Articles on Law, Science, and Engineering

How Effective Safety Devices Lead to Secondary Litigation

By Edward P. Richards and Charles Walter, 10 IEEE Engineering in Medicine and Biology Magazine #2, pg. 66 (June 1991)

This article discusses the paradox of safety devices that perform properly, but increase litigation against the physicians who use them. The phenomenon of secondary litigation has been neglected by device manufacturers and designers because they are not parties to the lawsuits. As the story on fetal monitors unfolds, the risk of secondary litigation will become a major consideration in the selection of new technologies by physicians and hospitals. This will put devices with a potential for secondary litigation at a competitive disadvantage, making secondary litigation an economic, if not legal, threat to their manufacturers.

Fetal Heart Monitors

We will use electronic fetal heart monitors and pulse oximeters to illustrate the continuum that spans the range of devices that increase secondary litigation to devices that decrease secondary litigation. The widespread use of fetal heart monitors was accompanied by a dramatic increase in obstetric malpractice litigation. It is certain that factors other than fetal heart monitors were primarily responsible for this increase. It is also certain, however, that the use of these monitors did not decrease litigation and, in the cases where fetal heart monitor records were available, these records increase the probability of litigation when a baby is born damaged. Conversely, the widespread adoption of pulse oximetry in the operating room was accompanied by a dramatic reduction in malpractice claims against anesthesiologists. These monitors appear to be among the primary causes of this reduction in claims. They clearly are not being used against anesthesiologists in the way that fetal heart monitors have been used against obstetricians.

It should be noted at the onset, however, that secondary litigation does not imply a defectively designed product, at least not in the traditional sense. Fetal heart monitors perform accurately and reliably within the constraints of what they measure. Devices that are unreliable or otherwise directly dangerous to patients will be the target of primary litigation against the device manufacturer. Pure secondary liability is an issue for otherwise safe and well engineered devices. Secondary liability becomes a problem when the device in question:

• Documents previously undocumented behavior indicating negligence on the part of the medical care providers;
• Records data with ambiguous interpretations; and/or
• Inappropriately leads to changes in medical care supervision, staffing, or patient contact because of reliance on the device.

The first problem, documentation of previously undocumented behavior, is common to all recording, as opposed to display only, instruments. (As discussed later in this article, it will prove especially interesting in the case of real-time clinical information systems.) Medical record systems that depend on people to record events, either on paper or computer, are highly filtered. Sometimes this filtering is conscious and intentional, as when personnel attempt to cover-up an error by not entering incriminating information into the medical chart. In the worst case, there may even be attempts to change previous entries. This intentional distortion of data is assumed to be relatively infrequent. Most commonly the filtering is unconscious and unintentional. It may occur because the recorder's memory fades between rendering the care or making the observations and recording them in the medical record. In many situations it occurs because the nurses do not make their primary entries into the medical record.

While generally forbidden by protocols on medical records management, off-chart records are commonly used by nurses to keep track of things to do, medications given, and patient observations. Most typically kept on index cards, these temporary records allow the nurses to batch enter data into medical records rather than keep contemporaneous records. This allows two stages of filtering: first the information is compressed into a minimal temporary record, then that minimal record is expanded into the permanent chart record. Such filtering makes it easy to accidentally transform an item from the to-do-list to the done-list without the task actually being performed. It also gives the filter an extra chance to remove non-conforming information.

Filters are defined by what they exclude. Human clinical filters tend to exclude things that do not easily fit into the expectations associated with the care of a given patient. This should not be seen as an act of deception, or even of carelessness. It is more a smoothing of data which tends to obscure anomalies. The smoothing is aided by the limited amount of information that can be recorded by periodic observations recorded in essentially narrative format. The end result of this process of smoothing and filtering is a medical record that is more often characterized by what it does not contain than by what it does document. This is borne out in litigation where records are usually incriminating because they fail to record things which were allegedly done.

When a real-time recording technology is introduced into a situation that previously depended on manual records, the amount of data recorded increases dramatically. A record that displayed fetal heart rates taken at 15 to 30 minute intervals and recorded at some later time by a nurse is suddenly displaced with a fetal heart monitor that generates a paper tape with a continuously recorded fetal rate. Previously unnoted short-term irregularities are now carefully preserved. Unfortunately, whenever an injured child is born, the fetal heart rate record will be scanned by plaintiff's counsel in hopes of finding some deviation from normal that can be used to build a case against the delivering physician. This search is seldom in vain, because of the second problem that leads to secondary liability: data with ambiguous interpretations.

