An article in the Journal of Patient Safety by a NASA toxicologist and patient safety advocate, John T. James, PhD, addressed the question of medical errors in hospital settings. Dr. James evaluated the results of 4 studies of patient adverse events (PAE) and extrapolated the results to estimate that 400,000 deaths yearly are associated with preventable patient harms. The Harvard Medical Practice study which reviewed records from several New York hospitals in 1984 estimated that deaths due to medical error numbered about 98,000 per year, and that number has been quoted widely and embraced as the truth by the Institute of Medicine, a nonprofit organization which seeks to advise patients and decision makers about important issues in medicine. This current article uses different methods and suggests that the number may be much higher.
After reading the article about the article on a public interest site on the internet, I perused the comments, which were primarily outrage, interspersed with a doctor or two who seemed pretty defensive and said unpopular things that were on the order of, “What do you expect? We’re doing the best we can!” My first reaction was to wonder what exactly the study meant, how was it done, and what were the definitions. I found the article and read it and came to some conclusions that probably ought to be part of the discussion.
How did they reach their conclusions?
The 4 studies that were evaluated were in hospitals in various areas of the country in which a “Global Trigger Tool” was used to identify patients who might have had an adverse event, something bad that happened in the hospital. This could have been a missed lab test, a procedure complication, a drug reaction or a hospital acquired infection. The charts of the patients who were identified this way were then reviewed by doctors and the doctors determined if it looked like the adverse event contributed to that patient’s death. Each of the studies was performed a little differently, but they shared this general method.
They found that about 0.65-1.4% of patients who were hospitalized had an adverse event that contributed to their death. Overall, in the 4 studies, 4,252 randomly selected patients’ charts were reviewed, and of that number, adverse events contributed to the deaths of 38 people. Based on other studies, 69% of these adverse events were felt to be preventable, and multiplied over the 34.4 million patients hospitalized yearly, they came up with 210,000 people yearly for whom preventable adverse events contributed to their deaths. Since their methods did not include errors of diagnosis or errors of omission, that is not figuring out what was wrong or not treating it according to accepted guidelines, and medical errors are not always well documented in the medical records, the author rounds this 210,000 up to about 400,000 patients who died partly because of medical error.
Why this is somewhat misleading
This is a big and very important issue. I see medical errors and, more often, adverse events from decisions that, in retrospect, were wrong, on a daily basis when I work in hospitals. But there is something very wrong in the way that this article is being reported. The author himself labels adverse events that contribute to death as “lethal errors.” Some of them are, but some of them just contribute to the lethal cascade of events, from life choices to decisions to pursue aggressive treatments, which end in death. The overdose of insulin that kills a patient is a rare event. The blood clot in the leg because of failure to order prophylactic heparin that leads to full anticoagulation that leads to gastrointestinal bleeding which leads to intensive care unit stay which leads to ventilator associated pneumonia and death due to drug resistant organisms because the patient was a diabetic who also dabbled in intravenous drugs, or some version of this narrative, is common. The failure to prescribe heparin in the first place should hardly be called a lethal error. In fact, the prescription of heparin in the first place, itself, might have led to a life-threatening complication.
The outrage which is beginning to erupt is taking the form of “Doctors kill 400,000 people a year in hospitals due to negligence.” That is not what this paper says. It says that 400,000 people who die each year in hospitals are made sicker by the things we do to them. Another issue brought up in this article is that 10 to 20 times this number of patients are significantly impacted by medical error in hospitals. These events might be a hospital-acquired infection or a surgical wound complication or a drug error from which a patient recovered, though not without suffering. If the estimate is correct, we are making 4-8 million people sicker in hospitals yearly in our attempts to heal them. We need to pay good attention to all of these things and identify the processes that make them happen and look for practical solutions. This falls under the heading of patient safety. Much is actively being done in hospitals to improve patient safety, but the complexity of what we do limits our ability to protect those in our care from harm.
An example from a hospital shift
I admitted 4 patients to the hospital yesterday and so had ample opportunity to make life altering mistakes in the lives of 4 individuals. The first was a morbidly obese woman with new diabetes who had developed a severe leg infection related to the fact that her legs were always swollen. The infecting organism was probably a staph or strep, and given the community in which I am working, it might be methicillin-resistant Staphylococcus aureus. I started her on two antibiotics, both of which might cause a life threatening allergic rash, and one of which might cause kidney failure. Both might also lead to a life-threatening Clostridium difficile infection of the colon. Starting insulin might lead to a hypoglycemic reaction that could result in death or disability. Simply putting the patient to bed might lead to bedsores, especially if I treat her pain with intravenous opiate medications which could lead to respiratory depression and death or disability.
