MPH 515 Week 1 Population Question Post Example

Week 1 Discussion: Initial Post

Saturday Nights in the Emergency Department: Turning a Run of Overdose Patients Into a Question About Bellamy County's Residents

Initial Post

The observation. I work night shifts as an emergency nurse in Bellamy County, a composite Midwestern county of about 415,000 people used throughout this course. Over two months last spring, it felt as though every Saturday brought three or four patients revived from an opioid overdose, several of them people I had seen before. My colleagues and I talked about it as "a bad stretch." That is a clinical impression, built from the patients in front of me, and it cannot tell me whether overdoses were becoming more common in the county or whether I happened to work the busy nights.

The population question. The epidemiologic version of my observation is: among residents of Bellamy County, how many new nonfatal opioid overdoses treated in emergency departments occur per year, and is that number changing? That question needs a numerator, a denominator and a period. The numerator is residents with a qualifying emergency visit; the denominator is the county's residents, counted as person-time; and the period has to be fixed in advance, such as a calendar year.

A Measure Named, With Its Denominator

The measure I would use is an incidence rate: new cases divided by person-years at risk, often reported per 100,000 person-years. If I counted 600 residents with a first overdose visit in a year and the county had 415,000 residents, the rate would be about 145 per 100,000 person-years. A count of visits would not work as the numerator, because one person can come in several times in a year; counting visits mixes how many people are affected with how often they return. Repeat visits are worth reporting, but as a separate measure.

What the Measure Misses

An emergency department rate leaves out people who overdosed and were never brought in, including those revived by a friend with naloxone and those who died before reaching care. It also misses residents treated in hospitals across the county line unless records are shared. And it depends on how visits are coded, so a change in the coding rules could look like a change in overdoses. The rate also says nothing about what happens after discharge. A Massachusetts study of more than 11,000 patients treated in emergency departments for nonfatal opioid overdose found that 5.5% died within one year and 1.1% within one month (Weiner et al., 2020). That is a reminder that the visit is not the end of the story, and that a rate of visits is only one part of the burden.

A Question for the Group

When you turn something you have seen at work into a rate, what do you use as the denominator? I struggled with whether to use all county residents or only adults, since almost no children appear in our overdose visits.

Reply to Keisha

Keisha, your observation about more falls among patients from one assisted living facility is a strong start. To make it a population question, you would need the facility's resident count and a period. Do you know how many residents it has? A rate of falls per 1,000 resident-days would let you compare it with other facilities fairly.

Reply to Devon

Devon, you report that your clinic saw 40 new diabetes diagnoses last year. Forty out of how many patients, and how many of them were tested? If testing increased, the count could rise even if the true occurrence did not. Naming the denominator, and how cases were found, would make the figure interpretable.

References

Weiner, S. G., Baker, O., Bernson, D., & Schuur, J. D. (2020). One-year mortality of patients after emergency department treatment for nonfatal opioid overdose. Annals of Emergency Medicine, 75(1), 13-17. https://doi.org/10.1016/j.annemergmed.2019.04.020