HOUSTON HOSPITALS EMERGENCY DEPARTMENT USE STUDY. January 1, 2007 through December 31, 2007 FINAL REPORT. Prepared By

Transcription

1 HOUSTON HOSPITALS EMERGENCY DEPARTMENT USE STUDY January 1, 2007 through December 31, 2007 FINAL REPORT Prepared By School of Public Health University of Texas Health Science Center at Houston Charles Begley, Patrick Courtney, Keith Burau April 2009

2 TABLE OF CONTENTS I. Executive Summary...3 II. Purpose...14 III. Methods...15 IV. Results...19 ED Visits in general Summary Tables and Figures ED Visits by Month, Day of Week, and Time of Day..23 Summary Tables and Figures ED Visits by Patient Characteristics.25 Summary Tables and Figures Health Conditions of Patients with ED Visits 44 Summary Tables and Figures Length of Stay of ED Visits Table Geographic Distribution of Patients with ED Visits Summary Maps Trends in ED Visits Over Time...55 Summary Tables and Figures V. Limitations

3 I. EXECUTIVE SUMMARY The University Of Texas School Of Public Health has been collecting and analyzing emergency department (ED) visit data in Harris County hospitals to monitor primary care-related use of the ED. Beginning with this year s report, non primary care related ED visits are examined as well. This report provides an analysis of 2007 ED visit data and describes trends over the last few years. Data and Analysis Twenty five hospitals which have EDs and provide a substantial amount of discounted and free care to the uninsured of Harris County have provided ED visit data for the year 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 (HCA): Bayshore Medical Center, Spring Branch Medical Center, East Houston Regional Medical Center, and West Houston Medical Center); St. Joseph Medical Center; Texas Children s Hospital; two hospitals of CHRISTUS Gulf Coast (St. Catherine and St. John); two hospitals of St. Luke s (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). 3

4 Each hospital supplied the following data elements on all ED visits made during the period January 1, 2007 through December 31, 2007: 1. Date and time of admission to ED 2. Primary and secondary discharge diagnosis 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 to (e.g. home, nursing home, etc.) To determine primary care-related ED visits, the probability that each non-hospitalized visit was one or more of the following types was assigned by applying the New York University ED Algorithm to the discharge diagnosis. 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 number of visits with each diagnosis was multiplied by the respective probabilities of visit type and then the number of visit types was aggregated across diagnoses to produce estimates of the total number of non-ed visits by type. ED visits in the first three categories are considered primary care-related use of the ED. Those in the fourth category reflect non-primary care-related use of the ED. 4

5 The demographic, coverage, and health conditions of patients with primary care and nonprimary care-related visits are also shown. ED visits were mapped by ZIP code according to both frequency and rate. Rates were based on 2007 population estimates provided by Environmental Systems Research Institute, Inc. ZIP codes with populations below 2,000 were excluded. The time and geographic pattern of primary care and non-primary care-related ED visits are summarized in the report. The Executive Summary and main body of the report present 2007 information for 25 hospitals. The comparison of ED visits over the years is for two groups of hospitals: a) 16 which have continuously been part of the study since 2004, and b) 23 which have continuously been part of the study since Results 1. Total ED Visits by Type During 2007, 1,060,825 total ED visits were made to all 25 participating hospitals. Of this total, 984,994, or 92.9%, visits were made by residents of the eight county area (surrounding and including Harris County), and 901,684 visits (85.0%) were made by Harris County residents. Non-hospitalized ED visits by residents of the eight county area numbered 835,251 (78.7% of total), and non-hospitalized ED visits by Harris County residents numbered 761,077. This figure of 761,077 represents 71.7% of total visits to participating hospitals (Table 1). Total ED visits to participating hospitals represent 71.0% of all ED visits to hospitals located in the Harris County area. When extrapolated to all hospitals and divided by the population, these visits translate to a total ED visit rate for Harris County residents of 32.6 per 100 residents (Table 1). 5

6 2. Primary Care-Related ED Visits Almost half (48.7%) of all non-hospitalized visits by Harris County residents were primary care-related. That figure represents 41.1% of total visits (Table 2 and Figure 2). 3. ED Visits by Month, Day, and Time Total ED visits by Harris County residents fluctuated during the year, peaking in March at roughly 78,000, with June having the fewest visits. While primary carerelated ED visits tended to remain at a low and stable level during the summer months of June through August, non primary care related ED visits during the same period rose. The number of hospitalized ED visits remained relatively level each month (Figure 3). The peak day of the week for ED visits was Tuesday (Figure 4). The number of hospitalized ED visits remained relatively level each day. ED visits by children peaked between 7-8 PM, with a secondary peak at noon (Figure 5a). ED visits by adults age 18 to 64 peaked at noon, dropped slightly and remained fairly level until 8 PM (Figure 5b). ED visits by adults age 65 and over peaked from 11 AM to 4 PM, then dropped during the evening (Figure 5c). Very few ED visits by children resulted in a hospitalization, whereas as many as half of ED visits by adults age 65 and over resulted in a hospitalization. 4. ED Visits by Patient Characteristics Payer source: Persons who were uninsured made up 32.8% of primary care related ED visits and 30.6% of all other ED visits (non-primary care-related non-hospitalized visits and hospitalized visits). Persons with Medicaid made up 27.4% of primary care 6

