hospital discharge data and vermont health surveillance charles bennett, ph.d. epidemiological...
TRANSCRIPT
Hospital Discharge Data and Vermont Health Surveillance
Charles Bennett, Ph.D.Epidemiological Surveillance Chief
Vermont Explor, Hospital Data Managers Meeting
March 14, 2004
Health Surveillance
• Health surveillance is the collection, assembly, analysis and timely reporting of health data to appropriate users.
• Hospital discharge and Emergency Department data are datasets of great importance in understanding health conditions affecting Vermonters.
Health SurveillanceFour Concerns
• Race and Ethnicity
• E-Codes
• More timely sharing of injury data
• Unique Personal Identifier Number (UPIN), a.k.a. Unique Patient Reference Number
Health SurveillanceFour Concerns
• Race and Ethnicity
• E-Codes
• More timely sharing of injury data
• Unique Personal Identifier Number (UPIN) a.k.a. Unique Patient Reference Number
Race and Ethnicity
• The Office of Management & Budget has required all federal databases that inquire about race and ethnicity to include, as a minimum, a category of Hispanic, preferably asked as a question of ethnicity separate from race.
• Information in hospital records → UHDDS, birth records, cancer registry, etc. (Intake questions are important)
Hospitalizations of Vermonters, 1997-2001
(Excluding newborns and MCD 14)
Group Per 1000 Pop. • Whites 75.0• Blacks 89.3• American Indian 7.7• Asian/Pacific Islander 30.0• Hispanic 11.2
Hospitalizations to Vermonters, 1997-2001 (Excluding newborns and MDC 14)
75
89.3
7.7
30
11.2
0
10
20
30
40
50
60
70
80
90
100
Whites Blacks AI or AN Asian or PI Hispanic
Dis
char
ges
per 1
,000
Pop
ulat
ion
Whites
Blacks
AI or AN
Asian or PI
Hispanic
Ethnicity QuestionsU.S. Standard Birth Certificate
When asked if mother/father is of Hispanic Origin, the responses are:
No, not Spanish/Hispanic/Latina Yes, Mexican, Mexican American, Chicana Yes, Puerto Rican Yes, Cuban Yes, other Spanish/Hispanic/Latina Specify___________________________
Questions on NHIS, BRFSS, YRBS, ATS, YTS (OMB)
• Which one or more of the following would you say is your race? Would you say: White, Black or African American, Asian, Native Hawaiian or Other Pacific Islander, American Indian or Alaska Native, or Other?
(Check all that apply)
1. White 2. Black or African American 3. Asian4. Native Hawaiian or Other Pacific Islander5. American Indian, Alaska Native or6. Other [specify]______________
Questions on NHIS, BRFSS, YRBS, ATS, YTS
• Which one of these groups would you say best represents your race?1. White 2. Black or African American 3. Asian4. Native Hawaiian or Other Pacific Islander5. American Indian, Alaska Native or6. Other [specify]______________
Minority Population Data in VT
Group (“Pre-bridged”) 1990 2000 ∆%Total 562,758 608,827 8.2
White 555,088 591,431 6.5
Black or AA 1,951 3,233 65.7
AI or AN 1,696 2,482 46.3
Asian or Pacific Isl. 3,134 5,538 72.3
Other Race 808 1,443 78.6
Two or More Races 0 7,335 --
Hispanic or Latino 3,661 5,245 43.3
(of any race)
Minority Population Data in VT
“Bridged” Group 1990 2000 ∆%
Total 562,758 608,827 8.2
White 555,088 595,963 7.4
Black or AA 1,951 4,051 107.6
AI or AN 1,696 3,077 81.4
Asian or Pacific Isl. 3,215 5,736 78.4
Other Race N.A. N.A.
Two or More Races N.A. N.A.
Hispanic or Latino 3,661 5,504 50.3
(of any race)
Population Change 1990-2000
8.2% 6.5%
65.7%
46.3%
72.3%
43.3%
8.2% 7.4%
107.6%
81.4%78.4%
50.3%
0.0%
20.0%
40.0%
60.0%
80.0%
100.0%
120.0%
Total White Black or AA AI or AN Asian or PI Hispanic
Perc
ent C
hang
e 19
90 to
200
0
Pre-bridged
Bridged
Issue of “Bridging”
• Current “bridging” is based on National model based upon National Health Interview Survey data.
• Vermont BRFSS yields very different result.
• Even first listed race, for persons of multiple race, is not always the one group that best represents respondent’s race.
Allocation of “Two or More Races” into “Bridge” Categories
Race NCHIS (NHIS) VT BRFSS
White 4532 74% 4661 76%
Black or African
American 818 13% 236 4%
American Indian/Native
Alaskan 595 10% 1069 17%
Asian/Pacific Islander 198 3% 177 3%
Total 6143 100% 6143 100%
Health SurveillanceFour Concerns
• Race and Ethnicity
• E-Codes
• More timely sharing of injury data
• Unique Personal Identifier Number (UPIN) a.k.a. Unique Patient Reference Number
E-Codes
• E-code reporting with injury diagnosis
• Location coding
• Perpetrator coding
E-Codes
• E-code reporting with injury diagnosis E-code reporting has greatly improved since
the addition of a special field in the data base. In recent years, approximately 97 percent of
all reports with a principle diagnosis of injury are accompanied with a correct E-code for nature of injury
The nature of injury code, however, is not always in the E-code field.
E-Codes
• E-code reporting with injury diagnosis
• Location coding
E-Codes• Location coding
For unintentional injuries, location is reported about 40 percent of the time.
Unintentional injuries are the “bulk” of injuries.
Perhaps a second E-code field would help to encourage reporting.
E-Codes
• E-code reporting with injury diagnosis
• Location coding
• Perpetrator coding
E-Codes• Perpetrator coding – Intentional assaults
The reporting of perpetrator codes for intentional assaults on children/youth (under 18 years of age) is, in recent years, running about 33 percent.
Reporting of perpetrator is called for in most recent ICD-9-CM code book for adults also.
Very few rapes have a perpetrator code. Some 29 hosp. discharges have perpetrator codes
without nature of injury e-code (1997-2001). In 2002, 124 ED records had perpetrator codes
without nature of injury e-code.
Health SurveillanceFour Concerns
• Race and Ethnicity
• E-Codes
• More timely sharing of injury data
• Unique Personal Identifier Number (UPIN) a.k.a. Unique Patient Reference Number
Timely Reporting of Injury Data
• Current reporting is about 15 to 27 months after the events in question.
• CDC is pressing state programs for more timely surveillance of injury conditions.
• Quarterly reporting of hospital and ED data would permit response to emergent conditions.
• Reporting of suicide attempts may help to minimize extent of contagious condition.
Health SurveillanceFour Concerns
• Race and Ethnicity
• E-Codes
• More timely sharing of injury data
• Unique Personal Identifier Number (UPIN), a.k.a. Unique Patient Reference Number
Unique Patient Reference Number• For purposes of health surveillance, it is very
important to distinguish between incidence and prevalence.
• UPIN facilitates accurate comparison of multiple records – to identify new cases as opposed to patients with repeated hospitalizations.
• UPIN would greatly enhance understanding of health conditions in VT – and guiding of programs in response.
Thank you for your attention . . . and your help.