HOUSTON HOSPITALS EMERGENCY DEPARTMENT USE STUDY. January 1, 2009 through December 31, 2009 FINAL REPORT. Prepared By

HOUSTON HOSPITALS EMERGENCY DEPARTMENT USE STUDY January 1, 2009 through December 31, 2009 FINAL REPORT Prepared By School of Public Health University of Texas Health Science Center at Houston Charles Begley, Patrick Courtney, Keith Burau May 2011

TABLE OF CONTENTS I. Executive Summary...3 II. Purpose...13 III. Methods...14 IV. Results,....18 A. Total ED Visits.................18 B. Primary Care Related ED Visits.....21 C. ED Visits by Month, Day, and Time....... 24 D. ED Visits by Patient Characteristics....... 31 E. Health Conditions of Patients with ED Visits........48 F. Behavioral Health Related ED Visits.....48 G. Length of Stay in the ED..... 52 H. Estimated Cost of Primary Care Related ED Visits. 52 I. Geographic Distribution of Patients with ED Visits..........58 J. Trends in ED Visits Over Time..68 V. Limitations........ 76 2

I. EXECUTIVE SUMMARY The University Of Texas School Of Public Health collects and analyzes emergency department (ED) visit data in Harris County hospitals and monitors primary care related (PCR) use of the ED. The main purpose of this report is to provide information on the frequency and type of ED visits in Houston hospitals in 2009, the medical conditions for which visits were made, the characteristics of patients that visited the ED, and the trends in ED use over the last five years. Data and Analysis Twenty four hospitals with EDs that serve the general public of Harris County (provide hospital based ED care and accept 911 deliveries) participated in the 2009 hospital ED study. They include: two hospitals of the Harris County Hospital District (Ben Taub General and Lyndon B. Johnson General); nine hospitals of the Memorial Hermann Health Care System (Hermann/Texas Medical Center, Southwest, Southeast, Northeast, Northwest, The Woodlands, Memorial City, Katy, and Sugar Land); four hospitals of the Hospital Corporation of America (Bayshore Medical Center, Spring Branch Medical Center, East Houston Regional Medical Center, and West Houston Medical Center); Texas Children s Hospital; two hospitals of CHRISTUS Gulf Coast (St. Catherine and St. John); two hospitals of the St. Luke s hospital system (Episcopal Hospital and Community Medical Center); and four hospitals of the Methodist Hospital System (Methodist/Texas Medical Center, San Jacinto Methodist, Methodist Willowbrook, and Methodist Sugar Land). Each hospital supplied the following information on each ED visit during the period January 1, 2009 through December 31, 2009: 3

1. Date and time of admission to ED 2. Primary and all secondary discharge diagnoses 3. Discharge date and time 4. Payment source 5. Patient age 6. Patient gender 7. Patient race/ethnicity 8. Patient ZIP code 9. Where discharged (e.g. home, nursing home, etc.) 10. Method of transport (e.g. ambulance, self, other) Based on this information, the number, type, and rate of ED visits to all Houston hospitals for various population groups was estimated and the number and percentage of primary care-related ED (PCRED) visits was estimated by applying the New York University ED Algorithm to each visit. The NYU Algorithm determines the likelihood that a visit is one of the following types. 1. Non-emergent: Immediate treatment was not required within 12 hours. 2. Emergent-Primary Care Treatable: Treatment was required within 12 hours, but could have been provided effectively and safely in a primary care setting. Continuous observation was not required, no procedures were performed or resources used that are not typically available in a primary care setting. 3. Emergent-ED Care Needed-Preventable/Avoidable: ED care was required within 12 hours, but the emergent nature of the condition was potentially preventable/ avoidable if timely/continuous primary care had been received for the underlying illness. 4. Emergent-ED Care Needed-Not Preventable/Avoidable: ED care was required within 12 hours and primary care could not have prevented the condition. The demographic, coverage, and health conditions of patients with ED visits and PCRED visits was determined. ED visits were mapped by ZIP code, population-based PCRED visit rates were calculated, and the pattern over time of ED visits and PCRED visits was summarized and assessed. 4

2009 Highlights 1. Total ED Visits a. During 2009, 1,127,557 ED visits were made to all 25 participating hospitals. Of this total, 84.8% were made by Harris County residents, 12.9% by residents of the seven surrounding counties, and 2.3% by patients from outside the area. Of all ED visits by residents of the eight county area, 84.0% were discharged home or elsewhere. Of all ED visits by Harris County residents, 72.9% were discharged home or elsewhere (Table 1a). b. The ED visits of participating hospitals represented 74.04% of all ED visits reported to the Texas Department of State Health Services by all hospitals in Harris County in 2009. Using this percentage for extrapolation, an estimated 1,287,710 ED visits were made by Harris County residents to all hospitals in the county. The ED visit rate of Harris County residents was 31.84 per 100 residents in 2009, exceeding the 30.91 and 30.0 ED visit rates in 2007 and 2008, respectively (Table 1a). 2. Primary Care Related ED Visits A little over 40% (41.0%) of ED visits by Harris County residents were PCR in 2009. This includes non-urgent, primary care treatable, and primary care preventable visits. The 2009 rate compares to 39.1% in 2008 and 41.1% in 2007 (Table 2). 3. ED Visit Rates by Population Sub-Group Females have a higher rate than males for both total visits and PCRED visits. Persons age 65 and over have the highest rate of total ED visits (34.33 per 100 elderly), but children and adolescents have the highest rate of PCRED visits (12.10 per 100 children) (Table 1b). Persons enrolled in Medicare have the highest rate of total ED visits (76.13 per 100 Medicare enrollees), but persons enrolled in Medicaid and/or CHIP have a slightly higher rate of PCRED visits (23.0 5

per 100 Medicaid/CHIP enrollees). Persons of other race or ethnicity have the highest rates of total visits and PCRED visits. Among those defined as one race, persons who were Black have the highest rate of total visits (37.43 per 100 Black persons) and PCRED visits (17.4 per 100 Black persons). 4. ED Visits by Month, Day, and Time The number of ED visits by Harris County residents was highest in May and September but vary relatively little from month to month (Figure 3). The percentage of ED visits that lead to hospitalizations was also relatively constant over the year. The peak days of the week for ED visits were Saturday and Sunday (Figure 4). The percentage of ED visits leading to hospitalizations remained relatively level each day. ED visits by children peaked between 7-8 PM, with a secondary peak at 11 AM-12 PM (Figure 5a). ED visits by adults age 18 to 64 peaked at noon, dropped slightly and remained fairly level until 7 PM (Figure 5b). ED visits by adults age 65 and over peaked from 12 PM to 4 PM, then dropped during the evening (Figure 5c). Relatively few ED visits by children resulted in a hospitalization, in contrast to adults age 65 and over. 5. ED Visits by Patient Characteristics a. Payer source: The uninsured accounted for 33.1% of PCRED visits and 29.2% of non- PCRED visits. Medicaid covered individuals made up 26.8% of PCRED visits and 21.2% of non-pcred visits. Children on CHIP made up 4.2% of PCRED visits (Figures 6 & 7). Patients with commercial insurance accounted for 24.1% of PCRED visits and 26.4% of non-pcred visits. Four in ten ED visits made by persons on Medicaid (39.8%) were PCR, and more than half of all ED visits by persons on CHIP (51.9%) were PCR. Medicare enrollees had the highest percentage of hospitalized ED visits (32.2%) (Figure 8). 6

b. Race/Ethnicity: Seven in ten (70.6%) PCRED visits were by persons who were Black or Hispanic, as were 60.2% of non-pcred visits. By contrast, persons who were White accounted for 24.1% of PCRED visits and 34.2% of non-pcred visits (Figures 9 & 10). Nearly half of all ED visits by persons who were black or Hispanic were PCR. Asians and Whites were most likely to have an ED visits resulting in a hospitalization (Figure 11). c. Age: Children age 17 and under made up more than one third (36.4%) of all PCRED visits, and just over one fourth (27.6%) of non-pcred visits. Adults age 65 and over were almost twice as likely to have non-pcred visits than a PCRED visit. Adults age 25-44 were more likely to have a PCRED visit than a non-pcred visit (Figures 12 & 13). More than three in ten (31.8%) ED visits by adults age 65 resulted in a hospitalization (Figure 14). d. Gender: Almost six in ten (58.5%) PCRED visits were by females and 41.5% by males (Figure 15). Of non-pcred visits, 52.8% were by females and 47.2% by males (Figure 16). The percentage of ED visits resulting in a hospitalization was higher among males (14.4%) than females (13.8) (Figure 17). e. Transport: One in ten (10.1%) PCRED visits were by persons with an ambulance transport to the ED. By contrast, more than one in five (21.8%) non-pcred visits were by persons with an ambulance transport to the ED (Tables 7a & 7b; Figures 18 & 19). One fourth (25.0%) of ambulance transports to the ED eventually resulted in a hospitalization (Table 7c and Figure 20) 6. Health Conditions of Patients with ED Visits The most common discharge diagnoses of patients with PCRED visits for the last three years has been acute upper respiratory infection not otherwise specified. The frequency of visits with this diagnosis was higher in 2009 than in previous years (Table 8). Chest pain not otherwise 7