I then moved on to the man who had been monthly to this hospital due to complications of ongoing methamphetamine abuse with diabetes for which he usually failed to take his medications and end-stage liver failure from hepatitis C. This time he wasn’t very sick, just had a very high blood sugar level and a very low potassium level. I gave him potassium first, because giving insulin would lower his potassium further and might cause a fatal heart arrhythmia. I gave lots of potassium, because the level was very low, and I risked raising the potassium level to a point where it might cause a fatal arrhythmia. The pharmacy called to see if I really wanted to give all of that potassium and I said yes, and thanks for checking on me. Pharmacy monitoring of physicians’ orders is standard, and they catch errors all the time and I really appreciate it. I had to decide whether to give intravenous fluid, and decided not to, despite some kidney failure, because of a long history of going into heart failure with just a little fluid overload.
I then headed to my third patient, a woman with chronic pain and ongoing alcohol abuse, a quart of whiskey a day, who was feeling like she was going into alcohol withdrawal again and also was short of breath because of her chronic lung disease and new pneumonia. She had just quit smoking but was using some kind of nicotine nebulizer she got in a tobacco shop. She was on warfarin, which she said had been perfectly stable for months. I ordered the antibiotic for her pneumonia which could cause tendon rupture and antibiotic associated diarrhea, the asthma meds which could make her heart beat erratically and make her delirious, decided against the sequential compression devices for her legs since she had known arterial occlusive disease with recent arterial clot which nearly resulted in amputation, even though I did risk her developing a clot in a vein. I ordered her regular warfarin dose and checked her protime. The pharmacy called a couple of hours later to tell me the protime was dangerously high, so I thanked them and discontinued the warfarin, asking the pharmacy to monitor it and restart when appropriate. I ordered high-dose lorazepam, which might lead to breathing problems with resultant need for a ventilator which could lead to pneumonia and ICU-related dementia.
Then up the stairs to the transfer patient, a 39-year-old man who used IV and subcutaneous methamphetamine and had developed abscesses all over his skin and a heart valve infection. He had been in hospital improving from this for 3 weeks, and now was in my care with a central line (intravenous catheter in his internal jugular vein) and open wounds on both hands and both legs after surgical treatments of his abscesses. He was on high dose pain meds and still complained of pain. He had a fever. His fever might have been from his known severe heart valve infection, from an internal abscess as yet undiagnosed or from the central line.
I ordered blood cultures which could only be obtained from the central line because he had no accessible peripheral veins and considered the option of replacing the central line with a new one which would not be a nidus of infection but the insertion of which could cause injury to an artery or puncture the lung, which might kill him. Another option would be a peripherally inserted central catheter which would have a pretty high risk of resulting in a blood clot in the arm which could migrate to the lung and cause his death.
I ordered most of the medications which had been working at the previous hospital. They included drugs that could cause kidney failure, respiratory failure and allergic reactions of all kinds. He was also on an antipsychotic medication that could cause a permanent neurological impairment called tardive dyskinesia in which the tongue involuntarily keeps sticking out of the mouth. When he didn’t take that, he had apparently been extremely anxious and agitated.
Keeping patients safer
And that wasn’t even a big day. Clearly protecting patients from harms in the hospital is a priority. The most important way to protect patients is to keep them out of the hospital and keep their treatments as streamlined as is possible without risking errors of omission.
Keeping patients out of the hospital means assigning priority to resources that keep people healthy. It also means having important conversations with patients about what they really want done and what being in a hospital means. Presenting every patient with the list of the all of terrible things our treatments can do to them is impractical, but we need to bring them in to the conversation. We need to have the time to do this, which means backing off on something else which fills our time. Deliberately simplifying care would help, and we could make this an item for action. If there were fewer medications prescribed, fewer specialists consulted, fewer intravenous lines inserted, fewer surgical procedures offered and performed, there would be fewer medical errors. We can make this happen, but it will involve a concerted decision by patients, families, doctors and hospitals that it is something we all value. Doing less will redirect the revenue stream which will be uncomfortable and disruptive, but so very worthwhile.
Janice Boughton, MD, ACP Member, practiced in the Seattle area for four years and in rural Idaho for 17 years before deciding to take a few years off to see more places, learn more about medicine and increase her knowledge base and perspective by practicing hospital and primary care medicine as a locum tenens physician. She lives in Idaho when not traveling. Disturbed by various aspects of the practice of medicine that make no sense and concerned about the cost of providing health care to every American, she blogs at Why is American Health Care So Expensive?, where this post originally appeared.