November 17, 2008
Comparing New Radiographs With Those Obtained Previously
By Leonard Berlin, MD, FACR
Vol. 9 No. 23 P. 48
Editor’s Note: Leonard Berlin, MD, FACR, is a professor of radiology at Rush University Medical College and chairman of the department of radiology at Rush North Shore Medical Center in Skokie, Ill. He began his writing on risk management and malpractice issues in a series of articles in the American Journal of Roentgenology. Those articles became the basis for his well-known book, Malpractice Issues in Radiology. This column is drawn from that book.
The third edition of Malpractice Issues in Radiology is scheduled for release this fall and will be available from the American Roentgen Ray Society at RSNA 2008.
A 57-year-old woman with a recent onset of cough and low-grade fever was referred by her family physician to the radiology department of a hospital for chest radiography. A radiologist interpreted the posteroanterior and lateral chest radiographs as disclosing a “subtle ill-defined infiltration in the left mid-lung field, seen only on the PA view, consistent with pneumonia” (Figs. 1A and 1B). The physician treated the patient with antibiotics, and her symptoms resolved. No immediate follow-up radiographs were obtained.
Three months later, the patient again consulted her family physician because of recurrence of cough and fever. After establishing a tentative diagnosis of pneumonia, the physician referred the woman for another radiographic examination. On this occasion, a radiologist who was associated in practice with the radiologist who had interpreted the study 3 months earlier reported seeing a “fairly well-circumscribed infiltration in the left lower lobe, most likely representing pneumonia” (Fig. 1C). The radiologist made no mention in his radiographic report of a comparison with the previous radiographs. The patient was again treated with antibiotics by her family physician, and her symptoms improved.
The woman did not return to her family physician until 9 months later, at which time she was experiencing hemoptysis and weight loss. She again underwent radiography of the chest. The same radiologist who had interpreted the second set of radiographs now reported that the left lower lobe infiltration had increased in size and, because of its “well-circumscribed appearance” and the fact that the left hilum appeared enlarged, concluded that the “findings are suspicious for tumor” (Figs. 1D and 1E). Additional diagnostic testing revealed primary lung carcinoma with metastases to the mediastinum and brain. Chemotherapy was begun, but the patient died 4 months later.
Just before her death, the woman filed a malpractice suit against the family physician and the radiologist who had interpreted the second and third sets of radiographs, claiming that her chance for cure was lost because both physicians had negligently delayed the diagnosis of lung cancer.
In the legal complaint filed by the attorney for the plaintiff, the radiologist who had interpreted the second and third sets of radiographs was accused of breaching the standard of care because he had not compared the second set of radiographs with the original radiographs that had been obtained 3 months earlier. The complaint charged that if a comparison had been made, the radiologist would have “easily recognized” that the abnormal density in the left lung was the same density, although larger, that was present on the initial radiographs; would have suspected a malignancy; and would have recommended additional studies that, in turn, would have led to a more timely diagnosis of lung cancer and to institution of curative treatment.
In a deposition conducted during the discovery period, an expert radiologist retained by the plaintiff characterized the defendant radiologist’s failure to compare the second set of chest radiographs with those of the initial examination as a “clear breach” of the standard of radiologic care. He used as the basis for his opinion the American College of Radiology (ACR) Standard for Communication: Diagnostic Radiology, which states, in part, “Comparisons with previous examinations and reports, when possible, are a part of the radiologic consultation and report.” The expert for the plaintiff also referred to the ACR Standard for the Performance of Pediatric and Adult Chest Radiography, which states, in part, “It is important that new films be compared with prior chest examinations and/or other pertinent studies that may be available.”