7 related ED visits but only 18.6% of all other ED visits. Children on SCHIP made up 3.1% of primary care related ED visits (Figures 6 & 7). Persons with commercial insurance accounted for 25.4% of primary care-related ED visits and 27.4% of all other ED visits. More than half of all ED visits by persons on Medicaid (52.7%) or SCHIP (53.5%) were primary care-related, and 35.6% of ED visits by persons on Medicare resulted in a hospitalization (Figure 8). Race: More than two thirds (67.5%) of primary care related ED visits were by persons who were Black or Hispanic, and they accounted for 57.5% of all other ED visits. By contrast, persons who were White made up 25.9% of primary care related ED visits and 35.4% of all other ED visits (Figures 9 & 10). Just under half of all ED visits by persons who were Black or Hispanic were primary care-related. Persons who were Asian or White were the ones most likely to have an ED visit resulting in a hospitalization (Figure 11). Age: Children age 17 and under made up more than one third (37.1%) of all primary care related ED visits, but only one fourth (25.4%) of all other ED visits. Adults age 65 and over were twice as likely to have some other type of ED visit as they were to have a primary care related ED visit. Adults age were almost as likely to have a primary care related ED visit as some other type of ED visit (Figures 12 & 13). More than half of all ED visits by children were primary care related, and roughly one third of ED visits by adults age 65 and over resulted in a hospitalization (Figure 14). Gender: Almost six in ten (57.8%) primary care-related visits were by females and 42.2% were by males (Figure 15). For all other ED visits, 52.5% were by females and 47.5% were by males (Figure 16). Close to half (45.4%) of all ED visits by females 7

8 were primary care-related, whereas percentages of ED visits resulting in a hospitalization were fairly equal between the genders (14.8% of visits by females and 14.5% of visits by males) (Figure 17). 5. Health Conditions of Patients with Primary and Non-Primary Care-Related Visits The top eight discharge diagnoses of patients with a primary care-related diagnosis in 2007 were the same top eight (in the same order) as in 2006 (Table 7). In 2007 these top eight diagnoses totaled 106,095 visits, or 28.6% of the 370,695 total primary care related visits in Harris County (cf. Table 2). Chest pain not otherwise specified has remained the top non-primary care-related diagnosis for the last two years (Table 8). 6. 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 lasted roughly half an hour less than all non-hospitalized visits (3.96 hours vs hours), whereas hospitalized visits lasted nearly five and a half hours (5.37 hours) (Table 9). 7. Geographic Distribution of Patients with ED Visits The following ZIP code clusters (in clockwise order) had the highest number of total ED visits (non-hospitalized and hospitalized), primary care-related ED visits, and primary care- related ED visits by the uninsured: 1) Aldine/Fifth Ward/northeast section of the 610 loop (both inside and outside the loop); 2) the Baytown/Ship Channel/far northeastern Harris County area; 3) Galena Park and parts of Pasadena; 4) Third Ward and South Park/central and southern part of the 610 loop; 5) 8

9 southwestern (Alief/Sharpstown) area; 6) Bear Creek/intersection of interstate 10 and highway 6; and 7) Acres Homes/Spring Branch (Maps 1, 2, and 3). Maps 2 and 3 in particular show geographic patterns similar to recent years. When examined in terms of population rates, the areas with the highest rates of overall ED visits, primary care related ED visits in general, and primary care related ED visits by the uninsured are the east/northeast and south central areas of Harris County (Maps 4, 5, and 6). Map 4 also shows an elevated rate of total ED visits in the Katy area, in contrast to primary care-related visits (cf. Maps 5 & 6). Trends in ED Visits Over Time Eleven hospitals have been part of the annual ED study since its inception in However, these hospitals now represent less than half of all hospitals participating in the study, and roughly one third of all hospitals with a general ED serving Harris County. Therefore, beginning this year the ED visits of 16 hospitals are tracked from 2004 to 2007, and the ED visits of 23 hospitals are tracked from 2005 to Due to changes in reporting, only non- hospitalized ED visits are tracked : 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, St. Joseph Medical Center (formerly CHRISTUS St. Joseph), Texas Children s Hospital, St. Luke s Episcopal Hospital : Above listed hospitals, plus MH Northeast (formerly Northeast Medical Center), HCA West Houston, CHRISTUS St. Catherine, CHRISTUS St. John, St. Luke s Community Medical Center, The Methodist Hospital, and San Jacinto Methodist. In both sets of analyses, there was a drop from 2006 to 2007 in the percentage of non-hospitalized visits that were primary care related. For the 16 hospitals, the percentage dropped from 54.1% in 2006 to 50.7% in 2007 (Table 10). In the 23 hospitals, the percentage 9

10 dropped from 52.3% in 2006 to 50.2% in 2007 (Table 11). There was little change in the percentage from 2004 to 2005, or from 2005 to Payer mix differences continue to figure predominantly in the number of primary carerelated visits (Figures 18 & 19). In the 16 hospitals, the percentage of primary care-related visits by the uninsured has declined since 2004 (Figure 18). In the 23 hospitals, persons with Medicaid and those with commercial insurance contributed between 25% and 30% of primary care-related ED visits. In both the 16 hospitals and the 23 hospitals, visits by persons with Medicare represent a small but steadily increasing percentage of primary care-related ED visits. Hispanics represent the largest proportion of persons with primary care related visits, and their proportion has remained relatively unchanged at 35% to 40% (Figures 20 & 21). Blacks represent the next largest group. Visits by children age 18 and under continue to account for between 35% and 40% of primary care-related ED visits (Figures 22 and 23). The percentage of primary care-related visits by those age 65 and over has increased in both groups of hospitals. The percentage of primary care-related ED visits by adults age 26 to 44 has decreased while the percentage of visits by adults age 45 to 64 has increased. There has been no change in the gender ratio (Figures 24 & 25). 10

