Drug Overdose ER Visits and Hospital Stays

Limitations

Overdose ER visits and hospital stays in Georgia non-federal hospitals, by home county: visits, not people (repeat visits count) or deaths: never compare or add them to drug_overdose_deaths. Border counties look low: out-of-state stays are missing. setting all = er + inpatient. drug_type rows overlap: never add across them. fentanyl has no rows before 2021 (no code until October 2020). OASIS withholds rates on 1-4 visits. Race-group rates run low: ethnicity is often unknown (NULL above 20%). Age-adjusted rates are OASIS's: weighted averages of age-specific rates with the year 2000 U.S. standard million as weights, so places and groups with different age mixes compare fairly; they never add or average across rows. OASIS publishes every count, however small, and withholds rates based on 1-4 events (NULL here); the platform adds no suppression of its own and never recomputes a withheld rate. A group with no residents in a place (num_population 0) has no rate (NULL): OASIS shows a crude rate of 0 there (and withholds the age-adjusted rate itself), but a rate on no population is undefined and would rank as the lowest (our mask; the count stays 0). 43% of rows with at least one visit have no age-adjusted rate, mostly 1-4 visits in small counties and groups; on race rows OASIS also withholds it where a group has no residents in some age group (-2) or more visits than residents in one (-3: a person can have several visits). The case definition is OASIS's (https://oasis.state.ga.us/dtt/content/DefDERDrug.html): any mention of an ICD-10-CM drug poisoning code T36-T50 (any over-the-counter, prescription or illicit drug; alcohol excluded) whose intent is unintentional, self-harm, assault or undetermined (6th character 1-4), initial encounter (7th character A or missing). er counts ER visits that did not lead to admission; inpatient counts hospital stays, including patients admitted through the ER, who count once; all is both (60% to 67% of Georgia's visits a year are er). Visits are dated by discharge and include patients discharged dead (about 1% in 2017-2019, OASIS says), whose deaths may also be in drug_overdose_deaths. The drug types are the poisonings listed on the visit: any_opioid (T40.0-T40.4, T40.6) contains heroin (T40.1) and fentanyl (T40.41), and a visit involving two counts in both. fentanyl is identified only by T40.41, a code that began 2020-10-01: OASIS shows 0 fentanyl visits everywhere through 2020 (even in October-December 2020), which are not real absences: fentanyl has no rows before 2021 (dropped, not served as zeros); fentanyl overdoses of those years are in any_opioid and all (T40.4) but cannot be told apart. drug_overdose_deaths' synthetic_opioid_other_than_methadone (T40.4: fentanyl, tramadol and other synthetic opioids) is a different definition. Georgia's opioid overdose visits (every setting): any_opioid 5,369 in 2016, 7,556 in 2021 (the peak), 4,510 in 2025; heroin 1,208, 2,350 and 173; fentanyl 518 in 2021, 1,861 in 2023 and 1,367 in 2025. T40.41 was new in October 2020, so part of fentanyl's rise from 2021 to 2023 may be hospitals starting to use the code (our reading). OASIS revised its counts: every overdose for 2016-2021 on 2023-02-28 (about 100 more a year) and any_opioid for 2020-2023 on 2025-07-03 (about 995 more a year statewide, for changing ICD-10-CM codes). Where race or ethnicity is unknown for more than 20% of a place-year's visits (race_unknown / all), the race groups' counts and rates for that place and year are NULL (our mask, OASIS's own 20% threshold for missing items): each group's count would miss its share of the unknowns, so its rate runs low and its trend follows the recording, not the group. The all, sex and race_unknown rows are kept, so the unknown share stays visible. It applies to 145 county-years (55 counties, never Georgia), measured on every overdose visit (drug_type and setting all) and applied to every drug type and setting. OASIS's hospital data records ethnicity from 2016, this module's first year, but some hospitals still do not (OASIS: about 9% of discharges; https://oasis.state.ga.us/dtt/content/KDIDERDrugOD.html): 6.3% (2023) to 16.5% (2016) of Georgia's overdose visits a year are of unknown ethnicity (race_unknown), and up to 96% of a county's among counties with 100 or more visits (Camden, 2018: 113 visits). In smaller counts every visit can be of unknown ethnicity (167 county, year, drug type and setting combinations, such as all 12 of Atkinson County's overdose visits in 2023 and all 51 of Glynn County's any_opioid visits in 2019): there every race group shows 0 visits and, where published, a rate of 0, which means ethnicity went unrecorded, not that no one in the group had a visit. The race mask NULLs the race rows of most of them (both examples); 27 remain, in county-years where 20% or fewer of all overdose visits lack ethnicity. So the race groups' counts and rates run low, unevenly by county and year; OASIS advises caution with Hispanic figures. race_unknown is our calculation, all minus the seven race groups: here exactly the visits of unknown ethnicity (OASIS shows no non-Hispanic visit of unknown race), a count with no population or rate. Stays in other states' hospitals and in federal (VA and military) hospitals are not counted: Dade County, next to Chattanooga, Tennessee, had 2 overdose visits in 2024 (16,143 residents), and Catoosa's and Walker's crude rates were 73.5 and 133.4 against Georgia's 229.7 per 100,000 (our reading: their residents use Tennessee hospitals). A small group's rates can be very high on few residents (the highest crude rate, 61,111 per 100,000: 11 visits of 18 asian residents of Johnson County, 2019; our reading: repeat visits or race recorded differently by hospitals and the Census): kept as published. num_population is OASIS's own population (its population module) for the place, year and group, the denominator of every crude rate (num_visits / num_population x 100,000 reproduces each published crude rate, in every setting); the race groups' populations add to all, as do male and female. 2025 is the newest year; OASIS marks no year provisional. Counties add up to Georgia, except race groups, whose counts are NULL in the county-years the race mask covers (Georgia's race rows count every county, so a race group's counties add up to less there). Not served: OASIS's payer and SES-vulnerability rows (their 'rates' divide by the whole place's population; payer counts belong in a separate payer topic), and age and education breakdowns. tract_geoid is always NULL.