Comparative Summary of States' Adverse Event
Reporting and Performance Experiences

• 13 states publish Adverse Events / Medical Error Data
Reporting by States of adverse event data on the Internet did not begin until 1999-2000. Aggregate data of reportable event performance are available for hospitals in 13 states in 2006. However, in Colorado and Washington, these aggregate data are only available to the Public upon request, and this raw facility-identified data must be collated and analyzed to make it interpretable.

Among these 13 States, facility-identifed data is available in three States. Only Minnesota provides a partial facility-level analysis on the Internet. Colorado publishes redacted adverse event data by occurrence and by facility on the Internet but does not collate or compare facilities. Washington data prior to June 7, 2006 is available for facility-level and aggregate data analysis.¹ Although not reporting in 2006, nascent programs in Illinois and Indiana indicate an intent to report facility-identified data on the Internet by 2008.

Ten (10) States (CT, FL, MA, MD, ME, NY, PA, TN, TX, UT) have published only aggregate State data for reportable events in variable depth. These States provide no facility-identified information. Wyoming's new program will publish only aggregate facility data beginning December 31, 2006, and Washington's revised program will begin reporting aggregate data on January 1, 2008.
• Numbers of Specified Events and Definition Origins
States collect data with different measures and/or different definition origins, including:
- ◊ NQF's 27 ‘Never’ Events Definitions: 2 states - (Connecticut and Minnesota) use NQF definitions; Connecticut also reports 6 state-specific adverse events. Illinois and Indiana anticipate reporting facility-identified NQF data beginnning in 2007.
- ◊ ICD-9CM Occurrence Codes Definitions: 2 States, New York and Tennessee, utilize Occurrence Codes based on International Classification of Diseases, 9th Edition, Clinical Modification (ICD-9-CM) taxonomy of administrative data. New York uses 54 discrete measures, whereas Tennessee includes 48 measures.
- ◊ Unique State-specific Definitions: 5 states (Colorado, Connecticut, Florida, Maryland, Massachusetts) employ unique state reportable event definitions; (Connecticut has 6 state-specific measures, in addition to 27 NQF measures).
- ◊ JCAHO Definitions: 5 states (Maine, Pennsylvania, Utah, Texas, Washington¹) utilize varying numbers and versions of JCAHO's ‘voluntary reportable sentinel events.’

¹ Washington facility-identified data are no longer available to the public after mid-2006 due to changes in the reporting program. Washington will provide only aggregate state data after the switch to NQF definitions on June 7, 2006.

– Figure 1. Specified Reportable Events –
Numbers and Definition Origins in
13 States that Publish Aggregate State Data





STATE	DEFINITION ORIGINS	NUMBER LISTED EVENTS


New York	ICD-9 Occurrence Codes	54
Tennessee	ICD-9 Occurrence Codes	48
Connecticut	NQF + Unique	33
Massachusetts	Unique	33
NQF	NQF	27
Minnesota	NQF	27
Pennsylvania	JCAHO + Unique	21
JCAHO*	JCAHO	19
Florida	Unique	17
Colorado	Unique	14
Utah	JCAHO + Unique	13
Washington	JCAHO + Unique	12
Texas	JCAHO	11
Maine	JCAHO	10
Maryland**	Unique	5



* JCAHO definitions based on JCAHO's ‘voluntary reportable sentinel events’ and other examples of sentinel events published on JCAHO's website.

** Maryland has released partially analyzed aggregate data for Level 1 Adverse Events only [PDF]. As defined by statute, each hospital is required to "list and describe examples of adverse events that shall be reported," thereby developing a unique hospital-specific list of Level 1 adverse events [COMAR 10.07.04B(2)]. As such there are no standard definitions for reportable events across hospitals in the state.