11 Conclusion Whether tracked through 16 hospitals or through 23 hospitals, the percentage of non-hospitalized ED visits that are primary care related has dropped. The uninsured continue to figure predominantly in the payer mix, yet a large number of primary care related ED visits are by persons with Medicaid or commercial/private insurance. The elderly and near elderly (age 45-64) make up an increasing percentage of primary care related ED visits. The geographic distribution of ED visits shows little variation, whether measured as total visits or as primary care related visits. The pattern of primary care related visits has changed little since the 2005 report. The distribution by population continues to distinguish the east/northeast and south central sections of Harris County as having high rates of primary care related ED visits by the uninsured. Hospitalized visits, when measured by month and day, seem to be fairly stable; the variability in volume seems to be in non-hospitalized visits. Persons on Medicaid show up disproportionately at the ED with primary care related visits compared to other ED visits. Likewise, persons who are Hispanic or Black are more likely to have a primary care related ED visit than some other type of ED visit. While childhood ED visits rarely resulted in a hospitalization, one third of all ED visits by the elderly resulted in a hospitalization. Primary care related visits wherein a patient was non-hospitalized actually lasted less time than other visits (possibly related to a lower level of acuity or severity). The peak hour of visits for children is different than the peak hours for adults and older adults. Continued participation by a large number of hospitals has allowed for a fuller picture of the nature and distribution of ED utilization in Harris County. Possible applications of this report include (but are not limited to): a) how and where to expand outpatient capacity, especially for 11

12 the uninsured, b) patient education about appropriate ED use, c) after hours clinics for children, d) hospital staffing at peak hours of ED use by children and the elderly, and e) intensive efforts to link Medicaid and SCHIP enrollees to a medical home. Focusing solely on primary care related ED visits by the uninsured ignores the fact that many primary care related ED visits are by persons with private insurance or Medicaid, and that a large percentage of ED visits are not primary care related. ED volume and ED overcrowding are many sided problems. Therefore, many solutions may be needed. 12

13 Limitations: The NYU ED algorithm does not classify hospitalized visits, nor mental health, drug/alcohol, and injury-related visits. It is assumed that hospitalized visits and a large percentage of mental health, drug/alcohol, and injury visits are not primary care-related ED visits. To the extent that some of the latter may be primary care-related, the algorithm understates the number of primary care-related ED visits. Because the data submitted is de identified, it is not possible to identify patients who make multiple ED visits. In past years it was believed that Medicaid also included SCHIP. A review of the 2007 data has revealed that most SCHIP visits were often counted as commercial or private insurance. Because some of the commercial visits could not be separated from SCHIP visits in a timely manner, actual SCHIP visits may be slightly higher in frequency. 13

14 II. PURPOSE The increasing number of hospital ED visits, many of which are primary care-related, 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 primary care-related. To achieve this purpose, the study obtained ED data from twenty five hospitals in Houston for 2006, classified the visits of Harris County residents in terms of primary care-related/non-primary care-related 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. 14

15 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 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) Working with the hospitals, a dataset was obtained, reviewed, and cleaned comprising a full set of ED visit information for the period January 1, 2007 December 31, 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. 2. Emergent-Primary Care Treatable: Treatment was required within 12 hours, but could have been provided effectively and safely in a primary care setting. 15

16 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 primary care-related 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 primary care-related use of the ED, while those in the fourth category reflect non-primary carerelated 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. 16

17 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 primary care-related and non-primary care-related 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 primary care-related categories should be interpreted as a conservative estimate. In order to get a complete picture of ED utilization, attention must also be 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,

18 paid to injuries, mental health and substance abuse, and hospitalized visits. In an effort to address this broader picture, the current report examines not only primary care-related visits, but non primary care related visits as well. 18

19 IV. RESULTS ED Visits in General During 2007, 1,060,825 total ED visits were made to all 25 participating hospitals. Of this total, 984,994, or 92.9%, visits were made by residents of the eight county area (surrounding and including Harris County), and 901,684 visits (85.0%) were made by Harris County residents. Non-hospitalized ED visits by residents of the eight county area numbered 835,251 (78.7% of total), and non-hospitalized ED visits by Harris County residents numbered 761,077. This figure of 761,077 represents 71.7% of total visits to participating hospitals (Table 1). Total ED visits to participating hospitals represent 71.0% of all ED visits to hospitals located in the Harris County area. When extrapolated to all hospitals and divided by the population, these visits translate to a total ED visit rate for Harris County residents of 32.6 per 100 residents (Table 1). Almost half (48.7%) of all non-hospitalized visits by Harris County residents were primary care-related. That figure represents 41.1% of total visits (Table 2 and Figure 2). 19

