Skip to content
FHWA Safety: First graphic from left courtesy of (http://www.pedbikeimages.org/Dan Burden)
Home > Five Percent Reports

Maryland 2007 Five Percent Report

This report is in response to the Federal requirement that each state describe at least 5 percent of its locations currently exhibiting the most severe highway safety needs, in accordance with Sections 148(c)(1)(D) and 148(g)(3)(A), of Title 23, United States Code. Each state's report is to include potential remedies to the hazardous locations identified; estimated costs of the remedies; and impediments to implementation of the remedies other than costs. The reports included on this web site represent a variety of methods utilized and various degrees of road coverage. Therefore, this report cannot be compared with the other reports included on this Web site.

Protection from Discovery and Admission into Evidence—Under 23 U.S.C. 148(g)(4) information collected or compiled for any purpose directly relating to this report shall not be subject to discovery or admitted into evidence in a Federal or State court proceeding or considered for other purposes in any action for damages arising from any occurrence at a location identified or addressed in the reports.

Additional information, including the specific legislative requirements, can be found in the guidance provided by the Federal Highway Administration,
http://safety.fhwa.dot.gov/safetealu/fiveguidance.htm.

Maryland
5 Percent Report

I. Introduction1

"Section 1401 of the Safe, Accountable, Flexible, Efficient Transportation Equity Act: A Legacy for Users (Public Law 109-59), or SAFETEA-LU, amended Section 148 of Title 23, United Sates Code, to create a new Highway Safety Improvement Program (HSIP) as a "core" FHWA program with specific funding. As a condition for obligating HSIP funds, Section 148(c)(1)(D) requires States to prepare an annual report that describes not less than 5 percent of their public road locations exhibiting the most severe safety needs. The legislation requires that these reports also include an assessment of potential remedies at the locations identified, the estimated costs of the remedies, and impediments to their implementation other than costs"2.

II. Data

  1. Current Data Systems Status and Years of Available Data

    1. Crash data availability

      Collision data that includes all severity levels (Fatal, Injury, Property damage) is available in an electronic database for CY 1994 through CY 2006. Accident data from CY 1980 to CY 1993 is available in an older, limited electronic format. Information on fatal accidents prior to CY 1980 may be available and can be researched on an individual basis. Note: site specific data on individual accidents occurring on roads maintained by Baltimore City (except for fatals) is not available through the SHA system.

    2. Traffic data availability

      Traffic data from CY 1994 through CY 2006 that is integrated with crash data is available in the current database format. Traffic data that was used with collision data prior to CY 1994 would have to be researched from electronic and non-electronic archive sources.

  2. Coverage of Local Roads

    1. Estimate of percentage of coverage of entire State's public roads

      Of the 30,633 miles of public highways in the State of Maryland, SHA has complete traffic volume and accident data by specific linear logmile location for the 5,235 (17%) miles of highways that are state maintained. On the 23,475 (77%) miles of non-state road systems (except those maintained by Baltimore City), SHA has full linear logmile location data on all reported accidents and general traffic volume data. Specific traffic volume data by location for all non-state maintained highways is unavailable at this time. For the 1,898 miles (6%) of Baltimore City maintained road systems, SHA has overall crash and volume totals but does not have specific locational accident or traffic volume data.

  3. Any Additional Limitations (if applicable)

    • Collision data for Baltimore City surface streets is not available to SHA in a format that is compatible with SHA's high accident location identification process.
    • Microfilm copies of MAARS collision reports on Baltimore City surface streets are not available to SHA.
    • Baltimore City is undertaking its own process of identifying hazardous locations and needed safety improvements.
    • Traffic volumes for non-SHA maintained highways in Maryland are incomplete, preventing the use of frequency/rate based model used for identifying hazardous locations on the SHA system.
    • Specific responsibilities for reviewing non-State maintained systems has not yet been determined.
    • The possibility of basing the selection of locations for the 5 Percent Report on fatalities and serious injuries, given that only 6,000 such accidents occur on 65,000 public roads, annually, is problematic.

