My company sells a large number of medical device implants per year. There are often “incidents” or minor issues, associated with them that can be easily categorized into an incident category.
I’m tasked with trending our data to look for any statistically significant increase in the frequency of these incidents, per incident category. Here is a mock-up of what I have been given.
| 2021 | 2022 | 2023 | 2024 |
| Implants sold | 189 | 74 | 180 | 154 |
|---|---|---|---|---|
| Incident A | 0 | 0 | 1 | 0 |
| Incident B | 0 | 0 | 8 | 4 |
| Incident C | 1 | 1 | 3 | 0 |
| Incident D | 0 | 2 | 0 | 0 |
| Incident E | 0 | 0 | 1 | 1 |
| Incident F | 0 | 2 | 0 | 2 |
Given that there are many implants sold per year, and relatively few incidents (and even fewer incidents per category), how could I go about trending this data? Which statistical techniques are optimal for a situation like this?
I’ve tried looking at a few tests that operate well for small sample sizes, but I struggle to find one that trends data over the years. Any help would be much appreciated.