20 Table Total ED Visits to Hospitals in Harris County Study Visits 25 Hospitals Frequency Percent of first row Pct. Of sub. total A Total Study 1 Visits (all areas) 2 1,060, % 100.0% B Total Non-Hospitalized Visits (all areas) 3 897, % 84.6% C Total Hospitalized Visits (all areas) 3 163, % 15.4% D Total Study Visits (eight county area) 3 984, % 100.0% E Total Non-Hospitalized Visits (eight county area) 3 835, % 84.8% F Total Hospitalized Visits (eight county area) 3 149, % 15.2% G Total Study Visits (Harris County) 901, % 100.0% H Total Non-Hospitalized Visits (Harris County) 761, % 84.4% I Total Hospitalized Visits (Harris County) 3 140, % 15.6% Total Visits 35 Hospitals J Total All 4 Visits (all areas) 2 K Total Visits in Study Hospitals as a Percentage of Total All Visits L Estimated Number of Total Visits for Harris County 5 M Est. Number of Non-Hospitalized Visits for Harris County 6 N Est. Number of Non-Hospitalized Visits for Harris County 7 O 2007 Population of Harris County 8 1,493, % 1,269,859 1,071, ,020 3,891,420 P Total Harris Co. ED Rate per 100 population Total visits reported by 25 hospitals participating in this report. 2 All areas = all states, all countries. 3 May or may not contain a small number of visits (est. 0.5% to 1%) lacking a ZIP code or an ICD9 code. 4 Source: 2007 Cooperative DSHS/AHA/THA Annual Survey of Hospital and Tracking Database. 5Row J x percentage in Row G (85.00%). 6Row J x percentage in Row H (71.74%). 7Row J x percentage in Row I (13.25%). 8 Source: Texas State Data Center. 20

21 Table ED visits to Hospitals (N=25) by Harris County Area Residents Type of Visit 25 Hospitals Frequency TOTAL VISITS Total All Visits, Harris County 1 901,684 Total Non-Hospitalized Visits, Harris County 1 761,077 Total Hospitalized Visits, Harris County 1 140,606 CATEGORIZED NON-HOSPITALIZED ED VISITS Non-Emergent 2 145,668 Emergent, Primary Care Treatable 2 166,693 Emergent, ED Care Needed - Preventable/Avoidable 2 58,334 Total Primary Care Related Visits, eight county area 370,695 Emergent, ED Care Needed - NOT Preventable/Avoidable 85,034 Total Categorized Visits eight county area 455,729 NON-CATEGORIZED NON-HOSPITALIZ ED VISITS Injury 166,531 Mental Health Related 10,720 Alcohol or Drug Related 5,648 Unclassified 96,525 Total Non-Categorized Visits, eight county area 279,424 Percent PERCENTAGE OF CATEGORIZED NON HOSP. ED VISITS Non-Emergent 32.0% Emergent - Primary Care Treatable 36.6% ED Care Needed - Prev./Avoid. 12.8% % Categorized That Are Primary Care Related 81.3% ED Care Needed - NOT Prev./Avoid. 18.7% PERCENTAGE OF TOTAL ED VISITS Non-Emergent 16.2% Emergent - Primary Care Treatable 18.5% ED Care Needed - Prev./Avoid. 6.5% % Total That Are Primary Care Related 41.1% ED Care Needed - NOT Prev./Avoid. 9.4% Injury 18.5% Mental Health Related 1.2% Alcohol or Drug Related 0.6% Unclassified 10.7% Hospitalized Visits (all causes) 15.6% PERCENTAGE OF NON-HOSPITALIZED ED VISITS Non-Emergent 19.1% Emergent - Primary Care Treatable 21.9% ED Care Needed - Prev./Avoid. 7.7% % Non-Hospitalized That Are Primary Care Related 48.7% ED Care Needed - NOT Prev./Avoid. 11.2% Injury 21.9% Mental Health Related 1.4% Alcohol or Drug Related 0.7% Unclassified 12.7% 1 May contain missing data. 2 Primary care related. 21

22 Figure 1. Categorized ED Visits by Type (N=25) ED Care Needed -NOT Prev./Avoid., 18.7% Non-Emergent, 32.0% ED Care Needed - Prev./Avoid., 12.8% Emergent - Primary Care Treatable, 36.6% Figure 2. Total ED Visits by Type Unclassified, 10.7% Alcohol or Drug Related, 0.6% Mental Health Related, 1.2% Admitted Visits (all causes), 15.6% Injury, 18.5% Non- Emergent, 16.2% Emergent - Primary Care Treatable, 18.5% ED Care Needed -Prev./Avoid., 6.5% ED Care Needed -NOT Prev./Avoid., 9.4% 22

23 ED Visits by Month, Day of Week, and Time of Day Total ED visits by Harris County residents fluctuated during the year, peaking in March at roughly 78,000, with June having the fewest visits. While primary care- related ED visits tended to remain at a low and stable level during the summer months of June through August, non primary care related ED visits during the same period rose. The number of hospitalized ED visits remained relatively level each month (Figure 3). The peak day of the week for ED visits was Tuesday (Figure 4). The number of hospitalized ED visits remained relatively level each day. Figure 5 represents ED visits by hour of day, and Figures 5a, 5b, and 5c represent ED visits by hour of day separately for three broad age groups: a) 0-17, b) 18-64, and c) 65 and over. ED visits by children peaked between 7-8 PM, with a secondary peak at noon (Figure 5a). ED visits by adults age 18 to 64 peaked at noon, dropped slightly and remained fairly level until 8 PM (Figure 5b). ED visits by adults age 65 and over peaked from 11 AM to 4 PM, then dropped during the evening (Figure 5c). Very few ED visits by children resulted in a hospitalization, whereas as many as half of ED visits by adults age 65 and over resulted in a hospitalization. 23