    1. Limitations (i.e., identification of crash severity, cannot identify intersection crashes)

      No additional limitations beyond those noted in previous items.

    2. Discussion of and schedule for data systems upgrade

      In 2005, an upgrade to the MAARS accident data collection and processing system was initiated in order to decrease the delay between the time an accident report is prepared and when the report becomes useful data in an electronic format. The initial stage involves the replacement of report microfilming with electronic scanning and converting the data storage to an Oracle platform on an SHA server. Equipment to implement the initial stage of that upgrade was delivered to MSP on the first of August 2007, which should enable that important stage to be completed by the end of the calendar year. Further information on this process will be reported in 2008.

      "The eMAARS system will be a real-time web application designed to improve all aspects of the data collection process. It will provide tools to increase accuracy, efficiency, and effectiveness for each collection function. While its first design will be to support the in-house data processes specifically receiving paper reports, it will also support future processes that deal with electronic collection of data. It will contain a series of integrated applications that provide an enterprise-wide solution for architecture, data capture, management and tracking, diagnostic tools, and trend analysis.3

III. Analysis (Determination of 5% List)

  1. Data Used in Analysis

    1. Years of data

      CY 2005 collision and traffic volume data were used for accident and severity frequency tabulations and rate calculations (at the time this report is being written this is the most current data year that has been fully analyzed). The prior two years (CY 2003 and CY 2004) collision data also is provided to the SHA District Offices for review.

    2. Crash severity levels

      Fatal, injury, and property damage accidents are used in accident and severity rate calculations. See Exhibit 7 for additional details.

    3. Use of corridors, or separation of crashes into segments and intersections

      Currently, intersections are considered for the 5 percent list because of the availability of severity rates for these sites and experience that show that these locations have the most severe safety problems. Highway corridors and sections may be used in future 5 percent listings if severity rates become available for these locations.

  2. Methodology to Identify 5% Locations

    1. Description of methodology

      1. Methodology and metric (frequency, rate, etc.) to identify locations with severe safety needs (i.e., more than twice the statewide average rate)

        See Exhibit 7 for the explanation of that methodology, which includes a crash rate criterion.

      2. Methodology and metric to identify to 5% (if not using 5%)

        Based on 2005 crash data, 138 Priority Candidate Safety Improvement Locations were identified. These are the locations along the state system demonstrating the greatest safety needs. Seven locations (5 percent of 138) were identified, using a weighted severity index that gives heavy weight to fatalities and incapacitating injuries, for inclusion in the 5 Percent report.

        The analysis of the 5 Percent locations is shown in Table IV. Estimated costs were the most current figures available at the time of this report and are subject to final adjustments.

      3. Discussion of development of threshold value to compare locations (i.e., Statewide average fatal and serious injury rate)

        Up to this point, Candidate Safety Improvement Locations have been identified only on the state maintained highway system. Although several local jurisdictions use the accident data, which SHA provides to them annually, to identify similar location of note, the practice is far from universal. Baltimore City has advised SHA that the City now is undertaking a process to identify its most hazardous locations and the needed remedial measures.

        In response to the requirements in 23 US Code §148 (§1401 of SAFETEA-LU), SHA will be ascertaining the extent to which local jurisdictions are determining hazardous locations and needed improvement on their road systems and, in cooperation with the local agencies, will be ascertaining the best approach to initiating such processes where they are missing. Decisions must be made as to the extent that SHA could or should perform the hazardous location identification processes for the local jurisdictions, how it otherwise could best assist the local jurisdictions that will perform the processes themselves, and the exact nature of those identification processes.

        Currently traffic volumes are not available for many local roads, a circumstance that likely will obviate the use of the number-rate methodology now used by SHA. Assuming that the appropriate federal agencies will be attending to their roads, the processes of identifying hazardous locations on all public roads throughout the state will be in place prior to the August 31, 2009 deadline.

