Emergency Room Visits by Cause (OASIS Detailed Causes, Nested)

Limitations

ER visits by Georgia residents at Georgia's non-federal hospitals, by home county, not ending in admission (those are hospital_discharges_by_cause): a person who visits twice counts twice; out-of-state visits are missing, so border counties look low. Causes nest: never add them. OASIS withholds the all-cause age-adjusted rate for 2-84 counties a year (never Georgia): use er_visit_rate_per_100k. Race groups start in 2016 and run low (unknown ethnicity is only in race_unknown; NULL above 20%). Some counties' visits drop for a year (missing hospital records, our reading): compare a county's years with care. OASIS withholds rates on 1-4 visits (NULL here). 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). OASIS codes the withheld all-cause age-adjusted rate -3: in at least one of the place's age groups its ER visits outnumber its residents (a person can visit several times; OASIS shows no age-specific figure, so which group is not known). It was withheld for the fewest counties in 2020 (2), when visits fell. 33% of the other rows with at least one visit have no age-adjusted rate, mostly 1-4 visits in small counties, small groups and rare causes. An ER visit is one visit to the emergency room of a non-federal hospital in Georgia, by the patient's county of residence, that did not end in an inpatient admission: OASIS counts a patient seen in the ER and admitted as a hospital discharge (hospital_discharges_by_cause) and keeps the two mutually exclusive, so no visit is counted in both. A person who visits twice counts twice, and visits of Georgia residents to other states' hospitals are not counted, so counties near another state's hospitals may look low. Counties beside out-of-state hospitals (Chattanooga, Tennessee; Eufaula, Alabama) show it most: all-cause crude rates per 100,000 in 2024 of Dade 3,122.1, Catoosa 10,455.7, Quitman 7,146.0, against Georgia's 35,977.1 (our reading: their residents use those hospitals). Some counties' visits drop for one year and recover, which looks like a hospital's ER records missing (our reading; kept as published, and in Georgia's total): the 11 county-years under half the mean of the years either side are Miller 2003, Burke 2004, Washington 2006, Walker, Catoosa and Dade 2007, Wayne 2009, Atkinson and Coffee 2011, Candler 2013, Tift 2015 (Walker: 15,713 visits in 2006, 1,512 in 2007, 15,927 in 2008; Candler: 5,481, 1,854 and 6,918 in 2012-2014, and OASIS notes Candler's hospital did not report discharges in 2013); Washington's visits run 1,525-1,992 a year through 2006 and 7,999-12,088 from 2007. OASIS also notes, for discharges, that Hart County Hospital closed in July 2012 without reporting that year; Hart's ER visits drop too (9,654 in 2011, 5,124 in 2012). Compare a county's years with care (https://oasis.state.ga.us/dtt/content/KDIMorb.html). Six southwest Georgia counties (Brooks, Colquitt, Decatur, Grady, Mitchell and Thomas) show cancer and anemia visits many times their other years' in 2011-2013 (Thomas County's cancer visits: 12 in 2010, 510 in 2012, 31 in 2014; the six had 1,353 of Georgia's 3,606 cancer visits in 2011): kept as published, probably a recording change (our reading). OASIS's hospital data records no ethnicity before 2016 (every visit of earlier years is unknown ethnicity), so the race groups, which are built from race x ethnicity (hispanic of any race, the rest non-Hispanic), start in 2016; earlier years have all, male and female rows only. From 2016 some hospitals still record no ethnicity: 5.9% (2024) to 16.4% (2016) of Georgia's visits a year are race_unknown, almost all of unknown ethnicity, and in some counties far more (up to 98% of a county's visits in a year), so the race groups' counts and rates run low, unevenly by county and year; OASIS itself advises caution with Hispanic figures. Small race groups' rates can also run implausibly high where hospital records and the Census estimates classify race differently: OASIS withholds the crude rate on 622 race-group rows whose visits outnumber the group's residents (Houston County's multiracial residents in 2017: 35,151 visits, population 5,559), and Georgia's native_american crude rate is 65,505.3 in 2021 against 36,640.3 in 2024. Visits outnumber the group on no all, male or female row but Clinch County's women in 2012-2014 (3,470 visits, 3,466 women in 2013): OASIS withholds both rates there. 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 146 county-years (49 counties, never Georgia). race_unknown is our calculation, all minus the seven race groups (unknown ethnicity, or non-Hispanic of unknown race): a count with no population or rate. Causes are coded in ICD-9-CM until September 2015 and ICD-10-CM from October 2015, which breaks some causes' series (OASIS's reporting-issues page), 2015 mixing both: Georgia's substance_use_disorders 4,329 in 2014 and 38,036 in 2017; other_mental_and_behavioral_disorders 102,578 in 2014 and 76,127 in 2017; hypertensive_heart_disease 221 in 2014 and 4,654 in 2017; hypertensive_heart_and_chronic_kidney_disease 106 in 2014 and 3,242 in 2017; hiv_disease 593 in 2014 and 293 in 2017; respiratory_distress_syndrome 16 in 2014 and 88 in 2017; accidental_shooting 1,272 in 2014 and 2,005 in 2017; other_chronic_lower_respiratory_diseases 3,971 in 2014 and 2,694 in 2017; unintentional_poisoning 8,806 in 2014 and 13,361 in 2017; alcoholic_liver_disease 312 in 2014 and 567 in 2017. lack_of_oxygen_to_the_fetus has no rows from 2016 (Georgia: 1 visit in 2014, 3 in 2015): OASIS defines it as P20-P21, codes ICD-10-CM does not have, so no visit can carry it (rows dropped, not served as zeros). sids is 0 in every place from 2016 (Georgia: 4 visits in 2014, 3 in 2015; the break is undocumented, and ICD-10-CM has R95, so its 0s are kept as published): its 0s from 2016 are probably not real absences (our reading). Neither has a series after 2015. Within a group the listed causes never overlap but need not add up to it: they do exactly only in endocrine_nutritional_and_metabolic_diseases, mental_and_behavioral_disorders, nervous_system_diseases, genitourinary_diseases; elsewhere the rest of the group is in no listed cause (heart failure in major_cardiovascular_diseases, for example). The 15 groups never overlap either, and 33% to 37% of Georgia's visits a year are in none of them, such as symptoms without a diagnosis (chest or abdominal pain), pregnancy, and skin, eye and ear conditions (our reading: OASIS's list has no group for them). Visits fell from 4,159,155 in 2019 to 3,246,307 in 2020 (the COVID-19 pandemic). On all rows, OASIS's rates of breast, cervical, uterine, ovarian, prostate and testicular cancer divide by the whole population, both sexes; the male and female rows divide by that sex's population. 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. 2025, the newest year, was added to OASIS on 2026-09-03; OASIS marks no year provisional. Counties add up to Georgia, except race groups from 2016, 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 deduplicated measures (people rather than stays or visits), its SES-vulnerability rows and payer 'rates' (each divides by the whole place's population; payer counts and shares are their own topic), age and education breakdowns, and the significance-flag and percent measures. tract_geoid is always NULL.