In his deposition, the defendant radiologist was asked by the attorney for the plaintiff why he had not compared the second set of radiographs with the initial study. He responded that the previous radiographs were stored in the basement of the hospital and that it would have taken “nearly a day” for them to be brought to him. He explained that because he believed the new radiographs indicated acute pneumonia, he believed it was imperative to relay that information to the referring physician immediately, so that appropriate treatment could be instituted. The defendant radiologist added that he had read the previous radiography report, and because it described pneumonia in the left mid-lung field rather than in the lower lobe, he had assumed that the original pneumonia had resolved. The radiologist explained that he had thus concluded that it was not necessary to make a comparison with the earlier set of radiographs. When asked whether he believed he had complied with ACR standards, the defendant radiologist responded in the affirmative, emphasizing that the insertion of the words “when possible” into the communication standard, and the use of the noncommittal phrase “it is important” in the chest radiography standard, indicated to him that the directive to compare current radiographs with previous ones was merely a recommendation rather than a requirement.
An expert radiologist retained by the defense agreed with the defendant radiologist. When questioned in depth about the ACR standards, the expert pointed out that the standard for adult chest radiography used the word “should” in certain statements and “may” in others. It was his opinion that “should” signifies a mandatory directive, “may” an optional one. The expert reasoned that because the portion of the standard that deals with comparison simply states, “it is important that… films be compared…”, it is a directive that is neither mandatory nor optional; thus, the standard does not require radiologists to make comparisons. As for the ACR standard for communication, the expert also concurred with the opinion of the defendant radiologist, emphasizing that the inclusion into the standard of the words “when possible” allows radiologists some latitude in determining when comparisons must be made.
The expert radiologist for the defense then raised a point that he believed bolstered the argument that the defendant radiologist had not breached the standard of radiologic care. He pointed out that the specific version of the ACR Standard for Communication: Diagnostic Radiology quoted by the expert radiologist for the plaintiff should not apply to the alleged act of malpractice committed by the defendant radiologist because that version had been adopted by the ACR in 1995, 1 year after the second set of chest radiographs had been interpreted. Therefore, the defense expert asserted, the correct version of the standard against which the defendant radiologist’s acts should be measured was the original version, first adopted by the ACR in 1991. Regarding comparisons with previous radiographs, the 1991 standard stated that the comparison may “optionally” be part of the impression section of the radiology report. The reference to optional comparison was removed from the 1995 revision of the communication standard, but the defense expert testified that the removal occurred after the alleged act of malpractice. He thus suggested that the plaintiff’s basis for charging the defendant radiologist with violation of the ACR standard was even further weakened.
As the malpractice trial date neared and each side was preparing its case, an event occurred in a Chicago courtroom that convincingly moved the defense team toward a negotiated settlement. There, on March 26, 1997, a jury [in Illinois] found two radiologists liable for failing to provide a timely diagnosis of lung cancer from the chest radiographs of a 30-year-old woman [Blankshain v Radiology and Nuclear Consultants Ltd]. During the trial, this plaintiff had claimed that the two radiologists incorrectly diagnosed pneumonia on her chest radiographs because the radiologists had failed to compare those radiographs with chest radiographs obtained previously. During the trial, the two radiologists insisted that they were under no obligation to seek out prior radiographs. However, the jury found that the two radiologists did indeed have the duty to compare new with previous radiographs and that their failure to do so caused them to miss the correct diagnosis of lung cancer. The jury awarded the plaintiff $4.5 million.
Upon the recommendation of the defense attorneys, and with the reluctant agreement of the radiologist and family physician, the insurance company settled the case for $1.5 million, apportioned equally between the two defendants.
As I have pointed out [in “Standard of Care”], what constitutes an appropriate standard of radiologic care in a given case of medical malpractice litigation is established by the testimony of expert witnesses. Although in years past these experts based their opinions primarily on their own subjective knowledge and experience, in recent years experts have to an increasing extent relied on guidelines or standards that have been formulated by professional organizations. Thus, standards published by the ACR now often play a major role in the determination of liability in medical malpractice cases involving radiologists.