        As previously noted, the identification of Candidate Safety Improvement Locations is based on all collisions included in the MAARS database, regardless of the crash severity. The number-rate method does not assign a higher weight to more severe crashes. However, when identifying candidate intersections, for example, severity is an additional consideration in that an intersection that might not otherwise be a Candidate Safety Improvement Location can become one if its severity index is unusually high (crashes are weighted 5, 4, 3, 2, and 1 for fatal, incapacitating injury, non-incapacitating injury, possible injury, and property damage only, respectively).

        SHA will be looking for ways to better account for accident severity in future years. First, it plans to change the aforementioned crash weighting to give greater weight to more severe crashes. Second, it will consider the extent to which Candidate Safety Improvement Locations identification might be based solely on more severe crashes, such as using fatally and incapacitating injury crashes as suggested in §148 and as required in the identification of High Risk Rural Roads. Several considerations must be recognized before such a step it taken.

        The identification of Candidate Safety Improvement Locations now is based on 102,000 4 accidents per year occurring along nearly 31,000 5 miles of highway, on the average for all roads more than 3 accidents per mile. For the state system, there are some 47,000 6 accidents over 5,200 7 miles of highway or more than 9 accidents per mile. If the selection were based solely on fatal and incapacitating injury accidents, 95 percent of the accidents would be ignored. The selection would consider only some 5,400 8 accidents per year or fewer than 1 crash per 5 miles for all roads. For state highways, 3,300 9 crashes or fewer than 1 accident per mile would be considered. For non-state roads, the identification process would be looking at an average of 1 fatal or incapacitating injury accident for every 17 miles.

        Although SHA has not conducted sensitivity analysis to ascertain the appropriateness of basing hazardous location information on such sparse data, doing so appears to be problematical. Engineering agencies long have decried the loss of valuable information on the frequency and characteristics of crashes when law enforcement agencies substantially reduced accident reporting. It should be recognized that the 102,000 accidents reported annually, reflect the filtering out of most minor collisions. It is estimated that only about 20 to 25 percent of the property damage only accidents are reported by police, mostly those accidents in which a vehicle is damaged to the extent it must be towed from the scene. Nevertheless, SHA will consider if limiting the Candidate Safety Improvement Location identification processes/analyses to severe accidents is feasible. Any such change obviously also would affect those locations selected for inclusion in the 5 percent report. (Note: cited numeric totals have been rounded)

    2. 5 Percent List(s)

      The locations selected for the 5 Percent report are shown in Table IV

    3. Limitations of methodology

      See item III b 3 for a discussion of limitations of the present system, which was basically designed for state maintained highways only.

IV. Remedies

  1. Discussion of process for identifying potential remedies

    1. Remedies for individual locations

      The locations selected for the 5 Percent report are shown in Table IV

  2. Cost of remedies

    1. Typical costs for general remedies

      The estimate costs for the proposed improvement to the selected locations are presented in Table IV

  3. Non-cost impediments to implementation (public or political issues, geometric design issues, high traf fic volumes, etc.)

    None are known for locations shown. Any non-cost issues would be reflected in special notations by site.

V. State Contacts for Additional Information

Robert Cunningham, Team Leader, Crash Analysis Safety Team, Traffic Safety Analysis Division, Office of Traffic and Safety, Maryland State Highway Administration, 7491 Connelley Drive, Hanover MD 21076

Table IV

1 Note: the outline format is based on the example shown in " 5 Percent Reports Noteworthy Practices and Examples", Attachment, Erin Kenley, FHWA, June 4 , 2007

2 Memorandum, ACTION: Highway Safety Improvement Program Reporting "5 Percent Report", Title 23, Section 148 (c)(1)(D), Michael L. Halladay, FHWA, April 10 2006