specified was the leading diagnosis of non-pcred visits, and also increased in frequency in 2009 (Table 9). 7. Behavioral Health-Related ED Visits The percentage of behavioral health visits dropped from 8.3% in 2008 to 7.7% in 2009. The number of patients who were seen with an alcohol-related diagnosis and no other behavioral diagnosis nearly doubled in 2009 (1,079) compared to 2008 (585). The number of patients whose primary diagnosis was medical, but who had a secondary diagnosis of an alcohol problem, rose by about 50% (from 2,417 in 2008 to 3,689 in 2009) (Table 10a). 8. Length of Stay in the ED The mean length of stay of ED visits (measured from time of admission to the ED to time of discharge from the ED) was calculated for visits of less than 24 hours each. PCR visits last roughly half an hour less than other non-hospitalized visits (3.70 hours in 2009 vs. 4.34 hours in 2009). The length of stay for a hospitalized ED visit increased from 5.37 hours in 2007 to 6.84 hours in 2009 (Table 12). Non hospitalized behavioral health visits continue to last longer than non hospitalized medical visits (Table 11). 8. Estimated Cost of Primary Care Related ED Visits Applying national estimates of the unit cost of ED visits to Houston ED visits suggests that the cost of PCRED visits in Houston hospitals increased from $187,310,573 in 2007 to $214,859,768 in 2009 (Table 13). The greatest potential cost savings from diverting PCRED visits would be among uninsured adults age 18 to 44 (Table 12 and 13, Figures 21and 22). This is based on the fact that PCRED visits of adults age 18 to 44 represented 36.8% of total PCRED visits the estimated savings among all persons in 2009 and uninsured adults accounted for 54.2% of the estimated cost of PCRED visits savings among the uninsured in 2009 (Tables 12 and 13). 8

9. Geographic Distribution of Patients with ED Visits The following ZIP code clusters had the highest number of ED visits: 1) the intersection of Highway 59 and the North Loop East; 2) the Ship Channel and the area along Interstate 10 and the East Belt; 3) an area just east of Highway 288 and on both sides of the South Loop; 4) an area along the Southwest Freeway and between the West Loop and West Belt; 5) the intersection of Interstate 10 and Highway 6; and 6) an area along Highway 290 and just outside the North Loop West (Map 1a). Clusters 1, 4, and 5 had the biggest increases from 2007 to 2009 (Map 1b). The geographic distribution of PCRED visits by the insured and uninsured was similar to that of ED visits (Maps 2a and 3a). An area between the North Loop and the North Belt had the biggest increase in PCRED visits by the insured and uninsured (area shaded in orange) compared to the increase by the uninsured (area shaded in yellow). (Maps 2b and 3b). The areas with the highest rates of ED visits, PCRED visits by all, and PCRED visits by the uninsured were the East/Northeast and South Central areas of Harris County (Map 4). The population rates of PCRED visits for all and the uninsured, were lowest in the Baytown area (Maps 4, 5, and 6). 10. Trends in ED Visits Over Time To examine trends over time, the ED visits in which patients were discharged home or elsewhere of 15 hospitals were tracked over the five-year period from 2004 to 2009 (Ben Taub General, Lyndon B. Johnson General, Memorial Hermann (MH) Hermann, MH Southwest, MH Southeast, MH Northwest, MH Woodlands, MH Memorial City, MH Katy, MH Sugar Land (formerly Fort Bend), Hospital Corporation of America (HCA) Bayshore, HCA Spring Branch, HCA East Houston, Texas Children s Hospital, St. Luke s Episcopal Hospital). The ED visits of 9

22 hospitals were tracked over the four-year period from 2005 to 2009 (above listed hospitals, plus MH Northeast, HCA West Houston, CHRISTUS St. Catherine, CHRISTUS St. John, St. Luke s Community Medical Center, The Methodist Hospital, and San Jacinto Methodist Hospital). There was a downward trend in non-hospitalized ED visits from 2004 through 2006 followed by an increase in 2008 and 2009. The number of PCRED visits was highest in 2009. The percentage of categorized visits that were PCR consistently declined over the period but increased in 2009. For the 15 hospitals, the PCRED visit percentage was 80.9% in 2009, compared to 72.4% in 2007 (Table 14). For the 22 hospitals, the percentage was 80.4% in 2009 (Table 15). Visits by persons with Medicare represented a small but steadily increasing percentage of PCRED visits for most of the study years, but decreased in 2009 (Figures 23 and 24). The uninsured have continued to account for the highest percentage of PCRED visits. Individuals with Medicaid have fluctuated between 25% and 32% of total PCRED visits. The percentage of PCRED visits by persons with private insurance was at its peak in 2006. Hispanics are steadily increasing as a percentage of PCRED visits. In both groups of hospitals they represented approximately 40% of PCRED visits in 2009 (Figures 25 & 26). Blacks represent the next largest group on a percentage basis. Visits by children age 18 and under continue to account for the largest proportion of PCRED visits. In 2008 that proportion dropped somewhat, but increased again in 2009 (Figures 27 & 28). Adults age 26 to 44 accounted for the next largest percentage. There has been no change in the gender ratio of PCRED visits over the last five years (Figures 29 & 30) 10

Conclusion ED visits of Harris County residents in area hospitals increased in 2009 both in terms of the total number and population rates. The percentage of visits that were PCR rose slightly when compared with 2008. The same areas of the county had the highest frequencies and rates of ED visits and PCRED visits as in past years. The ZIP codes with the largest increases in PCRED visits by the uninsured tend to be outside the 610 Loop. The East/Northeast and South/Central sections of Harris County continue to have the highest rates of PCRED visits by the uninsured. PCRED visits had a shorter length of stay than non-pcred visits. Behaviorrelated ED visits took longer to treat than medical visits. Diversion of PCRED visits of adults age 18 to 44 have the largest potential for cost-savings, especially among the uninsured. Children have the highest rate of PCRED visits as do Medicaid and CHIP enrollees. Further exploration of Census data is recommended to compare children with commercial insurance to children who are on Medicaid and CHIP, and to compare Medicare enrollees above 65 with Medicare enrollees below 65. Continued participation by a large number of hospitals has allowed for monitoring of community-wide trends ED and PCRED utilization in Harris County. The trend analysis over the last five years suggests the following: a) the need to expand outpatient primary and specialty care capacity, especially for the uninsured, b) the need for patient education about appropriate ED use, c) the need for after hours clinics for children, d) the need for hospital staffing at peak hours of ED use, and e) the need for efforts to link the uninsured, as well as Medicaid and CHIP enrollees to a medical home. Although the highest percentages and rates of PCRED are for the uninsured, Medicaid, and CHIP, the fact that a high number and rate of PCRED visits are by persons with private insurance or Medicare should not be ignored. Understanding trends in ED 11

patterns for all populations suggests that multiple diversion strategies are needed targeting multiple population groups. Limitations The NYU ED algorithm does not classify hospitalized visits, mental health, drug/alcohol, and injury-related visits. It may be assumed that most hospitalized visits and a large percentage of mental health, drug/alcohol, and injury visits are not PCRED visits. To the extent that some are PCR, the estimates understate the number of PCRED visits. This limitation does not affect the percentage or rate calculations. Because the data submitted is de-identified, it is not possible to identify patients who make multiple ED visits. Therefore, we are not able to identify frequent users of the ED. In past years it was believed that Medicaid ED visits also included CHIP. A review of the 2007 data has revealed that most CHIP visits were counted as commercial or private insurance by participating hospitals. In 2009, because some of the commercial visits could not be separated from CHIP visits, actual CHIP visits may be slightly higher in frequency than reported. New ZIP codes took effect during 2008 in the Baytown and Fort Bend County areas. For the aggregate report, what may appear to be a drop in visits from 77520 (for example) is a reflection of the fact that part of that ZIP code is now 77523, which is in Chambers County. 12

II. PURPOSE The increasing number of hospital ED visits, many of which are PCR, is leading to a multitude of associated issues regarding equity and access to care in the U.S. One of the most pressing issues in Houston is the capacity of hospitals to provide emergency care when emergency rooms are crowded with patients seeking basic care. The main purpose of this study is to provide information on the frequency, type, and distribution of ED visits in Houston hospitals that are PCR. To achieve this purpose, the study obtained ED data from twenty five hospitals in Houston for 2008, classified the visits of Harris County residents in terms of PCR/non-PCR use of the ED, and examined the demographic, coverage, and geographic characteristics of patients making the visits. The goal is to replicate the study over time in order to determine trends and evaluate primary care enhancement activities. 13

III. METHODS The study initially resulted from a partnership between Gateway to Care, the Harris County Hospital District (HCHD), and The University of Texas School of Public Health (UTSPH). In 2002, UTSPH worked with Gateway to Care on a pilot study to develop a process for monitoring ED use in Houston. The process developed in the pilot study was then applied to 11 hospitals in 2002. An increasing number of hospitals have participated each year. The current report uses data from 25 of the hospitals that serve the Houston 911 service area. Aggregate reports for each year are posted on the website for the Health Services Research Collaborative. The first step in the study involves requesting the following information on ED visits in Houston hospitals: 1. Date and time of admission to ED 2. Primary and secondary discharge diagnosis 3. Discharge date and time 4. Payment source (payer codes from the Patient Data Set of the Texas Hospital Association and the Texas Health Care Information Council) 5. Patient age 6. Patient gender 7. Patient race/ethnicity (Black, Asian, American Indian, Hispanic, White, Other, Unknown) 8. Patient ZIP code 9. Where discharged to (e.g. home, hospital, etc) 10. Method of transport (e.g. ambulance, self, other) Working with the hospitals, a dataset was obtained, reviewed, and cleaned comprising a full set of ED visit information for the period January 1, 2008 December 31, 2008. Visits that did not result in a hospitalization were analyzed. The second step involved the application of the New York University ED Classification Algorithm to classify ED visits of Harris County residents into the following four categories: 1. Non-emergent: Immediate treatment was not required within 12 hours. 14

