Are 25 Percent of Hospital Patients Really Harmed?

The devil is in the details for this OIG report.

Last week the OIG released a report entitled “A Quarter of Medicare Patients Experienced Harm in October 2018.” As one would expect, that title led to extensive media coverage suggesting hospitals are routinely harming patients. We saw the same in 2000 with the publication of “To Err is Human” which claimed that 98,000 patients a year are killed by medical errors. But while that headline is appealing, as with many government reports the devil is in the details. And in this case, there are a lot of details that the headline neglects.

As in most OIG audits, they used extrapolation, looking only at 800 admissions of the over 1,000,000 admissions that occurred that month. I am not a statistician but auditing 0.07 percent of all claims seems inadequate. But arguing bad extrapolation never works so instead the focus should be on their accusatory terminology. Every event is labeled as “harm.” This is an inappropriate characterization of what happened to patients and the implication of that term is that this harm is the fault of providers.

While in some cases, the harm may have been preventable, such as the “pneumonia that was missed and resulted in a readmission,” for the vast majority of the reported harms, the patient experienced an expected and well-known side effect of the treatment. For example, a patient who was admitted as an inpatient with a stroke and inability to swallow had a feeding tube placed. Of course, it was appropriate to provide nutrition. Unfortunately, the patient aspirated and developed pneumonia, with the OIG calling this patient harm. Feeding the patient with a feeding tube was absolutely standard of care. Despite doing everything perfectly right, patients with feeding tubes still aspirate.

It’s in the nature of medical care that side effects and adverse effects occur. They also described venous thromboembolism after surgery as harmful, even though there has never been a treatment to avoid 100% of all post-operative thromboemboli. It would be one thing if appropriate prophylaxis was not ordered, but a thrombus even with the standard of care prophylaxis is still considered in this report to be “harm” and suggests that the provider was responsible for the outcome.

In another case, a patient developed a rash from chemotherapy and it was called harm. The rash experienced by the patient was a known and common side effect of the chemotherapy. Suggesting the providers “harmed” that patient is not appropriate. Should that patient have not been given chemotherapy because it could cause a rash? Should the diabetic patient who developed mild hypoglycemia from insulin have not been given insulin so there would be no risk of being harmed by hypoglycemia? Should the patient with atrial fibrillation who developed a GI bleed while on an anticoagulant not have been given the anticoagulant and instead been allowed to have a stroke?

What the OIG is doing here is criticizing providers for treating patients with the standard of care. Their use of the word “harm” implies that treatment caused harm, it is the fault of the provider. Yet if the treatment was not administered and there was a bad outcome, that would be just fine. The crime of commission is punished but not the crime of omission. That’s a dangerous double standard.

In one case the OIG criticized the care of a patient with an ankle fracture who developed lethargy from opioids. The OIG determined that this event likely could have been prevented by using a less aggressive pain management regimen after surgery. Can you imagine the uproar if the patient’s pain was not adequately treated? Would the patient’s uncontrolled pain have been captured in the OIG’s audit as harm? I am certain the patient would classify it as such.

A theme we are hearing from CMS repeatedly this year is increasing health equity. Improving health equity is a complex journey that will require changing many of the relationships that exist in medicine today. Attacks like this won’t promote that cause but lead providers to be more suspicious of government oversight and are likely to lead to cherry-picking patients who are at lower risk of suffering from adverse effects of treatment.

Before accepting the conclusion that 25 percent of patients are harmed, read the actual report, take the time to read the cases outlined starting on page 73, and decide if that “harm” should have led to the provocative headlines. Finally, to lighten the mood a bit, let me present this preventable harm case patient with hyperkalemia leading to severe bradycardia per telemetry with an inadequate emergency response due to difficulty locating the patient. That one is a problem.