24 Figure 3 24

25 Figure 4 25

26 Figure 5 26

27 Figure 5a. Age

28 Figure 5b. Age

29 Figure 5c. Age 65 and over 29

30 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). Persons who were uninsured made up 32.8% of primary care related ED visits and 30.6% of all other ED visits (non-primary care-related non-hospitalized visits and hospitalized visits). Persons with Medicaid made up 27.4% of primary care related ED visits but only 18.6% of all other ED visits. Children on SCHIP made up 3.1% of primary care related ED visits (Figures 6 & 7). Persons with commercial insurance accounted for 25.4% of primary care-related ED visits and 27.4% of all other ED visits. More than half of all ED visits by persons on Medicaid (52.7%) or SCHIP (53.5%) were primary care-related, and 35.6% of ED visits by persons on Medicare resulted in a hospitalization (Figure 8). More than two thirds (67.5%) of primary care related ED visits were by persons who were Black or Hispanic, and they accounted for 57.5% of all other ED visits. By contrast, persons who were White made up 25.9% of primary care related ED visits and 35.4% of all other ED visits (Figures 9 & 10). Just under half of all ED visits by persons who were Black or Hispanic were primary care-related. Persons who were Asian or White were the ones most likely to have an ED visit resulting in a hospitalization (Figure 11). Children age 17 and under made up more than one third (37.1%) of all primary care related ED visits, but only one fourth (25.4%) of all other ED visits. Adults age 65 and over were twice as likely to have some other type of ED visit as they were to have a primary care related 30

31 ED visit. Adults age were almost as likely to have a primary care related ED visit as some other type of ED visit (Figures 12 & 13). More than half of all ED visits by children were primary care related, and roughly one third of ED visits by adults age 65 and over resulted in a hospitalization (Figure 14). Almost six in ten (57.8%) primary care-related visits were by females and 42.2% were by males (Figure 15). For all other ED visits, 52.5% were by females and 47.5% were by males (Figure 16). Close to half (45.4%) of all ED visits by females were primary care-related, whereas percentages of ED visits resulting in a hospitalization were fairly equal between the genders (14.8% of visits by females and 14.5% of visits by males) (Figure 17). 31

32 Table 3a. Frequency of ED Visit Types by Payment Source (N=25) Type of Visit Comm SCHIP M'Caid M'care Other Gov. Other Priv. Unins Unk All Total Visits (admitted and discharged) 228,715 21, , ,587 14,516 8, ,817 4, ,558 Total Admitted Visits 21, ,829 42,553 1, , ,405 Total Primary Care Related (dc) 94,038 11, ,638 34,958 4,441 1, ,705 1, ,695 Total Injury/BH/Unclass. Visits (dc) 84,850 7,276 56,569 29,595 7,314 6,244 85,312 2, ,424 Tot. non Primary Care Related (Total minus PCR) 134,677 9,887 91,189 84,629 10,075 6, ,112 3, ,863 Non-Emergent 36,704 4,535 38,469 11,882 1, , ,668 Emergent-Primary Care Treatable 43,541 5,144 46,672 15,298 2, , ,693 Emergent Care Needed-Preventable/Avoidable 13,793 1,676 16,497 7, , ,334 Emergent Care Needed-NOT Prev./Av. (dc) 28,216 1,693 15,791 12,481 1, , ,034 TOTAL Categorized Visits 122,254 13, ,429 47,439 5,890 1, ,615 1, ,729 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 (admitted and discharged) 26.5% 2.5% 22.4% 13.9% 1.7% 0.9% 31.5% 100.0% Total Admitted Visits 17.1% 0.7% 14.9% 33.7% 1.0% 0.3% 31.6% 100.0% Total Primary Care Related (dc)* 25.4% 3.1% 27.4% 9.4% 1.2% 0.3% 32.8% 100.0% Total Injury/BH/Unclass. Visits (dc) 30.4% 2.6% 20.2% 10.6% 2.6% 2.2% 30.5% 100.0% Tot. non Primary Care Related (Total minus PCR)* 27.4% 2.0% 18.6% 17.2% 2.1% 1.4% 30.6% 100.0% Non-Emergent 25.2% 3.1% 26.4% 8.2% 1.2% 0.5% 34.8% 100.0% Emergent-Primary Care Treatable 26.1% 3.1% 28.0% 9.2% 1.2% 0.2% 31.9% 100.0% Emergent Care Needed-Preventable/Avoidable 23.6% 2.9% 28.3% 13.3% 1.1% 0.2% 30.4% 100.0% Emergent Care Needed-NOT Prev./Av. (dc) 33.2% 2.0% 18.6% 14.7% 1.7% 0.3% 29.3% 100.0% TOTAL Categorized Visits 26.8% 2.9% 25.8% 10.4% 1.3% 0.3% 32.2% 100.0% * Reflected in figures on following page. 32

33 Figure 6. Primary Care-Related ED Visits by Payment Source (N=25) Unins, 32.8% Comm. 25.4% SCHIP 3.1% Oth. Priv. 0.3% Oth. Govt. 1.2% Medicare 9.4% Medicaid 27.4% Figure 7. All Other 1 ED Visits by Payment Source (N=25) Unins, 30.6% Comm. 27.4% Oth. Priv. 1.4% SCHIP 2.0% Oth. Govt. 2.1% Medicare 17.2% Medicaid 18.6% 1 Non primary care-related non-hospitalized visits and hospitalized visits. 33