2. Emergent-Primary Care Treatable: Treatment was required within 12 hours, but could have been provided effectively and safely in a primary care setting. Continuous observation was not required, no procedures were performed or resources used that are not typically available in a primary care setting. 3. Emergent-ED Care Needed-Preventable/Avoidable: ED care was required within 12 hours, but the emergent nature of the condition was potentially preventable/avoidable if timely/continuous primary care had been received for the underlying illness. 4. Emergent-ED Care Needed-Not Preventable/Avoidable: ED care was required within 12 hours and primary care could not have prevented the condition. The NYU Center for Health and Public Service Research and the United Hospital Fund of New York developed the ED Algorithm as a measure of PCR ED use. The ED Algorithm is a set of probabilities that when applied to the primary diagnosis (ICD-9 code) of the patient estimates the likelihood that the patient s ED visit was one or more of the types described above. The ED algorithm was developed with the advice of a panel of ED physicians and is based on information abstracted from a sample of complete ED records 3,500 cases in 1994 and 2,200 cases in 1999 from six Bronx, New York hospitals. The decision tree followed by the panel is summarized on the next page. The distribution of ED visits by type represents the weighted sum of all visits with a certain probability of being that type. ED visits in the first three categories are considered PCR use of the ED, while those in the fourth category reflect non-pcr use of the ED. A number of visits are not categorized using the Algorithm. These include injury, mental health-related, and alcohol or drug-related visits, and visits with missing data. The ED Algorithm for these visits has not yet been developed by the NYU researchers. 15

ED CLASSIFICATION PROCESS Step 1 Steps 2 and 3 Step 4 Not Preventable/Avoidable ED Care Needed Emergent Preventable/Avoidable Primary Care Treatable Non-emergent Primary Care Treatable Various analyses were conducted of the classified visit data to determine patterns of PCR and non-pcr visits. These included monthly, daily, and time of day patterns of visits; the distribution of visits by the coverage and demographic characteristics of patients (payment source, race/ethnicity, age, employment status, and sex); the distribution of visits by primary diagnosis (ICD9 Codes); and the distribution of visits by ZIP code of patient residence. The results of the analysis of ED visits should be treated cautiously and are best viewed as indicators of utilization rather than a definitive assessment. This is because only a portion of all visits are categorized by the Algorithm. ED visits that result in a hospitalization usually encompass no more than 10-20% of total visits. 1 Given these limitations in the methods, the percentage of visits that fall into the PCR categories should be interpreted as a conservative estimate. In order to get a complete picture of ED utilization, attention must also be paid to injuries, mental health and substance abuse, and hospitalized visits. In an effort to address this 1 Billings J, Using administrative data to monitor access, identify disparities, and assess performance of the safety net, U.S. Agency for Healthcare Research and Quality, 2003. 16

broader picture, the report examines not only PCR visits, but non primary care related visits as well. 17

IV. RESULTS A. Total ED Visits by Type During 2009, 1,127,557 ED visits were made to all 25 participating hospitals. Of this total, 84.8% were made by Harris County residents, 12.9% by residents of the seven surrounding counties, and 2.3% by patients from outside the area. Of all ED visits by residents of the eight county area, 84.0% were discharged home or elsewhere. Of all ED visits by Harris County residents, 72.9% were discharged home or elsewhere (Table 1a). The ED visits of participating hospitals represented 74.04% of all ED visits reported to the Texas Department of State Health Services by all hospitals in Harris County in 2009. Using this percentage for extrapolation, an estimated 1,287,710 ED visits were made by Harris County residents to all hospitals in the county. The ED visit rate of Harris County residents was 31.84 per 100 residents in 2009, exceeding the 30.91 and 30.0 ED visit rates in 2007 and 2008, respectively (Table 1a). Females have a higher rate than males for both total visits and PCRED visits. Persons age 65 and over have the highest rate of total ED visits (34.33 per 100 elderly), but children and adolescents have the highest rate of PCRED visits (12.10 per 100 children) (Table 1b). Persons enrolled in Medicare have the highest rate of total ED visits (76.13 per 100 Medicare enrollees), but persons enrolled in Medicaid and/or CHIP have a slightly higher rate of PCRED visits (23.0 per 100 Medicaid/CHIP enrollees). Persons of other race or ethnicity have the highest rates of total visits and PCRED visits. Among those defined as one race, persons who were Black have the highest rate of total visits (37.43 per 100 Black persons) and PCRED visits (17.4 per 100 Black persons) (Table 1b). 18

Table 1a. Total ED Visits to Hospitals in Harris County Study Hospital Visits 25 Hospitals 2007 25 Hospitals 2008 24 Hospitals - 2009 Frequency Percent of first row Percent of Frequency first row Percent of Frequency first row A Total 1 (all areas) 2 1,060,825 100.00% 1,060,592 100.00% 1,127,557 100.00% B Total Discharged to Home or Self (all areas) 3 897,264 84.58% 931,721 87.85% 961,707 85.29% C Total Admitted (all areas) 3 163,561 15.42% 128,871 12.15% 165,850 14.71% D Total (eight county area) 3 984,994 92.85% 1,013,343 95.55% 1,094,529 97.07% E Total Discharged to Home or Self (eight county area) 3 835,251 78.74% 890,817 83.99% 934,563 82.88% F Total Admitted (eight county area) 3 149,743 14.12% 122,526 11.55% 159,966 14.19% G Total (Harris County) 901,684 85.00% 899,673 84.83% 953,395 84.55% H Total Discharged to Home or Self (Harris County) 761,077 71.74% 789,995 74.49% 812,863 72.09% I Total Admitted (Harris County) 3 140,607 13.25% 109,678 10.34% 140,532 12.46% All Hospital Visits All Hospitals 2007 4 All Hospitals 2008 All Hospitals 2009 J Total All 5 (all areas) 2 1,414,863 1,401,423 1,522,943 K Total Visits as a Percentage of Total All Visits 74.98% 75.68% 74.04% L Estimated Number of Total That Are From Harris Co. 6 1,202,611 1,188,791 1,287,710 M Est. Number of Discharged That Are From Harris Co. 7 1,015,078 1,043,867 1,097,899 N Est. Number of Admitted That Are From Harris Co. 8 187,533 144,924 189,811 O Population of Harris County 9 3,891,240 3,965,716 4,044,032 P Harris Co. ED Rate per 100 population (adm. and disch.) 30.91 29.98 31.84 Q Total Rate of Discharged ED Visits per 100 26.09 26.32 27.15 R Total Rate of Admitted ED Visits per 100 4.82 3.65 4.69 1 Total visits reported by hospitals participating in this report. 2 All areas = all states, all countries. 3 22,691 total visits in 2008 lacked a primary ICD9 code, 20,533 of these were in Harris County. 4 Represents a correction to report on 2007 ED visits. 5 Source: Cooperative DSHS/AHA/THA Annual Survey of Hospital and Tracking Database. 6Row J x percentage in Row G. 7Row J x percentage in Row H. 8Row J x percentage in Row I. 9 Source: Texas State Data Center. http://txsdc.utsa.edu/tpepp/txpopest.php 19

Table 1b. 2009 ED Visits to Hospitals in Harris County by Demographics, 2009 Population Rates 1 For Demographics Males Females 2009 total population (by gender) 2,043,393 2,027,596 2009 total ED visits (by gender) 417,734 514,564 2009 total PCRED visits (by gender) 162,202 228,716 2009 total ED rate per 100 (by gender) 20.44 25.38 2009 PCRED rate per 100 (by gender) 7.94 11.28 0-17 18-44 45-64 65 and over 2009 total population (by age) 1,174,860 1,641,519 926,836 327,774 2009 total ED visits (by age) 291,892 342,380 185,602 112,511 2009 total PCRED visits (by age) 142,210 148,674 69,914 30,147 2009 total ED rate per 100 (by age) 24.84 20.86 20.03 34.33 2009 PCRED rate per 100 (by age) 12.10 9.06 7.54 9.20 Commercial Medicaid/SCHIP Medicare Uninsured 2009 total population (by payer) 2 2,111,270 527,469 162,607 1,093,235 2009 total ED visits (by payer) 237,291 251,068 123,786 287,290 2009 total PCRED visits (by payer) 94,500 121,166 34,270 129,434 2009 total ED rate per 100 (by payer) 11.24 47.60 76.13 26.28 2009 PCRED rate per 100 (by payer) 4.48 22.97 21.08 11.84 Asian Black Hispanic Am. Indian Other White 2009 total population (by race) 3 232,385 730,412 1,619,426 6,455 51436 1430875 2009 total ED visits (by race) 16,286 273,373 328,973 757 24505 279543 2009 total PCRED visits (by race) 6,040 127,095 149,100 331 9,698 94,205 2009 total ED rate per 100 (by race) 7.01 37.43 20.31 11.73 47.64 19.54 2009 PCRED rate per 100 (by race) 2.60 17.40 9.21 5.12 18.85 6.58 1 Population data comes from the 2009 American Community Survey 1 Year Estimates, Census Bureau. 2 Payer: For ED visits, commercial and other private were combined, as were Medicaid and CHIP. For ACS population, the following were combined: commercial (employer provided, direct purchase, employer and direct purchase, employer and Medicare, other private only), Medicare (Medicare only, Medicare and Medicaid). Even though ACS population is available for Tricare and VA coverage, the majority of other government ED visits are usually worker s comp. The ACS data did not have a separate category for CHIP. 3 Race: For ACS data, the following were combined: other (Native Hawaiian, Pacific Islander, some other race alone, two or more races). 20