• Literature Documenting Medical Errors / Adverse Events Problem

The two key publications supporting the IOM's To Err is Human are the Harvard Medical Practice Study (HMPS) published in 1991 that utilized 1984 New York discharge data^[1] and a study published in 2000 that used 1992 Colorado and Utah discharge data.^[2]

^[1] [Brennan TA, Leape LL, Laird NM, Hebert L, Localio AR, Lawthers AG, Newhouse JP, Weiler PC, Hiatt HH, "Incidence of adverse events and negligence in hospitalized patients. Results of the Harvard Medical Practice Study I." N Engl J Med 1991; 324:370-6.] (Reprint available at Qual Saf Health Care 2004;13:145-151)

^[2] Thomas EJ, Studdert DM, Burstin HR, Orav EJ, Zeena T, Williams EJ, Howard KM, Weiler PC, Brennan TA, "Incidence and types of adverse events and negligent care in Utah and Colorado." Med Care 2000; 38:261-71

The HMPS is useful in developing a scheme for standards and/or expectations for interpreting reported adverse events across States. The table below reveals the categories of incident adverse events (AEs) and negligent adverse events (NAEs) by % that were uncovered in the HMPS. A similar categorization was not created in the Utah/Colorado study.

Table 1.
HMPS Categories of incidence of
Adverse Events (AE) and Negligent Adverse Events (NAE)




Category	AE Timing relative to Index Hospitalization		AE (%)	Negligent AE (%)

	Time of Occurrence	Time of Discovery

1	During	During	50.6 %	51.0 %
2	During	After – OutPatient	6.1 %	2.3 %
3	During	After – Later Hospital Stay	5.2 %	6.2 %
4	During or Before	During	13.1 %	19.3 %
5	Before	During	25.0 %	21.2 %

The data published by both New York and Colorado/Utah studies provided incident values for AEs and NAEs for events that were discovered during the index hospitalization and which occurred either during or prior to the index hospitalization. HMPS categories 1, 4, and 5 satisfy this criterion, leaving out categories 2 and 3, and these are the incident categories that make up the data that is used as the foundation for reported incidence of medical errors in the IOM study, To Err is Human.

However, State reporting programs for adverse events and medical errors do not follow this temporal methodology. State programs provide instructions for TEMPORAL monitoring of only HMPS category 1 events that occur during and are discovered during the index hospitalization. In the table below, the temporal category of State reporting (category 1 only) is compared to all event categories identified in the HMPS (categories 1+2+3+4+5), and to the events reported in the HMPS and by the IOM (categories 1+4+5). These category 1 adverse events make up a little over half of all adverse events and negligent adverse events noted in the 1984 New York dataset and slightly more than less than 60% of the AEs and NAEs reported in the IOM study.

Table 2.
HMPS Adverse Event Incidence Summary by
Category of New York 1984 Discharge Data




Category	Category Timing Relative to Index Hospitalization		SAMPLE DATA		POPULATION DATA*

			Total AEs	Total NAEs	Total AEs	Total NAEs	Events per 100,000 DCs

	Occurrence	Discovery					AEs	NAEs

Sum (1,2,3,4,5)	During or Before	During or After	1,278	306	111,461	29,812	4,172	1,116
Sum (1,4,5)	During or Before	During	1,133	280	98,608	27,179	3,691	1,017
1	During	During	647	156	55,046	15,257	2,060	571


*	Population numbers based on 1984 New York Discharge volume of 2,671,863

Based upon this data from the HMPS, the interpretation of this table is that State adverse event reporting programs that predominantly measure events that both occur and are discovered during an index hospitalization will yield

2,060 Adverse events per 100,000 hospital discharges AND

571 Negligent Adverse Events per 100,000 hospital discharges.

These numbers assume that all categories and types of adverse events that were measured in the HMPS are measured and properly recorded. Since most State programs report a subset of all recognized adverse events, it is unlikely that most States' reporting programs would ever satisfy these criteria, even under an assumption of full reporting compliance.

