The downsides of automating health care

If you are a doctor, nurse, patient, or just someone interested in patient safety, you should read a five-part story called “The Overdose: Harm in a Wired Hospital” excerpted from a book The Digital Doctor by Dr. Robert Wachter.

Dr. Wachter and the hospital are to be commended for publicizing this incident so others may learn from it. The hospital staff, the patient, and his mother, also deserve credit for allowing their stories to be told.

A synopsis does not do justice to this well-written account of the boy’s near-death experience in a top hospital in San Francisco. In short, he somehow received a massive overdose of the antibiotic Septra despite the presence of a sophisticated electronic medical record and multiple systems in place that were supposed to prevent such a thing from happening.

After the patient recovered from receiving 38½ pills when he should have been given only one, a root cause analysis found numerous faulty system issues such as an electronic ordering program that was overly complex, a nurse “floating” to an unfamiliar floor, a satellite pharmacy that was too busy and susceptible to distractions, “alert fatigue” among hospital staff, and a culture, like that of most hospitals, that may have discouraged questioning both authority and the almighty computer.

Dr. Wachter contrasted the error-prone way we used to order medications on paper, which he said could take up to 50 different steps before the medication got to the patient, with the electronic process which even uses a “smart” robot instead of a human to count out the number of pills to be dispensed.

But, in this case, errors such as those caused by illegible handwriting, transcription errors, and the like were replaced with errors we never dreamed of.

Twenty years ago, a human pharmacist probably would have questioned the order as he was counting out 38½ pills of Septra to be given as a single dose. But the “smart” robot didn’t bat an eye. (Robots don’t have eyelids.)

And most of the nurses of that era would have balked at giving any patient 38½ pills of a single drug at one time.

A French airliner crashed because the pilots didn’t know what to do when the plane’s computer malfunctioned. The author of the lengthy Vanity Fair piece about it said, “Automation has made it more and more unlikely that ordinary airline pilots will ever have to face a raw crisis in flight — but also more and more unlikely that they will be able to cope with such a crisis if one arises.”

A brief article called “The case for dangerous roads and low-tech cars” (also from a book The World Beyond Your Head: On Becoming an Individual in an Age of Distraction by Matthew B. Crawford), discusses the possibility that so-called safety advances in automobile design may lull drivers into a false sense of security.

New options such as automatic braking when a car ahead slows down, or an alert warning about a car in your blind spot may isolate drivers too much. Crawford says, “The animating ideal seems to be that the driver should be a disembodied observer, moving through a world of objects that present themselves as though on a screen.”

On the subject of roads, he writes, “When roads look dangerous, people slow down and become more heedful” and says that some new roads deliberately built with “less safe” features yield fewer crashes.

Like pilots and drivers, are hospital personnel becoming less vigilant by trusting computers and automation too much?

“Skeptical Scalpel” is a surgeon blogs at his self-titled site, Skeptical Scalpel.

Do medical scribes threaten patient privacy?

Medical scribes are a burgeoning field with many institutions and practices exploring their use while the many commercial enterprises who lease out scribes are pushing for their widespread acceptance. There is no accepted definition of what scribes do or what their background or training should be. There is no mechanism for licensure of them in any state. They are poorly defined medical assistants. The field is in its infancy and its ultimate role in our health care system is unclear.

The usual use of a medical scribe is to follow a provider around in their clinical tasks for the purpose of data entry. This may or may not involve being present for the history and physical exam. Most commonly they are physically present in the room and witness the entire encounter. The need they fill is a function of our ever increasing mandates for electronic medical records (EMR). Before EMRs, the use of data processors in examining rooms was quite rare. Thus federal mandates have created another whole class of employees who are placed in the middle of health care interactions increasing the potential cost and complexity for all. Some practices will clearly be tempted to use the scribes for as many uses as possible including assisting with minor procedures which I’ve seen advertised and acting as chaperones, which I know happens. This can help to ameliorate some of the extra cost involved in their employment.

The background of scribes is not standardized. The minimum requirement appears to be competent data entry ability which translates into typing speed at a terminal. Some advertisements require applicants to have a high school diploma or a year or two of college. Many make no mention of education at all. Some commercial sites train the scribe for a period varying from a few weeks to months. At the high end, some advertisers require an existing knowledge of medical terminology with the preferred candidate being a medical student or premed student. This is more often the case when the scribes are being used by academic institutions that have much greater access to people with these qualifications. The majority of practices will not have access to premed or med students. The scribes are mostly young and it is likely that few will choose this as a permanent career. Thus they constantly have to train new ones. The advertised pay rates vary from a minimum wage of about $8/hr up to about $20-25/hr. The work may be full time but often is part time.

The touted benefits of scribes are to increase the efficiency of the practice by allowing the physician to see more patients while having a more personal interaction as they are freed from data entry. It is advertised as a money saving strategy though a practice I discussed this with wasn’t at all sure that it saved money in their practice. Scribes are a possible solution to the imposition of EMRs whose benefit to the patient and practice are frequently unclear or negative to begin with. The dangers to patient privacy are clear but not often emphasized. How much of a concern this is depends on the practice. My ophthalmologist uses them without difficulty. Few would object to an assistant hearing your ophthalmological history.

But many patients of general practitioners or internists would be inhibited from giving an intimate history by the presence of an assistant. A few have the assistant outside the room for this perhaps making the patient more comfortable. The presence of scribes during intimate physical exams is a further situation where many patients would be uncomfortable. One article by urologists studied this and found no problems with acceptance, but the study had many caveats which the authors document. This took place in an academic setting with medical students used as scribes. The majority of the scribes were in fact men. The one female urologist in the practice with a predominately female following refused to be part of the study. I think the results would have been different in a urology practice in a private non academic setting where scribes were predominately young women without a medical background. The routine use of opposite gender scribes would certainly make many patients uncomfortable in a private urology practice. The most common use of scribes to date is in emergency departments where there are certainly many issues of patient privacy. But in emergency settings patients are less likely to be concerned about their privacy.

The accuracy of scribes has not been studied to my knowledge. The results would depend greatly on the background and training of the personnel. But I do not see how anyone with a high school background and a few months of training could possibly understand all the medical conditions covered in a general practice. The physician is responsible for signing off all records of course, but if the record is full of errors, it would be of limited use. If premed or med students are being used, the results should be better, but this is partially mitigated by the fact that there would be immense turn over and new people would always need to be trained.

In summary, the routine use of scribes in a general practice has many potential problems of privacy and accuracy. I don’t see how scribes without a medical background could ever be competent to understand a general medical history. It would be far more accurate to have patients fill out detailed questionnaires on the initial visit and then have data entry performed on this later by scribes. This would also permit a far greater feeling of patient privacy. Patients have always understood that their medical records may be processed for administrative purposes.

The presence of scribes during intimate histories and exams should be very limited and only done with the express permission of the patient. They should not be used as chaperones or multipurpose medical assistants without further special training. In the long run the use of scribes is likely to be a temporary answer to a cumbersome system of mandated EMR which can still cause as many problems as it solves. There is no intrinsic reason why EMRs need be so intrusive that their use requires a staff of intermediaries. This makes it increasingly impossible for solo physicians and small groups to stay in practice.

Joel Sherman is a cardiologist who blogs at Patient Modesty & Privacy Concerns.