34 Table 3c. Percentage of ED Visit Types by Payment Source (read vertically) Type of Visit Commerc. SCHIP Medicaid Medicare Other Gov. Other Priv. Uninsured All Percentage Primary Care Related (DC) 41.1% 53.5% 52.7% 29.2% 30.6% 14.9% 44.8% 43.0% Percentage ED Care Needed, Not Prev. (DC) 12.3% 8.0% 8.2% 10.4% 10.0% 2.7% 9.2% 9.9% Percentage Injury/BH/Unclass. (DC) 37.1% 34.3% 29.3% 24.7% 50.4% 77.5% 31.4% 32.4% Percentage Admitted 9.4% 4.3% 9.8% 35.6% 9.0% 4.9% 14.7% 14.7% Total 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=25) 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Percentage Admitted Percentage Injury/BH/Unclass. (DC) Percentage ED Care Needed, Not Prev. (DC) Percentage Primary Care Related (DC) 34

35 Table 4a. Frequency of ED Visit Types by Race/Ethnicity (N=25) Type of Visit Asian Black Hispanic Am. Indian Other Unk White All Total Visits (adm itted and discharged) 10, , , ,856 23, , ,558 Total Adm itted Visits 1,824 38,773 30, ,441 2,735 48, ,405 Total Prim ary Care Related (dc) 3, , , ,741 9,803 96, ,695 Total Injury/BH/Unclass. Visits (dc) 3,538 73,106 88, ,859 8,852 96, ,424 Tot. non Primary Care Related (Total minus PCR) 6, , , ,115 13, , ,863 Non-Emergent 1,593 49,247 49, ,288 4,045 36, ,668 Emergent-Primary Care Treatable 1,709 53,585 58, ,867 4,519 43, ,693 Emergent Care Needed-Preventable/Avoidable ,176 18, ,586 1,240 15,737 58,334 Emergent Care Needed-NOT Prev./Av. (dc) 1,119 24,025 26, ,815 1,676 28,523 85,034 TOTAL Categorized Visits 4, , , ,556 11, , ,729 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 (adm itted and discharged) 1.2% 30.2% 31.6% 0.1% 3.0% 2.7% 31.3% 100.0% Total Adm itted Visits 1.4% 30.7% 24.4% 0.1% 2.7% 2.2% 38.5% 100.0% Total Prim ary Care Related (dc)* 1.0% 33.5% 34.0% 0.1% 2.9% 2.6% 25.9% 100.0% Total Injury/BH/Unclass. Visits (dc) 1.3% 26.2% 31.7% 0.1% 3.2% 3.2% 34.5% 100.0% Tot. non Primary Care Related (Total minus PCR)* 1.3% 27.7% 29.8% 0.1% 3.1% 2.7% 35.4% 100.0% Non-Emergent 1.1% 33.8% 34.3% 0.1% 2.9% 2.8% 25.0% 100.0% Emergent-Primary Care Treatable 1.0% 32.1% 34.8% 0.1% 2.9% 2.7% 26.3% 100.0% Emergent Care Needed-Preventable/Avoidable 0.9% 36.3% 30.9% 0.1% 2.7% 2.1% 27.0% 100.0% Emergent Care Needed-NOT Prev./Av. (dc) 1.3% 28.3% 31.5% 0.1% 3.3% 2.0% 33.5% 100.0% TOTAL Categorized Visits 1.1% 32.5% 33.5% 0.1% 3.0% 2.5% 27.3% 100.0% * Reflected in figures on following page. 35

36 Figure 9. Primary Care-Related ED Visits by Race/Ethnicity (N=25) White, 25.9% Asian, 1.0% Black, 33.5% Unk, 2.6% Other, 2.9% Am. Indian, 0.1% Hispanic, 34.0% Figure 10. All Other 1 ED Visits by Race/Ethnicity (N=25) Asian, 1.3% White, 35.4% Black, 27.7% Unk, 2.7% Other, 3.1% Am. Indian, 0.1% Hispanic, 29.8% 1 Non primary care-related non-hospitalized visits and hospitalized visits. 36

37 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 (DC) 37.1% 47.7% 46.3% 45.5% 41.5% 42.5% 35.6% 43.0% Percentage ED Care Needed, Not Prev. (DC) 10.9% 9.2% 9.8% 10.1% 10.9% 7.3% 10.6% 9.9% Percentage Injury/BH/Unclass. (DC) 34.3% 28.1% 32.5% 29.9% 34.3% 38.4% 35.8% 32.4% Percentage Admitted 17.7% 14.9% 11.3% 14.5% 13.3% 11.9% 18.0% 14.7% 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=25) 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Percentage Admitted Percentage Injury/BH/Unclass. (DC) Percentage ED Care Needed, Not Prev. (DC) Percentage Primary Care Related (DC) 37