• Comparison of State Reported Medical Errors / Adverse Events

Table 3.
Adverse Events and Hospital Discharges
with Indexed Incidence Values




State / Entity	Reporting Year(s) and Reportable Events		Population Totals		Events per 100,000 Discharges

			Hospital AEs	Hospitals Discharges	Aggregate Number	Indexed to MN (NQF)*



HMPS-NY¹	1984	All	55,046	2,671,863	2,060	141
New York	2001	54	28,689	2,473,930	1,160	80
Pennsylvania	2004-2005	21	28,689	1,702,971	407	28
Tennessee	2004	48	1,313	773,598	170	12
Massachusetts	2001-2004	33	2,010	3,305,147	60.8	4.2
Colorado	2000-2004	14	1,404	2,314,649	60.7	4.2
Connecticut²	2004-2005	33	206	500,169	41.2	2.8
Florida	2000-2004	17	4,287	11,838,434	36.2	2.5
Maryland	2004-2005	5	237	1,340,508	17.7	1.2
Minnesota³	2004-2005	27	202	1,384,892	14.6	1.0
Connecticut³	2004-2005	27	72	500,169	14.4	1.0
Maine	2004-2005	10	50	315,624	15.8	1.1
Utah	2002-2004	13	108	780,780	13.8	0.9
Washington	2000-2004	12	269	2,775,574	9.7	0.7
Texas⁴	2004-2005	11	184	2,565,007	7.2	0.5

*	Events per 100,000 discharges are indexed to Minnesota's NQF-based event definitions
¹	Harvard Medical Practice Study Events numbers based solely upon events that occurred and were discovered during index hospitalization
²	NQF + 6 Connecticut-specific Events
³	NQF Events only
⁴	Texas collects data with each facility's annual license renewal, so that first year of data represents composite of 12 different monthly collection periods.

Negligent Adverse Events
NY (1984) – HMPS

All Adverse Events
Occurring and Discovered during Hospitalization
NY (1984) – HMPS

Figure 2.
Index Of Hospital Reportable Events
per 100,000 Discharges by State

STATE & YEAR	INDEX RELATIVE TO MN^†


New York - HMPS (1984) Discharges - Adverse Events Discharges - Negligent AEs	141 39
Based upon projections from the Harvard Medical Practice Study, these adverse events represent the same proportion of adverse events that occurred during and were discovered during the index hospitalization. The HMPS did not measure adverse events per patient-day.
New York (2001) Discharges Patient-days	80 60
Pennsylvania (2004) Discharges Patient-days	28 26
Tennessee (2004) Discharges Patient-days	12 11
Massachusetts (2002-04) Discharges Patient-days	8.3 7.7
Colorado (2000-04) Discharges Patient-days	4.2 4.8
Florida (2000-04) Discharges Patient-days	2.5 2.3
Connecticut ALL (2004-05) Discharges Patient-days	2.8 2.6
Connecticut NQF (2004-05) Discharges Patient-days	1.0 0.9
Maryland (2004-05) Discharges Patient-days	1.2 1.1
Maine (2004) Discharges Patient-days	1.1 1.0
Minnesota (2004-05) Discharges Patient-days	1.0 1.0
Utah (2002-04) Discharges Patient-days	0.9 1.1
Washington (2004) Discharges Patient-days	0.7 0.8
Texas (2004-05) Discharges Patient-days	0.5 0.4


† Minnesota used as Referent Group (MN = 1.0)

Comparative Summary of States' Adverse EventReporting and Performance Experiences

– Figure 1. Specified Reportable Events –Numbers and Definition Origins in13 States that Publish Aggregate State Data

Table 1. HMPS Categories of incidence ofAdverse Events (AE) and Negligent Adverse Events (NAE)

Table 2. HMPS Adverse Event Incidence Summary byCategory of New York 1984 Discharge Data

Table 3. Adverse Events and Hospital Dischargeswith Indexed Incidence Values

Figure 2. Index Of Hospital Reportable Eventsper 100,000 Discharges by State