B. Primary Care Related ED Visits A little over 40% (41.0%) of ED visits by Harris County residents were PCR in 2009. This includes non-urgent, primary care treatable, and primary care preventable visits. The 2009 rate compares to 39.1% in 2008 and 41.1% in 2007 (Table 2). 21

Table 2. ED visits to Hospitals by Harris County Residents Type of Visit 25 Hospitals 2007 25 Hospitals 2008 24 Hospitals 2009 Frequency TOTAL VISITS Total All Visits, Harris County (discharged and admitted) 1 901,684 899,673 953,395 Total Discharged Visits (DC), Harris County 1 761,077 789,995 812,863 Total Admitted Visits, Harris County 1 140,606 109,678 140,532 CATEGORIZED VISITS (DC) Non-Emergent 2 145,668 141,507 158,180 Emergent, Primary Care Treatable 2 166,693 161,033 180,389 Emergent, ED Care Needed - Preventable/Avoidable 2 58,334 48,903 52,376 Total Primary Care Related Visits, eight county area 370,695 351,443 390,945 Emergent, ED Care Needed - NOT Preventable/Avoidable 85,034 94,429 95,463 Total Categorized Visits eight county area 455,729 445,872 486,408 NON-CATEGORIZED VISITS (DC) Injury 166,531 193,591 164,243 Mental Health Related 10,720 12,382 10,227 Alcohol or Drug Related 5,648 4,389 4,973 Unclassified 3 96,525 219,902 135,390 Total Non-Categorized Visits, eight county area 279,424 430,264 314,833 Percent PERCENTAGE OF CATEGORIZED ED VISITS (DC) Non-Emergent 32.0% 31.7% 32.5% Emergent - Primary Care Treatable 36.6% 36.1% 37.1% ED Care Needed - Prev./Avoid. 12.8% 11.0% 10.8% % Categorized That Are Primary Care Related 81.3% 78.8% 80.4% ED Care Needed - NOT Prev./Avoid. 18.7% 21.2% 19.6% PERCENTAGE OF TOTAL ED VISITS Non-Emergent (DC) 16.2% 15.7% 16.6% Emergent - Primary Care Treatable (DC) 18.5% 17.9% 18.9% ED Care Needed - Prev./Avoid. (DC) 6.5% 5.4% 5.5% % Total That Are Primary Care Related 41.1% 39.1% 41.0% ED Care Needed - NOT Prev./Avoid. (DC) 9.4% 10.5% 10.0% PERCENTAGE OF TOTAL ED VISITS Injury (DC) 18.5% 21.5% 17.2% Mental Health Related (DC) 1.2% 1.4% 1.1% Alcohol or Drug Related (DC) 0.6% 0.5% 0.5% Unclassified (DC) 3 10.7% 24.4% 14.2% Admitted Visits (all causes) 15.6% 12.2% 14.7% PERCENTAGE OF DISCHARGED ED VISITS Non-Emergent 19.1% 17.9% 19.5% Emergent - Primary Care Treatable 21.9% 20.4% 22.2% ED Care Needed - Prev./Avoid. 7.7% 6.2% 6.4% % Discharged That Are Primary Care Related 48.7% 44.5% 48.1% ED Care Needed - NOT Prev./Avoid. 11.2% 12.0% 11.7% PERCENTAGE OF DISCHARGED ED VISITS Injury 21.9% 24.5% 20.2% Mental Health Related 1.4% 1.6% 1.3% Alcohol or Drug Related 0.7% 0.6% 0.6% Unclassified 12.7% 27.8% 16.7% 1 May contain missing data. Therefore, percentages at the bottom will not total to 100%. 2 Primary care related. 3 The rise in unclassified visits in 2008 was due primarily to a coding problem. 22

Figure 1. 2009 Categorized ED Visits by Type (N=24) ED Care Needed - NOT Prev./Avoid., 19.6% Non-Emergent, 32.5% ED Care Needed - Prev./Avoid., 10.8% Emergent - Primary Care Treatable, 37.1% Figure 2. 2009 Total ED Visits by Type Unclassified (DC), 14.2% Admitted Visits (all causes), 14.7% Non- Emergent (DC), 16.6% Emergent - Primary Care Treatable (DC), 18.9% Alcohol or Drug Related (DC), 0.5% ED Care Needed - Prev./Avoid. (DC), 5.5% Mental Health Related (DC), 1.1% Injury (DC), 17.2% ED Care Needed - NOT Prev./Avoid. (DC), 10.0% 23

C. ED Visits by Month, Day, and Time The number of ED visits by Harris County residents was highest in May and September but vary relatively little from month to month (Figure 3). The percentage of ED visits that lead to hospitalizations was also relatively constant over the year. The peak days of the week for ED visits were Saturday and Sunday (Figure 4). The percentage of ED visits leading to hospitalizations remained relatively level each day. ED visits by children peaked between 7-8 PM, with a secondary peak at 11 AM-12 PM (Figure 5a). ED visits by adults age 18 to 64 peaked at noon, dropped slightly and remained fairly level until 7 PM (Figure 5b). ED visits by adults age 65 and over peaked from 12 PM to 4 PM, then dropped during the evening (Figure 5c). Relatively few ED visits by children resulted in a hospitalization, in contrast to adults age 65 and over. 24

Figure 3 25

Figure 4 26

Figure 5. All ages 27

Figure 5a. Age 0-17 28

Figure 5b. Age 18-64 29

Figure 5c. Age 65 and over 30

D. ED Visits by Patient Characteristics Tables 3a, 4a, 5a, and 6a represent the frequency counts for ED visits by payer source, race, age, and gender (respectively). Tables 3b, 4b, 5b, and 6b represent percentages of these characteristics when summed horizontally (e.g. total hospitalized visits by payer source). Tables 3c, 4c, 5c, and 6c represent percentages of these characteristics in four broad categories when summed vertically (e.g. total visits by persons who are Hispanic). Payer source: The uninsured accounted for 33.1% of PCRED visits and 29.2% of non- PCRED visits. Medicaid covered individuals made up 26.8% of PCRED visits and 21.2% of non-pcred visits. Children on CHIP made up 4.2% of PCRED visits (Figures 6 & 7). Patients with commercial insurance accounted for 24.1% of PCRED visits and 26.4% of non-pcred visits. Four in ten ED visits made by persons on Medicaid (39.8%) were PCR, and more than half of all ED visits by persons on CHIP (51.9%) were PCR. Medicare enrollees had the highest percentage of hospitalized ED visits (32.2%) (Figure 8). Race/Ethnicity: Seven in ten (70.6%) PCRED visits were by persons who were Black or Hispanic, as were 60.2% of non-pcred visits. By contrast, persons who were White accounted for 24.1% of PCRED visits and 34.2% of non-pcred visits (Figures 9 & 10). Nearly half of all ED visits by persons who were black or Hispanic were PCR. Asians and Whites were most likely to have an ED visits resulting in a hospitalization (Figure 11). Age: Children age 17 and under made up more than one third (36.4%) of all PCRED visits, and just over one fourth (27.6%) of non-pcred visits. Adults age 65 and over were almost twice as likely to have non-pcred visits than a PCRED visit. Adults age 25-44 were more likely to have a PCRED visit than a non-pcred visit (Figures 12 & 13). More than three in ten (31.8%) ED visits by adults age 65 resulted in a hospitalization (Figure 14). 31

