears on a monthly basis to identify and act upon areas of concern
Reduction in morbidity and mortality due to cervical cancer has been a miracle of this century to which cytopathology has made a decisive contribution. New technologies have been suggested to reduce the false-negative risk further such as thin layer preparations and selected hybrid capture testing for HPV DNA. Nevertheless, significant reductions in true-false negative screening can be wrenched from the existing technology. Concerns about liability exposure should not lead, as it has, to the promulgation of literature suggesting that the interpretation of cytology specimens is an artistic endeavor without a scientific basis. The scientific basis of cytologic interpretation, evolved laboriously over decades, creates the framework within which all instruction in this field is carried out. Academic nihilism suggests it is not possible in a particular given instance to correctly interpret a slide by resort to accepted principles of interpretation. It is human nature that drivers of cars will sometimes go through red lights and stop signs, cross center lines, tailgate, etc. But the inevitability of such carelessness does not mean that on each and every occasion where one goes through a stop sign that they should not be held responsible for the error where that error harms another.
Every time that a screener of a Pap smear is distracted and looks without seeing, we should examine the causes of the distraction and find ways to avoid such vacant observations rather than asserting that to look without seeing is the standard by which practice should be measured. If multiple individual cells are present with high nuclear cytoplasmic ratios, hyperchromasia, indistinct cell boundaries, and nuclear chromatin clumps, surely a tissue examination is warranted. In many academic centers ASCUS is a basis for colposcopy. For two or more ascus reports colposcopy is required.**[footnote required-American College of Obstetricians and Gynecologists: Cervical Cytology: Evaluation and Management of Abnormalities. Technical Bulletin no. 183, August 1993]** The distinction between whether a lesion is a low grade lesion which may disappear or a high grade lesion which will likely not disappear, should not be a distinction used to justify a classification system or practice which prevents a more accurate diagnostic procedure from being secured.
In criticism of the zero error standard, several authors have suggested that inter-observer variation is a justification for asserting that since there are differences of opinion about the precise diagnosis to be made on a particular cellular specimen, that no diagnosis made of the specimen is an error. If there is agreement that a Pap smear is abnormal, consensus as to the degree of abnormality is not critical. It is the failure of a given observer to recognize an abnormal smear is present which is the gravest of errors. Such errors need our fullest attention and introspection because it is precisely these errors that take patients out of the mainstream of continued clinical evaluation and lead to needless delay and destructive consequences. Thus, assuming that all women will undergo repeated cytologic screening and assuming that adequate material is obtained, correctly interpreted and correctly handled by the clinicians, one can anticipate a significant further decrease in the incidence of carcinoma of the uterine cervix and ultimately, even if there is no further therapeutic progress, a disappearance of carcinoma of the cervix as a cause of mortality.
Click here to view requisition for Pap Smear interpretation.
Figure 3 (click to return to paragraph)
Figure 4 (click to return to paragraph)
Figure 5 (click to return to paragraph)
Figure 6 (click to return to paragraph)