Inasmuch as new ACR standards are continually being issued and those already in existence are constantly being revised, it is important to determine which particular standard or revision of a standard was in effect at the time of, and therefore applicable to, a specific alleged act of negligence. For example, as was accurately brought out by the defense expert in the case described in this article, the ACR Standard for Communication: Diagnostic Radiology was originally adopted in 1991 but was revised in 1995. In theory and from a legal point of view, therefore, the determination of whether the conduct of a radiologist was negligent should not be influenced by the more recently adopted standard if the event in question occurred before 1995. Determination of negligence would be influenced by the 1991 original version of the standard if the event occurred between 1991 and 1995, and determination of negligence would be influenced by the 1995 revised standard if the event occurred after 1995. Of course, some ACR standards have undergone more than one revision. For example, the ACR Standard for the Performance of Adult Chest Radiography referred to in this article was originally issued in 1993 and has already undergone two revisions.
To date, no state or federal appeals court has ruled that ACR standards published after an alleged act of malpractice cannot be used retroactively to determine whether professional conduct occurring before the effective date of the revised standard did in fact violate the standard of radiology practice. Therefore, when the issue of standards is raised in a specific malpractice trial, the defense attorney generally files a request for the trial judge to rule that any reference to ACR standards adopted after an alleged act of malpractice is inadmissible. Although judges usually rule favorably on these requests, judges are not compelled to do so and may at their discretion admit into evidence testimony related to later standards. In this situation, a defense attorney must then attempt to convince a jury that determination of the appropriate standard of radiologic care applicable to the defendant radiologist should not be influenced by formal standards published after the fact (Grossman FJ, personal communication).
Notwithstanding the meticulous dissection and convoluted word analysis of various versions of the ACR standards for communication and chest radiography offered by the expert radiologist for the defense in the case described, who attempted to support his opinion that the standards do not require an interpreting radiologist to compare new with previous radiographs, I believe that it is difficult to justify the failure of any radiologist to make such a comparison. The general acceptance by the radiology community at large that the standard of radiologic care calls for comparing new radiographs with previous examinations is based on far more than the formal standards published by the ACR. The radiology literature is replete with references that support comparison. For example, Hunter and Boyle stated in 1988 [in “The Value of Reading the Previous Radiology Report”] that “all radiologists believe it is important to review a patient’s previous radiographs, at least the latest ones, before officially interpreting a study.” In 1994, White et al. [in “The Role of Previous Radiographs and Reports in the Interpretation of Current Radiographs”] analyzed the role of previous radiographs in the interpretation of current unenhanced radiographs and found that previous radiographs were judged valuable in interpreting current studies in 89% of cases.
The importance of obtaining previous mammograms to compare with new studies is less clear and is, indeed, more controversial. Brenner1 believes that “the duty to retrieve old films likely falls under a ‘reasonable’ standard.” Logan-Young et al.2 state that previous mammograms should always be obtained because “you never know whether you need the old films until you put them up on the viewbox and scrutinize them.” Hall et al.3,4 disagree, claiming that obtaining previous mammograms is not necessarily required and that the decision about whether to obtain them should be left to the interpreting mammographer. In fact, Hall et al. suggest that it is not always necessary to obtain previous chest radiographs if the current chest radiograph is normal. Bassett et al.5 echo this sentiment, for they found after reviewing 1,432 randomly selected screening mammograms that comparison with previous studies had a positive impact on clinical management and cancer detection in only 3% of cases. The ACR Standard for the Performance of Screening Mammography seems to take a middle-of-the-road stance, stating simply, “Prior mammograms should be obtained when practical.”