Table 3a. Frequency of ED Visit Types by Payment Source (N=24) Type of Visit Comm SCHIP M'Caid M'care Other Gov. Other Priv. Unins Unk All Total Visits (hospitalized & non hospitalized) 237,079 31,589 219,479 123,786 27,755 212 287,290 5,195 932,385 Total Hospitalized Visits 22,392 1,817 25,057 39,841 3,165 20 38,204 648 131,144 Total Primary Care Related 94,393 16,404 104,762 34,270 8,721 107 129,434 2,854 390,945 Total Injury/BH/Unclass. Visits 89,113 10,848 72,299 36,316 13,768 62 91,318 1,109 314,833 Tot. non Primary Care Related (Total minus PCR) 142,686 15,185 114,717 89,516 19,034 105 157,856 2,341 541,440 Non-Emergent 36,722 6,464 40,656 12,033 4,014 76 56,608 1,608 158,180 Emergent-Primary Care Treatable 44,956 7,830 48,961 15,569 3,689 28 58,257 1,099 180,389 Emergent Care Needed-Preventable/Avoidable 12,715 2,109 15,145 6,669 1,018 3 14,569 147 52,376 Emergent Care Needed-NOT Prev./Av. 31,181 2,520 17,361 13,359 2,101 23 28,334 584 95,463 TOTAL Categorized Visits 125,574 18,924 122,123 47,629 10,822 130 157,768 3,438 486,408 Table 3b. Percentage of ED Visit Types by Payment Source (read horizontally) Type of Visit Comm SCHIP M'Caid M'care Oth. Gov Oth. Priv Unins All Total Visits (hospitalized & non hospitalized) 25.4% 3.4% 23.5% 13.3% 3.0% 0.0% 30.8% 100.0% Total Hospitalized Visits 17.1% 1.4% 19.1% 30.4% 2.4% 0.0% 29.1% 100.0% Total Primary Care Related* 24.1% 4.2% 26.8% 8.8% 2.2% 0.0% 33.1% 100.0% Total Injury/BH/Unclass. Visits 28.3% 3.4% 23.0% 11.5% 4.4% 0.0% 29.0% 100.0% Tot. non Primary Care Related (Total minus PCR)* 26.4% 2.8% 21.2% 16.5% 3.5% 0.0% 29.2% 100.0% Non-Emergent 23.2% 4.1% 25.7% 7.6% 2.5% 0.0% 35.8% 100.0% Emergent-Primary Care Treatable 24.9% 4.3% 27.1% 8.6% 2.0% 0.0% 32.3% 100.0% Emergent Care Needed-Preventable/Avoidable 24.3% 4.0% 28.9% 12.7% 1.9% 0.0% 27.8% 100.0% Emergent Care Needed-NOT Prev./Av. 32.7% 2.6% 18.2% 14.0% 2.2% 0.0% 29.7% 100.0% TOTAL Categorized Visits 25.8% 3.9% 25.1% 9.8% 2.2% 0.0% 32.4% 100.0% * Reflected in figures on following page. 32

Figure 6. PCR ED Visits by Payment Source (N=24) Unins, 33.1% Comm. 24.1% SCHIP 4.2% Oth. Priv. 0.0% Oth. Govt. 2.2% Medicaid 26.8% Medicare 8.8% Figure 7. All Other 1 ED Visits by Payment Source (N=24) Unins, 29.2% Comm. 26.4% Oth. Priv. 0.0% SCHIP 2.8% Oth. Govt. 3.5% Medicare 16.5% Medicaid 21.2% 1 Non PCR non-hospitalized visits and hospitalized visits. 33

Table 3c. Percentage of ED Visit Types by Payment Source (read vertically) T ype o f Visit Commerc. SCHIP Medicaid Medicare Other Gov. Other Priv. Uninsured Unk All P ercentage P rimary C are R elated 39.8% 51.9% 47.7% 27.7% 31.4% 50.4% 45.1% 54.9% 41.9% P ercentage ED C are N eeded, N o t P rev. 13.2% 8.0% 7.9% 10.8% 7.6% 10.9% 9.9% 11.2% 10.2% P ercentage Injury/ B H / Unclass. 37.6% 34.3% 32.9% 29.3% 49.6% 29.3% 31.8% 21.3% 33.8% P ercentage H o spitalized 9.4% 5.8% 11.4% 32.2% 11.4% 9.4% 13.3% 12.5% 14.1% T o tal 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% Figure 8. Payment Source by ED Visit Type (N=24) 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Percentage Hospitalized Percentage Injury/BH/Unclass. Percentage ED Care Needed, Not Prev. Percentage Primary Care Related 34

Table 4a. Frequency of ED Visit Types by Race/Ethnicity (N=24) Type of Visit Asian Black Hispanic Am. Indian Other Unk White All Total Visits (hospitalized & non hospitalized) 16,286 273,373 328,973 757 24,505 8,948 279,543 932,385 Total Hospitalized Visits 2,754 38,756 37,904 105 3,351 917 47,357 131,144 Total Primary Care Related 6,040 127,095 149,100 331 9,698 4,477 94,205 390,945 Total Injury/BH/Unclass. Visits 5,634 80,293 109,231 230 8,813 2,881 107,751 314,833 Tot. non Primary Care Related (Total minus PCR) 10,246 146,278 179,873 426 14,807 4,471 185,338 541,440 Non-Emergent 2,529 53,088 60,500 131 3,859 1,845 36,229 158,180 Emergent-Primary Care Treatable 2,767 56,008 70,509 158 4,599 2,230 44,118 180,389 Emergent Care Needed-Preventable/Avoidable 745 17,999 18,091 41 1,240 402 13,858 52,376 Emergent Care Needed-NOT Prev./Av. (dc) 1,858 27,229 32,738 91 2,643 673 30,230 95,463 TOTAL Categorized Visits 7,898 154,324 181,838 422 12,341 5,150 124,435 486,408 Table 4b. Percentage of ED Visit Types by Race/Ethnicity (read horizontally) Type of Visit Asian Black Hispanic Am. Indian Other Unk White All Total Visits (hospitalized & non hospitalized) 1.7% 29.3% 35.3% 0.1% 2.6% 1.0% 30.0% 100.0% Total Hospitalized Visits 2.1% 29.6% 28.9% 0.1% 2.6% 0.7% 36.1% 100.0% Total Primary Care Related* 1.5% 32.5% 38.1% 0.1% 2.5% 1.1% 24.1% 100.0% Total Injury/BH/Unclass. Visits 1.8% 25.5% 34.7% 0.1% 2.8% 0.9% 34.2% 100.0% Tot. non Primary Care Related (Total minus PCR)* 1.9% 27.0% 33.2% 0.1% 2.7% 0.8% 34.2% 100.0% Non-Emergent 1.6% 33.6% 38.2% 0.1% 2.4% 1.2% 22.9% 100.0% Emergent-Primary Care Treatable 1.5% 31.0% 39.1% 0.1% 2.5% 1.2% 24.5% 100.0% Emergent Care Needed-Preventable/Avoidable 1.4% 34.4% 34.5% 0.1% 2.4% 0.8% 26.5% 100.0% Emergent Care Needed-NOT Prev./Av. (dc) 1.9% 28.5% 34.3% 0.1% 2.8% 0.7% 31.7% 100.0% TOTAL Categorized Visits 1.6% 31.7% 37.4% 0.1% 2.5% 1.1% 25.6% 100.0% * Reflected in figures on following page. 35

Figure 9. Primary Care-Related ED Visits by Race/Ethnicity (N=24) White, 24.1% Asian, 1.5% Black, 32.5% Unk, 1.1% Other, 2.5% Am. Indian, 0.1% Hispanic, 38.1% Figure 10. All Other 1 ED Visits by Race/Ethnicity (N=24) Asian, 1.9% White, 34.2% Black, 27.0% Unk, 0.8% Other, 2.7% Am. Indian, 0.1% Hispanic, 33.2% 1 Non PCR non-hospitalized visits and hospitalized visits. 36

Table 4c. Percentage of ED Visit Types by Race/Ethnicity (read vertically) Type of Visit Asian Black Hispanic Am.Indian Other Unk White All Percentage Primary Care Related 37.1% 46.5% 45.3% 43.7% 39.6% 50.0% 33.7% 41.9% Percentage ED Care Needed, Not Prev. 11.4% 10.0% 10.0% 12.0% 10.8% 7.5% 10.8% 10.2% Percentage Injury/BH/Unclass. 34.6% 29.4% 33.2% 30.4% 36.0% 32.2% 38.5% 33.8% Percentage Hospitalized 16.9% 14.2% 11.5% 13.9% 13.7% 10.2% 16.9% 14.1% Total 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% Figure 11. Race/Ethnicity by ED Visit Type (N=24) 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Percentage Hospitalized Percentage Injury/BH/Unclass. Percentage ED Care Needed, Not Prev. Percentage Primary Care Related 37

Table 5a. Frequency of ED Visit Types by Age (N=24) Type of Visit 0-17 18-25 26-44 45-64 65+ All Total Visits (hospitalized & non hospitalized) 291,892 115,968 226,412 185,602 112,511 932,385 Total Hospitalized Visits 22,867 12,349 26,802 33,297 35,829 131,144 Total Primary Care Related 142,210 51,806 96,867 69,914 30,147 390,945 Total Injury/BH/Unclass. Visits 108,286 39,870 74,614 58,392 33,671 314,833 Tot. non Primary Care Related (Total minus PCR) 149,682 64,162 129,545 115,688 82,364 541,440 Non-Emergent 53,425 23,182 42,041 28,845 10,688 158,180 Emergent-Primary Care Treatable 68,024 23,156 44,345 31,302 13,561 180,389 Emergent Care Needed-Preventable/Avoidable 20,762 5,468 10,481 9,767 5,898 52,376 Emergent Care Needed-NOT Prev./Av. (dc) 18,529 11,943 28,129 23,999 12,864 95,463 TOTAL Categorized Visits 160,739 63,749 124,996 93,913 43,011 486,408 Table 5b. Percentage of ED Visit Types by Age (read horizontally) Type of Visit 0-17 18-25 26-44 45-64 65+ All Total Visits (hospitalized & non hospitalized) 31.3% 12.4% 24.3% 19.9% 12.1% 100.0% Total Hospitalized Visits 17.4% 9.4% 20.4% 25.4% 27.3% 100.0% Total Primary Care Related* 36.4% 13.3% 24.8% 17.9% 7.7% 100.0% Total Injury/BH/Unclass. Visits 34.4% 12.7% 23.7% 18.5% 10.7% 100.0% Tot. non Primary Care Related (Total minus PCR)* 27.6% 11.9% 23.9% 21.4% 15.2% 100.0% Non-Emergent 33.8% 14.7% 26.6% 18.2% 6.8% 100.0% Emergent-Primary Care Treatable 37.7% 12.8% 24.6% 17.4% 7.5% 100.0% Emergent Care Needed-Preventable/Avoidable 39.6% 10.4% 20.0% 18.6% 11.3% 100.0% Emergent Care Needed-NOT Prev./Av. (dc) 19.4% 12.5% 29.5% 25.1% 13.5% 100.0% TOTAL Categorized Visits 33.0% 13.1% 25.7% 19.3% 8.8% 100.0% * Reflected in figures on following page. 38