3 eMAARS Project Plan Version 1.3, October 2005, OOTS

4 Statewide Crash Profile Sheet, Pg 1, May 1 2007, Maryland Highway Safety Office

5 Highway Mileage on State, State Toll, County and Municipal Systems, Pg 4, 2005, Highway Information Services Division

6 State System (IS,MD,US) Routes Crash Profile Sheet, Pg 1, May 1 2007, Maryland Highway Safety Office

7 Highway Mileage on State, State Toll, County and Municipal Systems, Pg 4, 2005, Highway Information Services Division

8 Incapacitating Injuries & Fatal Crash Profile Sheet, Pg 1, May 30 2007, Maryland Highway Safety Office

9 State System (IS,MD,US) Incapacitating & Fatality Crash Profile Sheet, Pg 1, May 30 2007, Maryland Highway Safety Office

EXHIBIT 7

Identifying Candidate Safety Improvement Locations

To provide a safe highway system to our constituents and to take advantage of securing available federal safety funding, the Maryland State Highway Administration has developed procedures for identifying Candidate Safety Improvement Locations on State maintained highways. Listings of Candidate Safety Improvement Locations are developed annually. Those locations having the highest accident experience and/or accident rate are examined to determine the appropriate safety improvements. Candidate Safety Improvement Locations are stratified into two categories: Intersections (CSII) and Sections (CSIS). The criteria for the two categories vary for obvious reasons. The following is an explanation of how the criteria were developed for each grouping.

Candidate Safety Improvement Intersections

Initially, it is necessary to determine the number of intersection related accidents on State designated highways for each of the 23 counties. The data collected is stratified by county because of the diversity in population and traffic volume characteristics existing in the various geographical locations throughout Maryland. This process is normally done by computer software and very simply involves selecting those accidents having been coded as being intersection related by the investigating police officer. The route designations we use in this selection have the following prefixes: MD, US, and IS.

It is necessary to determine the number of intersections that are located on highways having the prefixes indicated above, for each county. This count of intersections does not include driveway intersections, interchanges, or ramp merges on fully controlled access highways. From the information gathered, the number of intersections per county and the number of intersection accidents per county, the average number of intersection accidents per intersection is derived for each county. For example, in Baltimore County there were a total of 2,497 intersections and 1,987 intersection related accidents, yielding an average of 0.80 accidents per intersection in that county. This is repeated for each county. The next step in the process is to determine what intersection accident frequency is significantly higher than the average for its respective county.

Determining the level of significance is done by using the Cumulative Poisson Distribution Table and selecting the appropriate proportion level desired. It has been our practice to use a Poisson proportion level having a confidence interval of 1.000 as a first step in identifying those intersections experiencing an abnormally high number of accidents. To illustrate, if the average number of intersection accidents (0.80) in Baltimore County is applied to the Poisson Table using confidence interval of 1.000, 5 or more intersection accidents at a location would be considered significant and 5 would be used as the initial cut-off number.

Using the Cumulative Poisson Distribution table to illustrate that if the average number of intersection accidents (0.80) in Baltimore County is applied to the Poisson Table using confidence interval of 1.000, 5 or more intersection accidents at a location would be considered significant and 5 would be used as the initial cut-off number.

This initial cut-off number is considered the minimum number of intersection accidents for a location to be included in the preliminary listing of locations of interest In that some of our counties are very rural and experience very low traffic volumes, those counties would achieve a very low minimum cut-off number (less than 3 accidents). Since it would not be wise to allocate funding for implementation of improvements to a location having less than 3 accidents, a minimum cut-off number for any counties is set at 3 accidents.

To insure that we are identifying only the most hazardous locations in each county, the criteria now becomes more stringent. From the listing of locations meeting the minimum cut-off numbers, a new, reduced listing of those locations experiencing an accident frequency two times the minimum cut-off number for its respective county is developed. The locations on this listing are considered Candidate Safety Improvement Intersections (CSIIs). For example, in Baltimore County, which has a cut-off number of 5 intersection accidents, an intersection must have a minimum of 10 accidents to be included on the listing of Candidate Safety Improvement Intersections. For each of these intersections, traffic volume data for all approaches are collected and accident rates per million vehicles entering (rate/MVE) the intersections are calculated. The following is an example of the rate calculation.