38 Table 5a. Frequency of ED Visit Types by Age (N=25) Type of Visit All Total Visits (admitted and discharged) 262, , , , , ,558 Total Admitted Visits 14,184 13,408 29,866 32,210 36, ,405 Total Primary Care Related (dc) 137,619 48,105 91,024 65,042 28, ,695 Total Injury/BH/Unclass. Visits (dc) 92,481 38,518 70,951 51,425 26, ,424 Tot. non Primary Care Related (Total minus PCR) 124,643 62, , ,177 74, ,863 Non-Emergent 50,994 21,106 38,422 25,463 9, ,668 Emergent-Primary Care Treatable 64,093 20,967 40,468 28,353 12, ,693 Emergent Care Needed-Preventable/Avoidable 22,532 6,032 12,135 11,225 6,409 58,334 Emergent Care Needed-NOT Prev./Av. (dc) 17,978 10,592 24,704 20,542 11,218 85,034 TOTAL Categorized Visits 155,597 58, ,728 85,584 40, ,729 Table 5b. Percentage of ED Visit Types by Age (read horizontally) Type of Visit All Total Visits (admitted and discharged) 30.4% 12.8% 25.1% 19.6% 11.9% 100.0% Total Admitted Visits 11.2% 10.6% 23.6% 25.5% 29.1% 100.0% Total Primary Care Related (dc)* 37.1% 13.0% 24.6% 17.5% 7.8% 100.0% Total Injury/BH/Unclass. Visits (dc) 33.1% 13.8% 25.4% 18.4% 9.3% 100.0% Tot. non Primary Care Related (Total minus PCR)* 25.4% 12.7% 25.6% 21.2% 15.1% 100.0% Non-Emergent 35.0% 14.5% 26.4% 17.5% 6.6% 100.0% Emergent-Primary Care Treatable 38.4% 12.6% 24.3% 17.0% 7.7% 100.0% Emergent Care Needed-Preventable/Avoidable 38.6% 10.3% 20.8% 19.2% 11.0% 100.0% Emergent Care Needed-NOT Prev./Av. (dc) 21.1% 12.5% 29.1% 24.2% 13.2% 100.0% TOTAL Categorized Visits 34.1% 12.9% 25.4% 18.8% 8.8% 100.0% * Reflected in figures on following page. 38

39 Figure 12. Primary Care-Related ED Visits by Age (N=25) 45-64, 17.5% 65+, 7.8% 26-44, 24.6% 0-17, 37.1% 18-25, 13.0% Figure 13. All Other 1 ED Visits by Age (N=25) 65+, 15.1% 45-64, 21.2% 0-17, 25.4% 26-44, 25.6% 18-25, 12.7% 1 Non primary care-related non-hospitalized visits and hospitalized visits. 39

40 Table 5c. Percentage of ED Visit Types by Age (read vertically) Type of Visit All Percentage Primary Care Related (DC) 52.5% 43.5% 42.0% 38.4% 28.1% 43.0% Percentage ED Care Needed, Not Prev. (DC) 6.9% 9.6% 11.4% 12.1% 10.9% 9.9% Percentage Injury/BH/Unclass. (DC) 35.3% 34.8% 32.8% 30.4% 25.3% 32.4% Percentage Admitted 5.4% 12.1% 13.8% 19.0% 35.7% 14.7% Total 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% Figure 14. Age by ED Visit Type (N=25) 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% All Percentage Admitted Percentage Injury/BH/Unclass. (DC) Percentage ED Care Needed, Not Prev. (DC) Percentage Primary Care Related (DC) 40

41 Table 6a. Frequency of ED Visit Types by Gender (N=25) Type of Visit Female Male Unk All Total Visits (admitted and discharged) 471, , ,558 Total Admitted Visits 69,875 56, ,405 Total Primary Care Related (dc) 214, , ,695 Total Injury/BH/Unclass. Visits (dc) 138, , ,424 Tot. non Primary Care Related (Total minus PCR) 257, , ,863 Non-Emergent 88,037 57, ,668 Emergent-Primary Care Treatable 94,997 71, ,693 Emergent Care Needed-Preventable/Avoidable 31,220 27, ,334 Emergent Care Needed-NOT Prev./Av. (dc) 48,966 36, ,034 TOTAL Categorized Visits 263, , ,729 Table 6b. Percentage of ED Visit Types by Gender (read horizontally) Type of Visit Female Male Unk All Total Visits (admitted and discharged) 54.8% 45.2% 0.0% 100.0% Total Admitted Visits 55.3% 44.7% 0.0% 100.0% Total Primary Care Related (dc)* 57.8% 42.2% 0.0% 100.0% Total Injury/BH/Unclass. Visits (dc) 49.7% 50.3% 0.0% 100.0% Tot. non Primary Care Related (Total minus PCR)* 52.5% 47.5% 0.0% 100.0% Non-Emergent 60.4% 39.6% 0.0% 100.0% Emergent-Primary Care Treatable 57.0% 43.0% 0.0% 100.0% Emergent Care Needed-Preventable/Avoidable 53.5% 46.5% 0.0% 100.0% Emergent Care Needed-NOT Prev./Av. (dc) 57.6% 42.4% 0.0% 100.0% TOTAL Categorized Visits 57.8% 42.2% 0.0% 100.0% * Reflected in figures on following page. 41

42 Figure 15. Primary Care Related ED Visits by Gender (N=25) Male, 42.2% Female, 57.8% Figure 16. All Other 1 ED Visits by Gender (N=25) Male, 47.5% Female, 52.5% 1 Non primary care-related non-hospitalized visits and hospitalized visits. 42

43 Table 6c. Percentage of ED Visit Types by Gender (read vertically) Type of Visit Female Male Unknown All Percentage Primary Care Related (DC) 45.4% 40.2% 40.7% 43.0% Percentage ED Care Needed, Not Prev. (DC) 10.4% 9.3% 12.4% 9.9% Percentage Injury/BH/Unclass. (DC) 29.4% 36.1% 32.7% 32.4% Percentage Admitted 14.8% 14.5% 14.3% 14.7% Total 100.0% 100.0% 100.0% 100.0% Figure 17. Gender by ED Visit Type (N=25) 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Female Male Unknown Percentage Admitted Percentage Injury/BH/Unclass. (DC) Percentage ED Care Needed, Not Prev. (DC) Percentage Primary Care Related (DC) 43