Fetal heart rate is monitored in women in labor to determine if the fetus's well-being is compromised. If the fetus is in trouble, the usual indication is for an emergency Cesarean section. Fetal heart monitors provide a reasonably accurate record of fetal heart rate. The problem is in interpreting these records. Dramatic, prolonged slowing of the fetal heart rate clearly means trouble. There are many other patterns of fetal heart rate irregularities that do not so clearly point to trouble that they unambiguously call for an emergency Cesarean section with its attendant risks and costs. On the other hand, if the baby is born damaged, such irregularities will seem very important in hindsight. In a strict sense, much of what a fetal heart monitor records is not information: it does not reduce the physician's uncertainty over the selection and timing of Cesarean sections.

In retrospect, the major factor in secondary litigation from fetal heart monitors may be the third factor: the shift in patient care patterns that accompanied the routine use of the devices. The traditional method of determining fetal fitness was to directly auscultate the fetal heart with a stethoscope. This requires someone to closely observe the laboring woman at frequent intervals. The premise of electronic fetal monitoring was that the heart rate itself was the critical parameter in this evaluation. It may be that other observations that accompanied this direct and intimate contact with the patient provided a necessary context for interpreting the significance of changes in the heart rate. As physicians and nurses came to rely on fetal heart monitors, they could evaluate the fetus by looking at the monitor strip and ignore the patient entirely. It is also likely that these cursory evaluations decreased in frequency because the monitor allowed the retrospective review of the heart rate. While it is difficult to sort out the causal factors, the most recent research indicates that the use of fetal monitors increases the probability of adverse fetal outcomes.

Pulse Oximeters

Pulse oximetry is a relatively simple technology that measures arterial oxygen saturation in real-time. The oximeters that are routinely used in clinical care are display only instruments. They do not produce a continuous historical record of oxygen saturation. When saturation falls below a certain threshold, an alarm is sounded. This alarm unambiguously means that the patient needs more oxygen or that the instrument has become detached or dysfunctional. In any case, definitive action can be taken at once. These factors make oximetry an ideal safety technology. It helps prevent injuries while not otherwise affecting record keeping or staffing practices. This is not entirely due to special virtues of oximetry.

It is rare that the patient will be left completely alone in the operating room. The genius of oximetry is that it is a monitor that even a surgeon can understand. When the alarm goes off, the surgeon can yell for help if the anesthesiologist has drifted away. It can be assumed that oximetry used as a remote sensing, continuous recording technology will pose the same documentation and staffing problems as fetal heart monitors. These may be outweighed by the clear intervention signal provided by oximetry. If, however, hospitals use recording remote oximeters in situations where this clear signal is ignored, they will suddenly find oximetry to be the fertile source of litigation.

Lessons for Technology Assessment

Continuous recording instruments are invaluable research tools. If there is a central criticism of fetal heart monitor usage, it is that the monitors were adopted for routine use without the research background necessary to understand their limitations. As illustrated by fetal heart monitors, more data (as opposed to information) increases the risk of litigation. This happens because plaintiff's attorneys thrive on ambiguity. If a patient is injured, the plaintiff's attorney will comb the records for anything that cannot be clearly explained to portray as the cause of the injury. This is countered to some extent by the highly filtered nature of traditional medical records. Continuously recorded streams of ambiguous instrument output, in contrast, provide a goldmine of exploitable ambiguity.

It is critical that new instruments be evaluated for the reliability of their intrinsic measurements and for the clinical significance of those measurements. Does the data resolve clinical questions, or merely complicate them? This question becomes more interesting as hospitals purchase clinical information systems that feed the outputs of various existing monitors into a computer-based continuous recording system. The assumption is that since these instruments are already in place, their output must be useful. It is possible, however, that many existing measurements are only useful in the larger context of patient care. Like a fetal heart rate observed without the context of the mother, their readings may be much less valuable as retrospective records. Given that efficacy testing is a recent innovation for medical devices, it is also possible that some existing measurements are clinically irrelevant. This means that their outputs will be randomly related to the patient's condition, creating that ambiguity so valuable to plaintiff's attorneys.

Clinical instruments provide data, not necessarily information that aids in clinical decisionmaking. This data creates noise that can obscure significant clinical information while providing 20/20 hindsight to those who would second guess the physician's actions. If clinical information systems are not carefully thought through and tested, it may be that the IEEE-488 connector will someday be renamed the "plaintiff's best friend."