Reviewing previous radiology reports in addition to reviewing the radiographs themselves, or in the absence of those radiographs, is also recommended. Hunter and Boyle found that reading a previous report was useful in 60% of cases and helped significantly in another 24% of cases. Likewise, in his classic book on radiographic errors, Smith6 emphasized that failure to compare a current radiograph with a previous one is an important and common cause of radiographic error. However, Smith also directed the attention of the physician to another common type of radiographic error, one that he called the alliterative error. It is a phenomenon caused by the influence that one radiologist can exert on another. Smith pointed out that if one radiologist fails to detect and report a radiographic abnormality, the chance that a second or subsequent radiologist will also miss the same abnormality is increased. Smith explained that alliterative errors more frequently occur when radiologists read the reports of previous examinations before looking at a new set of radiographs, thereby concentrating their attention on areas of the radiographs that may not harbor a new abnormality. Smith also believed that an interpreting radiologist’s perception of a radiographic abnormality is influenced by the reports rendered by radiology colleagues who have interpreted previous studies of the same patient.
Circumstances may exist, such as when it is noted that a lesion was missed on a previous study, under which radiologists may choose deliberately not to compare new with previous radiographs. Radiologists who elect not to disclose a radiologic missed diagnosis by omitting documentation of comparison with previous radiologic studies may escape legal exposure or criticism. However, such radiologists should be aware that omitting comparisons for this purpose violates the spirit, if not the letter, of ACR standards. The preponderance of legal opinion and ethical commentaries favors comparison with previous studies whenever possible, even if such comparison results in disclosure of errors committed by the interpreting radiologists themselves or their colleagues.7
Summary and Risk Management
On the basis of the preponderance of articles and commentaries published in the radiology literature, the formal standards issued by the ACR, and an informal consensus among practicing radiologists, the radiology community generally accepts that the standard of radiologic care requires that, when rendering radiographic interpretations, radiologists compare new radiographs with those obtained previously. Any delay in establishing the diagnosis of serious illness that can be attributed to a radiologist’s failure to compare current with prior radiographs may generate a medical malpractice lawsuit and can result in a finding of liability against the radiologist. Risk management regarding comparison with previous radiographs can lessen the likelihood of incurring a medical malpractice lawsuit and maximize chances for a successful defense if such a suit is filed. Risk management can also enhance good patient care. The following risk management pointers will help radiologists meet all three of these objectives:
• When interpreting general radiologic studies, radiologists should make reasonable attempts to obtain previous radiographs and compare them with current examinations. The official radiography report should include the results of this comparison.
• If the radiologist is unable to make a comparison because previous radiographs cannot be located or are otherwise unavailable, the interpreting radiologist should contemporaneously document that fact in the report. I suggest inserting into the report a simple phrase such as “previous radiographs are unavailable for comparison.”
• The duty to obtain previous mammograms for comparison with current studies is less clear and remains controversial. Although the ACR standard on screening mammography favors obtaining previous radiographs “when practical,” radiology groups should adopt their own individual policies to deal consistently with previous mammograms and to which all members can adhere.
• If previous radiographic studies are not available within an appropriate predetermined time but are still in the process of being sought by the patient or by radiology personnel, I suggest issuing a radiology report indicating that the previous radiographs are not yet available. Later, if previous studies become available, radiologists can consider issuing an addendum report that includes comparison.
• Radiologists should be aware that deliberately omitting the comparison of a current with a previous radiographic examination because they do not want to call attention to a lesion that was missed on a previous study is fraught with a myriad of medical-legal, ethical, and professional concerns. Radiologists who choose to omit comparisons for this reason do so at great risk.
— “Comparing New Radiographs With Those Obtained Previously” appeared in its original form in the American Journal of Roentgenology. It is reprinted here with permission of the American Roentgen Ray Society.
1. Brenner RJ. Medicolegal aspects of breast imaging. Radiol Clin North Am. 1992;30(1):277-286.
2. Logan-Young WW, Hoffman NY, Janus J, Destounis S. Obtaining previous mammograms for comparison (letter). Am J Roentgenol. 1996;166(4):989.
3. Hall FM, Baum JK, Raza S. Obtaining previous mammograms for comparison (reply to letter). Am J Roentgenol. 1996;166:989.