Rate calculation example

The accident rate/MVE is used to further stratify the CSIIs. Any location meeting a threshold level of 1.0 accident/MVE is categorized as a Priority Candidate Safety Improvement Intersection. Doing this identifies the most critical locations and enables the District staffs to give priority attention to those locations where safety improvements likely are of the greatest need.

Finally, a severity rate has been developed as yet another tool to prioritize CSII listings. The severity rate is used in determining the rank of locations that are otherwise equal in terms of accident rate. The severity rate also is used to further address certain intersections that have met the accident frequency cutoffs, but whose accident rate is less than 1.0

The severity rate is an accident rate that has been weighted to adjust for accident severity. Each of the severity classifications (fatality, incapacitating injury, non-incapacitating injury, possible injury, and property damage only) are assigned weighting factors as follow: (Note it is anticipated that greater weights will be given to more severe crashes in future years.)

Severity Weighting Factor
Fatality 5
Incapacitating Injury 4
Non-incapacitating Injury 3
Possible Injury 2
Property damage only 1

Example of determining accident severity (using figures from the previous example):

If a location experienced 20 accidents that were distributed by severity as follows:

Category Weighting Factor Frequency Weighted Frequency
Fatality 5 1 5
Incapacitating injury 4 2 8
Non-incapacitating injury 3 5 15
Possible injury 2 7 14
Property damage only 1 5 5
Total   20 47

The severity rate would then be computed using the severity index of 47 as shown:

Severity Rate is egual to the product of 47 time 1,000,000 divided by the product of 37,500 time 365, which is egual to 3.43

The District staff is requested to study those locations whose severity rates are 2.5 or larger.

Candidate Safety Improvement Sections

The listing of Candidate Safety Improvement Sections is totally computer generated. This process requires developing a listing of all non-intersection accidents on the State system throughout the State. From this listing and using the Highway Performance Monitoring System's traffic volume data, statewide accident rates per 100 million vehicle miles of travel are calculated and grouped by type of access control (number of lanes; divided/non-divided; full control, partial control, or no control of access; urban or rural designation). This information is used in the Critical Rate Formula further along in the process. At this point it is necessary to arrive at an accident frequency cut-off number to be used in eliminating those sections of highway having few accidents and low traffic volumes -- sections having only 1 or 2 accidents, but because of their low traffic volumes, have relatively high accident rates.

Annually, there are approximately 50,000 accidents on the State maintained system. The State of Maryland maintains approximately 5,000 miles of highways. This results in an average of approximately 10 accidents per mile. A section length of ½ mile has been selected. To identify the most hazardous sections, i. e., those experiencing large clusters of accidents, a 1/2-mile section is slide along each route seeking to find every section having 10 or more accidents or twice the average per mile rate. The sliding scale program in essence slides a half-mile section every 100th of a mile. For example, if there is an accident at logmile 0.00 the section will cover the area of 0.00 to 0.50. If the next accident is at 0.01, the start of the 1/2-mile section will slide to 0.01 miles and cover the portion of the route from 0.01 to 0.51. This process will continue for each accident along each route. Those sections having 10 or more accidents (a per mile rate of 20 accidents per mile) become Candidate Safety Improvement Sections (CSISs). This file is merged with the Highway Performance Monitoring System (HPMS) file, which incorporates the traffic volume data along state roads. Using the accident data from the CSIS file and the traffic volume data from the HPMS file, accident rates per 100 million vehicle miles of travel are computed for each CSIS.

A statistical analysis then is performed on each CSIS in each district. An important part of the analysis is the segregation of candidate safety improvement sections by number of lanes and the type of access control for each section so that the analysis identifies those sections that are experiencing significantly high accident rates when compared to highways with similar design characteristics. The statistical procedure used in this step is known as Donald A. Morin's Rate Quality Control Method (Upper Control Limit). This procedure determines an upper control limit accident rate using the following equation:

upper control limit accident rate equation

Where:
Rcu = the computed upper control limit accident rate (critical value) in 100 miles of travel.
Ra = the statewide average accident rate for specific category of highway.
K = a level of probability constant. This value is set at 1.645 and provides a 95 percent probability level that the upper control limit could be exceeded by chance variation in the observed accident rate.
M = vehicle miles of travel in the study section for the study period divided by 100,000,000.