Figure 12. PCR ED Visits by Age (N=24) 45-64, 17.9% 65+, 7.7% 26-44, 24.8% 0-17, 36.4% 18-25, 13.3% Figure 13. All Other 1 ED Visits by Age (N=24) 65+, 15.2% 45-64, 21.4% 0-17, 27.6% 26-44, 23.9% 18-25, 11.9% 1 Non PCR non-hospitalized visits and hospitalized visits. 39

Table 5c. Percentage of ED Visit Types by Age (read vertically) Type of Visit 0-17 18-25 26-44 45-64 65+ All Percentage Primary Care Related 48.7% 44.7% 42.8% 37.7% 26.8% 41.9% Percentage ED Care Needed, Not Prev. 6.3% 10.3% 12.4% 12.9% 11.4% 10.2% Percentage Injury/BH/Unclass. 37.1% 34.4% 33.0% 31.5% 29.9% 33.8% Percentage Hospitalized 7.8% 10.6% 11.8% 17.9% 31.8% 14.1% Total 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% Figure 14. Age by ED Visit Type (N=24) 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 0-17 18-25 26-44 45-64 65+ All Percentage Hospitalized Percentage Injury/BH/Unclass. Percentage ED Care Needed, Not Prev. Percentage Primary Care Related 40

Table 6a. Frequency of ED Visit Types by Gender (N=24) Type of Visit Female Male Unk All Total Visits (hospitalized & non hospitalized) 514,564 417,734 87 932,385 Total Hospitalized Visits 71,129 60,000 15 131,144 Total Primary Care Related 228,716 162,202 27 390,945 Total Injury/BH/Unclass. Visits 159,138 155,657 38 314,833 Tot. non Primary Care Related (Total minus PCR) 285,848 255,532 60 541,440 Non-Emergent 95,755 62,414 11 158,180 Emergent-Primary Care Treatable 104,381 75,996 12 180,389 Emergent Care Needed-Preventable/Avoidable 28,580 23,792 4 52,376 Emergent Care Needed-NOT Prev./Av. (dc) 55,581 39,875 7 95,463 TOTAL Categorized Visits 284,297 202,077 34 486,408 Table 6b. Percentage of ED Visit Types by Gender (read horizontally) Type of Visit Female Male Unk All Total Visits (hospitalized & non hospitalized) 55.2% 44.8% 0.0% 100.0% Total Hospitalized Visits 54.2% 45.8% 0.0% 100.0% Total Primary Care Related* 58.5% 41.5% 0.0% 100.0% Total Injury/BH/Unclass. Visits 50.5% 49.4% 0.0% 100.0% Tot. non Primary Care Related (Total minus PCR)* 52.8% 47.2% 0.0% 100.0% Non-Emergent 60.5% 39.5% 0.0% 100.0% Emergent-Primary Care Treatable 57.9% 42.1% 0.0% 100.0% Emergent Care Needed-Preventable/Avoidable 54.6% 45.4% 0.0% 100.0% Emergent Care Needed-NOT Prev./Av. (dc) 58.2% 41.8% 0.0% 100.0% TOTAL Categorized Visits 58.4% 41.5% 0.0% 100.0% * Reflected in figures on following page. 41

Figure 15. Primary Care Related ED Visits by Gender (N=24) Male, 41.5% Female, 58.5% Figure 16. All Other 1 ED Visits by Gender (N=24) Male, 47.2% Female, 52.8% 1 Non PCR non-hospitalized visits and hospitalized visits. 42

Table 6c. Percentage of ED Visit Types by Gender (read vertically) Type of Visit Female Male Unk All Percentage Primary Care Related 44.4% 38.8% 31.1% 41.9% Percentage ED Care Needed, Not Prev. 10.8% 9.5% 8.0% 10.2% Percentage Injury/BH/Unclass. 30.9% 37.3% 43.7% 33.8% Percentage Hospitalized 13.8% 14.4% 17.2% 14.1% Total 100.0% 100.0% 100.0% 100.0% Figure 17. Gender by ED Visit Type (N=24) 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Female Male Unk All Percentage Hospitalized Percentage Injury/BH/Unclass. Percentage ED Care Needed, Not Prev. Percentage Primary Care Related 43

ED Visits by Method of Transporation to the ED Visits were classified according to the following groupings: ambulance (any form of ground or air ambulance), self (automobile, walk in, public transporation), other (e.g. police, newborns), and unknown (field was empty or missing). Even though the hospitals invited to take part in the study are primarily the ones serving the Houston 911 service area, non HFD ambulance transports are reflected in these results as well. One in ten (10.1%) PCRED visits were by persons with an ambulance transport to the ED. By contrast, more than one in five (21.8%) non-pcred visits were by persons with an ambulance transport to the ED (Tables 7a & 7b; Figures 18 & 19). One fourth (25.0%) of ambulance transports to the ED eventually resulted in a hospitalization (Table 7c and Figure 20). 44

Table 7a. Frequency of ED Visit Types by Transport (N=24) Type of Visit Ambulance Self Other Unk All Total Visits (hospitalized & non hospitalized) 157,466 749,225 5,661 20,033 932,385 Total Hospitalized Visits 42,779 82,759 1,179 4,427 131,144 Total Primary Care Related 39,428 342,385 1,682 7,449 390,945 Total Injury/BH/Unclass. Visits 58,825 247,556 2,382 6,070 314,833 Tot. non Primary Care Related (Total minus PCR) 118,038 406,840 3,979 12,584 541,440 Non-Emergent 13,633 140,483 816 3,248 158,180 Emergent-Primary Care Treatable 17,300 159,134 667 3,288 180,389 Emergent Care Needed-Preventable/Avoidable 8,494 42,769 199 913 52,376 Emergent Care Needed-NOT Prev./Av. (dc) 16,434 76,525 418 2,087 95,463 TOTAL Categorized Visits 55,862 418,910 2,100 9,536 486,408 Table 7b. Percentage of ED Visit Types by Transport (read horizontally) Type of Visit Ambulance Self Other Unk All Total Visits (hospitalized & non hospitalized) 16.9% 80.4% 0.6% 2.1% 100.0% Total Hospitalized Visits 32.6% 63.1% 0.9% 3.4% 100.0% Total Primary Care Related* 10.1% 87.6% 0.4% 1.9% 100.0% Total Injury/BH/Unclass. Visits 18.7% 78.6% 0.8% 1.9% 100.0% Tot. non Primary Care Related (Total minus PCR)* 21.8% 75.1% 0.7% 2.3% 100.0% Non-Emergent 8.6% 88.8% 0.5% 2.1% 100.0% Emergent-Primary Care Treatable 9.6% 88.2% 0.4% 1.8% 100.0% Emergent Care Needed-Preventable/Avoidable 16.2% 81.7% 0.4% 1.7% 100.0% Emergent Care Needed-NOT Prev./Av. (dc) 17.2% 80.2% 0.4% 2.2% 100.0% TOTAL Categorized Visits 11.5% 86.1% 0.4% 2.0% 100.0% * Reflected in figures on following page. 45

Figure 18. Primary Care Related ED Visits by Transport (N=24) Other, 0.4% Unk, 1.9% Ambulance, 10.1% Self, 87.6% Figure 19. All Other 1 ED Visits by Transport (N=24) Other, 0.7% Unk, 2.3% Ambulance, 21.8% Self, 75.1% 1 Non PCR non-hospitalized visits and hospitalized visits. 46

Table 7c. Percentage of ED Visit Types by Transport (read vertically) Type of Visit Ambulance Self Other Unk All Percentage Primary Care Related 25.0% 45.7% 29.7% 37.2% 41.9% Percentage ED Care Needed, Not Prev. 10.4% 10.2% 7.4% 10.4% 10.2% Percentage Injury/BH/Unclass. 37.4% 33.0% 42.1% 30.3% 33.8% Percentage Hospitalized 27.2% 11.0% 20.8% 22.1% 14.1% Total 100.0% 100.0% 100.0% 100.0% 100.0% Figure 20. Transport by ED Visit Type (N = 24) 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Percentage Hospitalized Percentage Injury/BH/Unclass. Percentage ED Care Needed, Not Prev. Percentage Primary Care Related 47

E. Health Conditions of Patients with ED Visits Categorized (primary and non primary care related) ED visits The most common discharge diagnoses of patients with PCRED visits for the last three years has been acute upper respiratory infection not otherwise specified. The frequency of visits with this diagnosis was higher in 2009 than in previous years (Table 8). Chest pain not otherwise specified was the leading diagnosis of non-pcred visits, and also increased in frequency in 2009 (Table 9). F. Behavioral Health Related ED Visits The percentage of behavioral health visits dropped from 8.3% in 2008 to 7.7% in 2009. The number of patients who were seen with an alcohol-related diagnosis and no other behavioral diagnosis nearly doubled in 2009 (1,079) compared to 2008 (585). The number of patients whose primary diagnosis was medical, but who had a secondary diagnosis of an alcohol problem, rose by about 50% (from 2,417 in 2008 to 3,689 in 2009) (Table 10a). 48