44 Health Conditions of Patients with ED Visits The top eight discharge diagnoses of patients with a primary care-related diagnosis in 2007 were the same top eight (in the same order) as in 2006 (Table 7). In 2007 these top eight diagnoses totaled 106,095 visits, or 28.62% of the 370,695 total primary care related visits in 2007 in Harris County (cf. Table 2). Chest pain not otherwise specified has remained the top non-primary care-related diagnosis for two years in a row (Table 8). 44

45 Table 7. Most Frequent Conditions of Patients with Primary Care-Related ED Visits (N=25) 2006 visits 2007 visits Icd COUNT Description ICD COUNT Description ,247 acute uri nos ,755 acute uri nos ,870 otitis media nos ,582 otitis media nos ,069 fever ,950 fever ,277 noninf gastroenterit nec ,834 noninf gastroenterit nec ,137 urin tract infection nos ,023 urin tract infection nos ,220 acute pharyngitis ,122 acute pharyngitis 486 8,744 pneumonia, organism nos 486 9,897 pneumonia, organism nos ,644 headache ,933 headache ,599 asthma nos w(ac) exacerb ,559 abdmnal pain unspcf site ,438 convulsions nec ,526 asthma nos w(ac) exacerb ,348 abdmnal pain unspcf site ,880 vomiting alone ,718 cellulitis of leg ,718 cellulitis of leg ,143 vomiting alone ,420 hypertension nos ,810 hypertension nos ,274 acute bronchitis ,684 chest pain nec ,962 chest pain nec ,532 viral infection nos ,687 convulsions nec ,142 oth curr cond-antepartum 490 5,571 bronchitis nos ,067 lumbago ,282 oth curr cond-antepartum 490 5,039 bronchitis nos ,117 lumbago ,741 dizziness and giddiness ,546 dizziness and giddiness ,642 pain in limb ,502 viral infection nos ,554 acute bronchitis ,257 pain in limb V642 4,322 no proc/patient decision ,829 painful respiration ,848 painful respiration ,709 threaten abort-antepart ,477 threaten abort-antepart ,471 cellulitis of arm 45

46 Table 8. Most frequent Conditions of Patients with Non Primary Care-Related ED Visits (N=25) 2006 visits 2007 visits Icd COUNT Description ICD COUNT Description ,903 chest pain nos ,623 chest pain nos ,623 abdmnal pain unspcf site ,794 chest pain nec ,617 chest pain nec ,727 abdmnal pain unspcf site ,365 syncope and collapse ,572 syncope and collapse ,187 fever ,402 fever ,012 calculus of ureter ,384 calculus of ureter ,803 calculus of kidney ,854 threaten abort-antepart ,739 threaten abort-antepart ,789 calculus of kidney ,580 abdmnal pain oth spcf st ,613 acute appendicitis nos V715 1,536 observ following rape ,600 abdmnal pain oth spcf st ,471 acute appendicitis nos ,474 vomiting alone ,329 palpitations ,392 palpitations ,316 vomiting alone ,364 cholelithiasis nos ,290 headache ,356 acu brnchlts d/t oth org ,226 croup ,333 headache ,219 atrial fibrillation ,265 croup ,198 acute pancreatitis ,244 trans cereb ischemia nos ,149 acu brnchlts d/t oth org ,229 atrial fibrillation ,087 cholelithiasis nos ,225 acute pancreatitis ,074 trans cereb ischemia nos ,132 dvrtcli colon w/o hmrhg ,048 gstr/ddnts nos w/o hmrhg V715 1,006 observ following rape ,028 dvrtcli colon w/o hmrhg gastrointest hemorr nos hb-s disease with crisis gstr/ddnts nos w/o hmrhg abdmnal pain epigastric hb-s disease with crisis gastrointest hemorr nos sciatica 46

47 Length of Stay of ED Visits 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 lasted roughly half an hour less than all non-hospitalized visits (3.96 hours vs hours), whereas hospitalized visits lasted nearly five and a half hours (5.37 hours) (Table 9). Table 9. Hospitals in the Houston 911 Service Area. Mean length of stay (in hours) in the emergency department for visits of less than 24 hours. 1 Residence Discharge status Mean (sd) Harris County All Visits 4.64 (4.18) Harris County Primary care related 3.96 (3.57) Harris County All non-hospitalized 4.57 (4.08) Harris County Other (5.07) 1 All diagnoses 2 Admitted, transferred to another facility, etc. 47

48 Geographic Distribution of Patients with ED Visits The following ZIP code clusters (in clockwise order) had the highest number of total ED visits (non-hospitalized and hospitalized), primary care-related ED visits, and primary carerelated ED visits by the uninsured: 1) Aldine/Fifth Ward/northeast section of the 610 loop (both inside and outside the loop); 2) the Baytown/Ship Channel/far northeastern Harris County area; 3) Galena Park and parts of Pasadena; 4) Third Ward and South Park/central and southern part of the 610 loop; 5) southwestern (Alief/Sharpstown) area; 6) Bear Creek/intersection of interstate 10 and highway 6; and 7) Acres Homes/Spring Branch (Maps 1, 2, and 3). Maps 2 and 3 in particular show geographic patterns similar to recent years. When examined in terms of population rates, the areas with the highest rates of overall ED visits, primary care related ED visits in general, and primary care related ED visits by the uninsured are the east/northeast and south central areas of Harris County (Maps 4, 5, and 6). Map 4 also shows an elevated rate of total ED visits in the Katy area, in contrast to primary carerelated visits (cf. Maps 5 & 6). 48

49 Map 1 49

50 Map 2 50

51 Map 3 51

52 Map 4 52