The final step is to compare the accident rate for each CSIS to the upper control limit accident rate for that highway category. When the difference between the study section rate and upper control limit rate exceeds the upper control limit rate, the location is considered a Priority (P) Candidate Safety Improvement Section. To illustrate:

Study RateUpper Control RateDifferenceCSIS
305116189Priority

Again, the priority location procedure identifies the most critical locations and enables the District staffs to give priority attention to those locations where safety improvements likely are of the greatest need.

Three-year Combined Candidate Safety Improvement Sections

Additionally, ½ -mile sections experiencing abnormal numbers or proportions of accidents having various characteristics are identified. Referred to as three-year combined CSISs, these listings are based on half-mile sections of state highway having at least 10 accidents during the three years. Sections where the percentage of accidents involving fixed objects, heavy trucks, trucks parked along controlled access highways, or wet pavement surfaces or that occur at nighttime are twice the average statewide percentage for that respective category are identified. The wet surface listing is also merged with the Office of Materials friction testing results to better identify pavement overlay needs.

Listings of locations where other accident characteristics are abnormally high also are produced using the three-year combined accident data. Location listings have been developed for the following categories:

Maryland's Roundabout Accident Experience 2007 Update

This report evaluates the effectiveness on safety of Roundabouts in Maryland. Studies have found that one of the benefits of roundabout installations is the improvement of overall safety performance. Single-lane roundabouts have been found to perform better than two-way, all-way stop and signalized intersections. Although the frequency of crashes is not always lower at roundabouts, particularly multi-lane roundabouts, reduced injury rates have been reported.

Cover of Marylamd's Roundabouts: Accident Experience and Economic Evaluation

The first segment of this report involves general accident data (that is in our database) of all roundabouts in Maryland. This includes accident data for single-lane roundabouts; two-lane roundabouts and also data for roundabouts that have been installed at new locations where no intersection previously existed.

The second segment of this report concentrates solely on single-lane roundabouts. This segment consists of a cost effective analysis and the benefit/cost ratio analysis of those roundabouts evaluating the before construction period to the after construction period.

The analysis method utilizes the before and after approach to assess the safety affects roundabouts have on accidents. The methodology consists of data collection, comparison of data, significance testing and economic evaluation.

This report indicates the results of two techniques used in the economic evaluation. The cost effectiveness technique indicates the numbers of dollars spent to reduce a single accident. The benefit/cost technique indicates the ratio of the rate of return in safety benefits for the entire after period (Annual Benefit) as compared to the amount of dollars spent on the project taking into account the invested capital that could be recuperated over the 20-year service of the project (Equivalent Uniform Annual Cost).


Findings:

This analysis indicates that at locations in Maryland where single lane roundabouts have been installed there has been a 68% decrease in the total accident rate/million vehicles entering the intersection (acc/mve), 100% decrease in the fatal accident rate/mve, an 86% reduction in the injury accident rate/mve and a 40% reduction in the property damage only accident rate/mve.

The weighted average accident cost for a single accident (Maryland Accident Cost figures) is estimated at approximately $100,000. The results of the cost effectiveness evaluation of the 19 single-lane roundabouts indicated that for a single accident reduced, a total of $16,469 dollars was spent. The benefit/cost analysis indicated that for every dollar spent on these projects, considering the 20-year service life of the roundabouts, there is a return of approximately $13.00 to be realized through accident reduction. These calculations are based on the anticipated accident experience expected to occur had no roundabouts been installed compared to the actual after period accident experience of the roundabout locations.

 

Office of Operations FHWA Safety Home