4. Hall FM, Baum JK, Raza S. Re: Obtaining previous mammograms for comparison (letter). Am J Roentgenol. 1995;165(2):488.
5. Bassett LW, Shayestehfar B, Hirbawi I. Obtaining previous mammograms for comparison: usefulness and costs. Am J Roentgenol. 1994;163(5):1083-1086.
6. Smith MJ. Error and variation in diagnostic radiology. Springfield, Ill.: Thomas; 1967:75-89.
7. Berlin L. Reporting the “missed” radiologic diagnosis: medicolegal and ethical considerations. Radiology. 1994;192(1):183-187.
Letter to the Editor
This article generated the following letter to the editor and reply:
I applaud Dr. Berlin for clarifying many important points with regard to comparing new radiographs with previous examinations in his recent article in the Malpractice Issues in Radiology series. He does, however, leave one area open and ambiguous. It is an area that I believe leaves the referring physician vulnerable and one that we should therefore make every effort to address.
In his article, Berlin states: “If previous radiographic studies are not available within an appropriate predetermined time but are still in the process of being sought … I suggest issuing a radiology report indicating that the previous radiographs are not yet available. Later, if previous studies become available, radiologists can consider issuing an addendum report that includes comparison.”
I agree wholeheartedly with the first sentence, but I think that his willingness to leave what happens thereafter optional (i.e., “can consider”) is unfair to the referring physician and a breach of our duty. Once we have gone on record stating that we will attempt to obtain previous studies, we must eventually issue an addendum report by some predetermined time, whether or not the films were obtained. The report should state either that the previous films were obtained and our recommendation is based on our comparison, or that our efforts to obtain the earlier images were unsuccessful and therefore our current recommendation is based on the information available from the current study alone.
Especially with regard to mammography, if it is important enough to attempt to obtain prior studies, then it is our responsibility to the patient and the referring physician to document, at some point, the results of these efforts. To do anything less leaves the referring physician, with respect to his records, in a state of limbo, assuming that a comparison is yet to come.
Edwin S. Gerson, MD
Radiology Associates of Clayton, PC
Dr. Berlin’s Reply
I thank Dr. Gerson for his comments and commend him for the manner in which he deals with the question of comparing new radiographs with those obtained previously. The issuing of a report that documents whether the effort to obtain previous radiographs for comparison has or has not been successful clearly adds closure to the subject and is therefore helpful to the referring physician. This process appears to work very well in Gerson’s practice, but it may not be feasible or easy to implement in many other, if not most, hospital- or office-based radiology practices. It is for that reason that I used in my article the words “can consider” with regard to my suggestion of issuing addendum reports if and when previous radiographs are found and comparisons are made. Gerson suggests that addendum reports be issued whether or not previous radiographs are located. Either of these suggestions represents, in my opinion, good risk management advice. I should emphasize, however, that there are no written ACR standards, nor any other national practice standards of which I am aware, that mandate the issuing of addendum reports. As I indicated in my article, radiology groups should adopt their own individual policies on how to deal with previous radiographs, a policy to which all members of the group can adhere consistently.
Leonard Berlin, MD, FACR
Fig. 1: 57-year-old woman with recurrent cough and low-grade fever. A and B, Posteroanterior (A) and lateral (B) radiographs of chest obtained at presentation were interpreted by initial radiologist as disclosing “subtle ill-defined infiltration in left mid-lung field, seen only on the PA view, consistent with pneumonia.” C, Posteroanterior radiograph of chest obtained 3 months after A and B were interpreted by defendant radiologist as disclosing a “fairly well-circumscribed infiltration in the left lower lobe, most likely representing pneumonia.” Defendant radiologist did not compare these radiographs with A and B. D and E, Posteroanterior (D) and lateral (E) radiographs of chest obtained 9 months later as revealing that left lower lobe infiltration had increased in size and that because of its “well-circumscribed appearance” and fact that “left hilum appears enlarged,” findings were “suspicious for tumor.”