Table 8. Most Frequent Conditions of Patients with PCR ED Visits 2007 data 2008 visits 2009 visits 25 hospitals 25 hospitals 24 hospitals ICD COUNT Description icd COUNT Description ICD COUNT Description 4659 19,755 acute uri nos 4659 15,118 acute uri nos 4659 26,401 acute uri nos 3829 15,582 otitis media nos 3829 11,934 otitis media nos 5990 14,559 urin tract infection nos 7806 13,950 fever 7806 11,790 fever 3829 12,517 otitis media nos 5589 13,834 noninf gastroenterit nec 462 10,652 acute pharyngitis 78900 12,450 abdmnal pain unspcf site 5990 13,023 urin tract infection nos 5589 9,249 noninf gastroenterit nec 7999 11,195 viral infection nos 462 11,122 acute pharyngitis 78900 8,341 abdmnal pain unspcf site 7840 11,115 headache 486 9,897 pneumonia, organism nos 486 7,938 pneumonia, organism nos 462 9,868 acute pharyngitis 7840 8,933 headache 49392 7,674 asthma nos w(ac) exacerb 78703 8,462 vomiting alone 78900 7,559 abdmnal pain unspcf site 7295 7,109 pain in limb 486 7,767 pneumonia, organism nos 49392 7,526 asthma nos w(ac) exacerb 78703 7,063 vomiting alone 5589 7,490 noninf gastroenterit nec 78703 6,880 vomiting alone 78659 6,816 chest pain nec 64893 7,238 oth curr cond-antepartum 6826 6,718 cellulitis of leg 6826 6,815 cellulitis of leg 490 6,672 bronchitis nos 4019 6,420 hypertension nos 7242 6,436 lumbago 49392 6,617 asthma nos w(ac) exacerb 4660 6,274 acute bronchitis 7804 6,398 dizziness and giddiness 78659 5,982 chest pain nec 78659 5,962 chest pain nec 64893 6,317 oth curr cond-antepartum 7242 5,852 lumbago 78039 5,687 convulsions nec 4019 6,091 hypertension nos 6826 5,673 cellulitis of leg 490 5,571 bronchitis nos 7840 5,323 headache 4871 5,587 flu w resp manifest nec 64893 5,282 oth curr cond-antepartum 490 5,219 bronchitis nos 7295 5,572 pain in limb 7242 5,117 lumbago 78909 4,951 abdmnal pain oth spcf st 4660 5,533 acute bronchitis 7804 4,546 dizziness and giddiness 5990 4,746 urin tract infection nos 4019 5,402 hypertension nos 7999 4,502 viral infection nos 78039 4,565 convulsions nec 7804 5,374 dizziness and giddiness 7295 4,257 pain in limb 78652 4,323 painful respiration 78909 5,065 abdmnal pain oth spcf st 78652 3,829 painful respiration 78701 4,300 nausea with vomiting 7862 4,490 cough 64003 3,709 threaten abort-antepart 64003 4,255 threaten abort-antepart 78650 4,421 chest pain nos 6823 3,471 cellulitis of arm 7862 4,067 cough 64003 4,172 threaten abort-antepart 49

Table 9. Most frequent Conditions of Patients with Non PCR ED Visits 2007 data 2008 visits 2009 visits 25 hospitals 25 hospitals 24 hospitals ICD COUNT Description icd COUNT Description ICD COUNT Description 78650 6,623 chest pain nos 78650 4,635 chest pain nos 78650 9,225 chest pain nos 78659 3,794 chest pain nec 5921 4,607 calculus of ureter 78900 6,140 abdmnal pain unspcf site 78900 3,727 abdmnal pain unspcf site 7802 4,382 syncope and collapse 5921 4,024 calculus of ureter 7802 3,572 syncope and collapse 78659 4,338 chest pain nec 78659 3,807 chest pain nec 7806 3,402 fever 78900 4,113 abdmnal pain unspcf site 7802 3,719 syncope and collapse 5921 3,384 calculus of ureter 7806 2,876 fever 78909 2,498 abdmnal pain oth spcf st 64003 1,854 threaten abort-antepart 78909 2,442 abdmnal pain oth spcf st 64003 2,086 threaten abort-antepart 5920 1,789 calculus of kidney 7999 2,270 unkn cause morb/mort nec 4871 1,862 flu w resp manifest nec 5409 1,613 acute appendicitis nos 78605 2,242 shortness of breath 78703 1,813 vomiting alone 78909 1,600 abdmnal pain oth spcf st 64003 2,127 threaten abort-antepart 7840 1,659 headache 78703 1,474 vomiting alone 42731 1,911 atrial fibrillation 57420 1,531 cholelithiasis nos 7851 1,392 palpitations 5409 1,908 acute appendicitis nos 46619 1,504 acu brnchlts d/t oth org 57420 1,364 cholelithiasis nos 56211 1,684 dvrtcli colon w/o hmrhg 42731 1,495 atrial fibrillation 46619 1,356 acu brnchlts d/t oth org 7851 1,680 palpitations 7851 1,459 palpitations 7840 1,333 headache 46619 1,648 acu brnchlts d/t oth org 5920 1,452 calculus of kidney 4644 1,265 croup 4359 1,572 trans cereb ischemia nos 56211 1,331 dvrtcli colon w/o hmrhg 4359 1,244 trans cereb ischemia nos 78703 1,513 vomiting alone 4644 1,294 croup 42731 1,229 atrial fibrillation 41071 1,314 subendo infarct, initial 78605 1,265 shortness of breath 5770 1,225 acute pancreatitis 5789 1,298 gastrointest hemorr nos 5770 1,262 acute pancreatitis 56211 1,132 dvrtcli colon w/o hmrhg 43491 1,292 crbl art ocl nos w infrc 7999 1,256 viral infection nos V715 1,006 observ following rape 28262 1,261 hb-s disease with crisis 53550 1,226 gstr/ddnts nos w/o hmrhg 5789 991 gastrointest hemorr nos 7243 1,218 sciatica 5409 1,219 acute appendicitis nos 53550 974 gstr/ddnts nos w/o hmrhg V715 1,207 observ following rape V715 1,183 observ following rape 28262 919 hb-s disease with crisis 78609 1,142 respiratory abnorm nec 5789 1,159 gastrointest hemorr nos 7243 889 sciatica 7295 1,016 pain in limb 7243 1,105 sciatica 50

Table 10. Total Behavioral Health 1 Visits To Harris County Area Hospitals Frequency - Primary diagnosis and any other diagnosis 2008 (25) 2009 (24) Total Visits 2 899,673 953,395 Total Visits With Any Behavioral Health Diagnosis 74,672 73,383 Total Visits With No Behavioral Health Diagnosis 825,001 880,012 Total Visits With a Behavioral Health Diagnosis as the primary diagnosis 3 9,021 10,311 Total Visits With a Behavioral Health Diagnosis as both primary and any other diagnosis 3 9,134 8,211 Total Visits With a Behavioral Health Diagnosis as any other diagnosis 3,4 56,287 54,861 Percentage of First Row Total Visits 100.0% 100.0% Total Visits With Any Behavioral Health Diagnosis 8.3% 7.7% Total Visits With No Behavioral Health Diagnosis 91.7% 92.3% Total Visits With a Behavioral Health Diagnosis as the primary diagnosis 1.0% 1.1% Total Visits With a Behavioral Health Diagnosis as both primary and any other diagnosis 3 1.0% 0.9% Total Visits With a Behavioral Health Diagnosis as any other diagnosis 3,4 6.3% 5.8% Frequency - Primary and secondary diagnoses Total Visits With a Mental Health Diagnosis as the primary diag. & no secondary behavioral diag. 5 2,674 2,761 Total Visits With an Alcohol Related Diagnosis as the primary diag. & no secondary behavioral diag. 5 585 1,079 Total Visits With a Drug Related Diagnosis as the primary diag. & no secondary behavioral diag. 5 226 137 Total Visits With a medical diagnosis as the primary diag. & Mental Health as the secondary diag. 5 11,028 11,155 Total Visits With a medical diagnosis as the primary diag. & Alcohol Related as the secondary diag. 5 2,417 3,689 Total Visits With a medical diagnosis as the primary diag. & Drug Related as the secondary diag. 5 1,820 1,056 1 Behavioral Health = mental health, alcohol, or drug related. 2 Harris County, all diagnoses, all discharge dispositions. 3 Primary diagnosis = first ICD code. Any other diagnosis = ICD codes 2 through 10. 4 A medical diagnosis is the primary diagnosis. 5 Unlike above, secondary diagnosis in this case refers only to the second ICD code. Person may or may not have a secondary medical diagnosis. 51

