First do no harm part 2: How can quality efforts harm patients?

As I noted on my previous blog, the Hippocratic Oath has been in existence since anywhere from 200-500 AD and is still relevant today. I discussed how a proposed eCQM was not enacted over concerns that the measure might actually harm patients.

The proposed but not finalized eCQM, the Hospital Harm – Opioid-Related Adverse Events eCQM, was designed to reduce adverse events associated with administration of opioids in the hospital setting. The measure was intended to lead to safer care by incentivizing hospitals to track and improve monitoring of patients receiving opioids during their hospitalization.

The eCQM would be measured by assessing administration of naloxone as an indicator of harm with naloxone utilized as a proxy of administration of excessive doses of opioid medications.

This measure was submitted to the National Quality Forum (NQF) MAP Hospital workgroup and did not receive approval for several reasons:

• Inaccurate denominator and under reporting of volume of events due to methodology
• Measurement was too broad as naloxone is also used for adverse reactions, diagnostic purposes, and side effects of opioids

In addition to the concerns from NQF, CMS solicited feedback and several of the commenters expressed concerns that:

• Hospitals would underutilize naloxone and would turn to use of BiPAP or intubation
• Hospitals would underutilize opioids for legitimate pain control needs

The stakeholders and CMS noted it would be unethical to withhold administration, but were concerned that providers and institutions may be overly concerned with hospital performance.

This is not the first time quality improvement efforts have been reviewed for their harmful (or potentially harmful effects). One example is an article in JAMA published in 2018 which noted a potential link between the CMS Hospital Readmission Reduction Program and 30-day mortality following hospitalizations for heart failure and pneumonia.

It is also not the first time that underreporting of adverse effects has been raised. CMS noted that previous efforts to detect adverse drug events focused on voluntary reporting and tracking of errors and referenced studies demonstrating that only 10 to 20 percent of Sentinel Event errors are ever reported.

I will also note that I have heard of anecdotal reports of decreasing screening as to not identify potential hospital acquired conditions such as DVTs and catheter-associated UTIs (CAUTI).

But what about under-reporting in general? 

CMS and AHRQ have reported decreases in adverse effects/hospital acquired conditions (HAC) over the last 20 years. The question has remained if it was due to:

• Improved patient care
• Improved documentation
• Both improved care and documentation

As an incidental note, CDI programs have been in place approximately 25 years, coinciding with the advent of quality improvement efforts following the Institute of Medicine’s “To Err is Human.” It would almost impossible to determine what percentage of the decrease in adverse effects is due to documentation improvement efforts as compared to improved care. 

It is also unclear what the volume of underreported conditions is. However, if mandated Sentinel Events are grossly underreported, it should not be a huge leap to suppose patient safety events and HAIs are also underreported.

But where is the harm?

In the case of the adverse opioid events and the under-surveillance of DVTs and CAUTIs, it should appear obvious that chasing performance metrics can cause direct harm to patients. It should also be obvious that decreased screening may lead to poor patient outcomes.

But I want to remind everyone of one more “harm”: bad data. Which quality issues should our country or our organization focus on?  If we underreport complications, it may appear that we don’t have a problem, or that there is less of a problem and we may turn our attention to other quality issues. 

And finally, I want to share one of the best representations of accurate quality data. I had a client who had significantly poor PSI rates and of course, fingers were pointed at coding. A five-year targeted effort was put into place to clean up the coding and documentation. I will note it was done very transparently and ethically.  After five years, there were legitimate decreases in complications in most of the PSIs except postoperative DVTs. With this clean data in hand, providers were able to realize they needed practice changes to reduce the incidence.

This is what quality measures and quality data is meant to do: Drive quality improvement efforts to reduce real patient harm.

Cheryl Manchenton is a senior inpatient consultant and project manager for 3M Health Information Systems.

References

Hippocrates of Cos (1923). "The Oath". Loeb Classical Library. 147: 298–299. doi:10.4159/DLCL.hippocrates_cos-oath.1923. Retrieved 20 November 2020.

Kessler, E.R., Shah, M., Gruschkkus, S.K., et al. (2013). Cost and quality implications of opioid -based postsurgical pain control using administrative claims data from a large health system: opioid-related adverse events and their impact on clinical and economic outcomes. Pharmacotherapy, 33(4): 383-91.

The Joint Commission. (2012.) Safe Use of Opioids in Hospitals. The Joint Commission Sentinel Event Alert, 49:1-5.

Rozich, J., Haraden, C., & Resar, R. (2003). Adverse drug event trigger tool: A practical methodology for measuring medication related harm. Quality and Safety in Health Care, 12(3), 194-200

Institute for Healthcare Improvement (IHI). Measures, Adverse Drug Events Per 1,000 Doses. http://www.ihi.org/resources/Pages/Measures/ADEsper1000Doses.aspx.

Association of the Hospital Readmissions Reduction Program With Mortality Among Medicare Beneficiaries Hospitalized for Heart Failure, Acute Myocardial Infarction, and Pneumonia, JAMA. 2018;320(24):2542-2552. doi:10.1001/jama.2018.1923,, https://jamanetwork.com/journals/jama/fullarticle/2719307