Dementia ER Visits (Alzheimer's Disease and Related Dementias)

Limitations

Counts ER visits, not people (a person who visits twice counts twice), whose principal diagnosis is an ADRD cause; visits that end in admission are inpatient stays, not counted here. 2025 is held back: OASIS's 2025 visits fall 26% from 2024, mostly unspecified dementia, while all-cause ER visits rose, so far unexplained. Causes nest: never add them; use cause = 'all'. These visits are also in er_visits_by_cause: never add the two. Race-group rates run low (some lack ethnicity; 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). On cause = 'all', OASIS publishes Georgia's age-adjusted rate every year (all, male and female) and withholds it for 29-43 counties a year, those with 1-4 visits. 70% of rows with at least one visit have no age-adjusted rate, all but 2 of them on 1-4 visits (small counties, small groups and rare causes); the other 2 are small race groups' county rows whose age-adjusted rate OASIS cannot compute (null_meaning). What counts: OASIS assigns each visit the cause of its principal (first-listed) diagnosis (the module page; its definitions page, https://oasis.state.ga.us/dtt/content/DefADRDER.html, words it as the underlying cause of illness), so a visit with dementia only as a secondary diagnosis is not counted; alzheimers_disease equals the Alzheimer's visits of OASIS's all-cause ER module in every place and year (checked 2026-10-28). An ER visit is one visit to the emergency room of a non-federal acute-care hospital in Georgia (billing data from the Georgia Hospital Association), by the patient's county of residence, in the year the patient left the ER, that did not end in an inpatient admission: OASIS counts an admitted ER patient as a hospital discharge and keeps ER and discharge data mutually exclusive. Visits of Georgia residents to other states' or to federal (VA and military) hospitals are not counted, so counties near another state's hospitals may look low. The 2025 hold-back (decided 2026-10-27): Georgia's ADRD ER visits were 4,141 in 2024 and 3,070 in 2025 (unspecified_dementia 2,750 and 1,541); 2025 stays out until a later OASIS release or DPH explains the drop. How causes add up: the 8 groups never overlap and add up to all exactly, and so do the subtypes of alzheimers_disease and of dementia_in_other_diseases. In other_degenerative_nervous_system_diseases (G31), neurocognitive_disorder_with_lewy_bodies (G31.83) and corticobasal_degeneration (G31.85) are inside other_specified_degenerative_nervous_system_diseases (G31.8; OASIS's definitions), so the group's seven listed subtypes add to more than it on 734 rows (by up to 47 visits); without those two they never exceed it and fall at most 5 visits short on 56 rows (probably Pick's disease, G31.01, which is not listed: our reading). OASIS's Selected Causes Total (895 for Georgia in 2024) counts only the visits in listed subtypes, each once, and is not served. dementia_with_anxiety, dementia_with_mood_disturbance and dementia_with_psychotic_disturbance are ICD-10-CM codes added in October 2022, so no visit could carry them before: they have no rows before 2022 (rows dropped, not served as zeros), 2022 covers October-December only, and Georgia's first visits come in 2022-2023; the mix of the other dementia_in_other_diseases subtypes shifts then (our reading). Coding also moved visits between causes after ICD-10-CM expanded its dementia codes in October 2022 (our reading): Georgia's vascular_dementia visits fell from 110 in 2022 to 51 in 2023 (62 in 2024), across counties, while dementia_in_other_diseases rose from 66 to 103 and unspecified_dementia from 2,516 to 2,734; OASIS's dementia hospital discharges break the same way and dementia deaths (ICD-10) do not, so don't read vascular_dementia's fall as real: follow it with the other dementia causes, or all, across 2022-2023. senile_degeneration_of_brain (G31.1) has 4 Georgia visits in 2016-2024, while it is a growing underlying cause of death in dementia_deaths: hospital and death records code dementia differently, so compare causes across the two with care. creutzfeldt_jakob_disease is OASIS's A81, the whole category of atypical virus infections of the central nervous system. Race groups follow the platform convention from OASIS's race x ethnicity cross: hispanic is Hispanic of any race; white, black, asian, native_american (American Indian or Alaska Native), pacific_islander (Native Hawaiian or Other Pacific Islander) and multiracial are non-Hispanic. race_unknown is our calculation, the total minus those seven (unknown ethnicity, or non-Hispanic of unknown race), so the race groups add up to all; it has no population and so no rate. 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 124 county-years (53 counties, never Georgia). Ethnicity is recorded from 2016, but some hospitals leave it out: 13.6% of Georgia's visits in 2016 are race_unknown (unknown ethnicity or race), 0.8% in 2024, and up to 92% of a county's (Hall, 2016: 79 visits), so the race groups' counts and rates run low, most in 2016; OASIS advises caution with Hispanic figures (https://oasis.state.ga.us/dtt/content/KDIADRDER.htm). Houston County's visits were coded non-Hispanic multiracial in 2017 only (34 of its 56 dementia ER visits, none in 2016 or 2018 in OASIS's figures, though the race mask NULLs Houston's 2016 race rows here; its all-cause ER visits in er_visits_by_cause jump the same year), so Georgia's 2017 multiracial figures (103 visits, age-adjusted 85.4, against 53 and 32 in 2016 and 2018; Houston 33% of them) are inflated: don't read them as a change, and Houston's own 2017 multiracial rate (2,452.6) is a coding artifact. Hospital records and OASIS's population estimates can classify race differently (OASIS's definitions say so), so a small group's rate can be far too high: the highest age-adjusted rate, 5,253.2 (asian, Washington County, 2023: 7 visits, population 96), is 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_er_visits / num_population x 100,000 reproduces each published crude rate); the race groups' populations add to all, as do male and female. 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 SES-vulnerability rows and payer 'rates' (each divides by the whole place's population; payer counts stay in bronze), and age breakdowns. tract_geoid is always NULL.