G. Length of Stay in the ED The mean length of stay of ED visits (measured from time of admission to the ED to time of discharge from the ED) was calculated for visits of less than 24 hours each. Primary care related visits continue to last roughly half an hour less than all non-hospitalized visits (3.70 hours in 2009 vs. 4.34 hours in 2009). The length of stay for a hospitalized ED visit increased from 5.37 hours in 2007 to 6.84 hours in 2009 (Table 12). Non hospitalized behavioral health visits continue to last longer than non hospitalized medical visits (Table 11). H. Estimated Cost of Primary Care Related ED Visits Applying national estimates of the unit cost of ED visits to Houston ED visits suggests that the cost of PCRED visits in Houston hospitals increased from $187,310,573 in 2007 to $214,859,768 in 2009 (Table 13). The greatest potential cost savings from diverting PCRED visits would be among uninsured adults age 18 to 44 (Table 12 and 13, Figures 21and 22). This is based on the fact that PCRED visits of adults age 18 to 44 represented 36.8% of total PCRED visits the estimated savings among all persons in 2009 and uninsured adults accounted for 54.2% of the estimated cost of PCRED visits savings among the uninsured in 2009 (Tables 12 and 13). 1 1 Machlin, S.R. Trends in health care expenditures: 2006 versus 1996. Statistical Briefs #253-256. August 2009. Agency for Healthcare Research and Quality, Rockville, MD. Available at http://www.meps.ahrq.gov/mepsweb/data_stats/meps_topics.jsp?topicid=5z-1. 2006 estimates were used for this report. These national estimates are based on all payer sources (e.g. insured and uninsured). They may not reflect actual hospital charges or procedures administered during treatment. Outpatient estimate may or may not cover both primary care and specialty care. MEPS estimates were adjusted for inflation to reflect 2007 to 2009 expenses. 52

Table 11. Hospitals in the Houston 911 Service Area. Mean length of stay (in hours) in the emergency department for visits of less than 24 hours. Residence Discharge status 2007 Mean (sd) 25 hospitals 2008 Mean (sd) 25 hospitals 2009 Mean (sd) 24 hospitals Harris County All Visits 1 4.64 (4.18) 4.55 (4.20) 4.54 (4.27) Harris County Primary care 3.96 (3.57) 3.86 (3.65) 3.70 (3.49) related 1 4.57 (4.08) 4.44 (4.10) 4.34 (4.04) Harris County All nonhospitalized 1 Harris County Other 1,2 5.37 (5.07) 5.91 (5.11) 6.84 (5.83) Harris County Harris County Non-hospitalized medical 3 4.50 (3.95) 4.48 (4.09) 4.28 (3.94) Non-hospitalized 6.05 (4.75) 5.52 (4.64) 6.01 (5.02) behavioral 3 1 All diagnoses 2 Admitted, transferred to another facility, etc. 3 Medical or behavioral as the primary discharge diagnosis. 53

Table 12. Estimated costs of primary care related emergency room visits, estimated costs of treating those same patients in an outpatient setting all payer sources 1 Age Group Prim. Care Visits Estimated national ED cost per visit 2007 (24 hospitals) Est. ED cost Estimated national outpt. cost per visit Est. Outpt cost Difference 17 and under 138,639 $529 $73,367,547 $140 $19,464,859 $53,902,688 18 to 44 137,623 $689 $94,828,021 $194 $26,753,987 $68,074,034 45 to 64 63,850 $1,023 $65,302,817 $224 $14,274,217 $51,028,600 65 and over 28,122 $703 $19,772,227 $194 $5,466,975 $14,305,252 Total Primary Care Visits 368,234 na 2 $253,270,611 na 2 $65,960,038 $187,310,573 2008 (24 hospitals) 17 and under 118,936 $541 $64,397,681 $144 $17,085,099 $47,312,582 18 to 44 120,347 $705 $84,843,269 $199 $23,936,973 $60,906,297 45 to 64 58,343 $1,046 $61,052,356 $229 $13,345,130 $47,707,226 65 and over 25,411 $719 $18,279,868 $199 $5,054,341 $13,225,527 Total Primary Care Visits 323,037 na 2 $228,573,174 na 2 $59,421,542 $169,151,632 2009 (24 hospitals) 17 and under 142,210 $568 $80,832,363 $151 $21,445,321 $59,387,042 18 to 44 148,674 $740 $110,030,654 $209 $31,043,131 $78,987,523 45 to 64 69,914 $1,099 $76,801,513 $240 $16,787,659 $60,013,854 65 and over 30,147 $755 $22,766,132 $209 $6,294,783 $16,471,349 Total Primary Care Visits 390,945 na 2 $290,430,662 na 2 $75,570,894 $214,859,768 1 1 Machlin, S.R. Trends in health care expenditures: 2006 versus 1996. Statistical Briefs #253-256. August 2009. Agency for Healthcare Research and Quality, Rockville, MD. Available at http://www.meps.ahrq.gov/mepsweb/data_stats/meps_topics.jsp?topicid=5z-1. 2006 estimates were used for this report. These national estimates are based on all payer sources (e.g. insured and uninsured). They may not reflect actual hospital charges or procedures administered during treatment. Outpatient estimate may or may not cover both primary care and specialty care. MEPS estimates were adjusted for inflation to reflect 2007 to 2009 expenses. 2 MEPS estimates were only created for specific age groups. 54

Figure 21. Estimated savings for treating all PCRED visits in an outpatient setting all payer sources (final column of previous table) $90,000,000 $80,000,000 $70,000,000 $60,000,000 $50,000,000 $40,000,000 $30,000,000 $20,000,000 $10,000,000 $0 2007 2008 2009 17 and under 18 to 44 45 to 64 65 and over 55

Table 13. Estimated costs of primary care related emergency room visits, estimated costs of treating those same patients in an outpatient setting uninsured only 1 Estimated national ED cost per visit 2007 (24 hospitals) Estimated national outpt. cost per visit Est. Outpt cost Difference Age Group Prim. Care Visits Est. ED cost 17 and under 24,289 $529 $12,853,797 $140 $3,410,191 $9,443,606 18 to 44 70,363 $689 $48,482,984 $194 $13,678,585 $34,804,399 45 to 64 23,761 $1,023 $24,301,350 $224 $5,311,911 $18,989,440 65 and over 1,437 $703 $1,010,235 $194 $279,328 $730,907 Total Primary Care Visits 119,850 na 2 $86,648,365 na 2 $22,680,014 $63,968,351 2008 (24 hospitals) 17 and under 19,168 $541 $10,378,779 $144 $2,753,554 $7,625,225 18 to 44 60,691 $705 $42,786,562 $199 $12,071,444 $30,715,118 45 to 64 22,296 $1,046 $23,331,535 $229 $5,099,924 $18,231,611 65 and over 1,634 $719 $1,175,146 $199 $324,925 $850,220 Total Primary Care Visits 103,789 na 2 $77,672,021 na 2 $20,249,846 $57,422,175 2009 (24 hospitals) 17 and under 24,221 $568 $13,767,256 $151 $3,652,537 $10,114,719 18 to 44 76,425 $740 $56,560,747 $209 $15,957,578 $40,603,170 45 to 64 27,009 $1,099 $29,669,421 $240 $6,485,291 $23,184,131 65 and over 1,779 $755 $1,343,279 $209 $371,413 $971,865 Total Primary Care Visits 129,434 na 2 $101,340,703 na 2 $26,466,819 $74,873,884 1 1 Machlin, S.R. Trends in health care expenditures: 2006 versus 1996. Statistical Briefs #253-256. August 2009. Agency for Healthcare Research and Quality, Rockville, MD. Available at http://www.meps.ahrq.gov/mepsweb/data_stats/meps_topics.jsp?topicid=5z-1. 2006 estimates were used for this report. These national estimates are based on all payer sources (e.g. insured and uninsured). They may not reflect actual hospital charges or procedures administered during treatment. Outpatient estimate may or may not cover both primary care and specialty care. MEPS estimates were adjusted for inflation to reflect 2008 expenses. 2 MEPS estimates were only created for specific age groups. 56

Figure 22. Estimated savings for treating all PCRED visits in an outpatient setting uninsured (final column of previous table) $45,000,000 $40,000,000 $35,000,000 $30,000,000 $25,000,000 $20,000,000 $15,000,000 $10,000,000 $5,000,000 $0 2007 2008 2009 17 and under 18 to 44 45 to 64 65 and over 57

I. Geographic Distribution of Patients with ED Visits The following ZIP code clusters had the highest number of ED visits: 1) the intersection of Highway 59 and the North Loop East; 2) the Ship Channel and the area along Interstate 10 and the East Belt; 3) an area just east of Highway 288 and on both sides of the South Loop; 4) an area along the Southwest Freeway and between the West Loop and West Belt; 5) the intersection of Interstate 10 and Highway 6; and 6) an area along Highway 290 and just outside the North Loop West (Map 1a). Clusters 1, 4, and 5 had the biggest increases from 2007 to 2009 (Map 1b). The geographic distribution of PCRED visits by the insured and uninsured was similar to that of ED visits (Maps 2a and 3a). An area between the North Loop and the North Belt had the biggest increase in PCRED visits by the insured and uninsured (area shaded in orange) compared to the increase by the uninsured (area shaded in yellow). (Maps 2b and 3b). The areas with the highest rates of ED visits, PCRED visits by all, and PCRED visits by the uninsured were the East/Northeast and South Central areas of Harris County (Map 4). The population rates of PCRED visits for all and the uninsured, were lowest in the Baytown area (Maps 4, 5, and 6). 58

Map 1a 59

Map 1b 60