[Federal Register: January 14, 1998 (Volume 63, Number 9)] [Rules and Regulations] [Page 2172-2177] From the Federal Register Online via GPO Access [wais.access.gpo.gov] [DOCID:fr14ja98-19] ----------------------------------------------------------------------- DEPARTMENT OF TRANSPORTATION Federal Highway Administration 49 CFR Part 382 Controlled Substances and Alcohol Testing Management Information System (MIS) Statistical Data AGENCY: Federal Highway Administration (FHWA), DOT. ACTION: Controlled substances and alcohol testing rates. ----------------------------------------------------------------------- SUMMARY: The FHWA is announcing the motor carrier industry's 1994, 1995, and 1996 controlled substances and alcohol testing positive rates. The controlled substances testing positive rate was 2.6 percent in calendar year 1994, 2.8 percent in 1995, and 2.2 percent in 1996. The alcohol testing ``violation'' rate was 0.14 percent in 1995, and 0.18 percent in 1996. Because the violation rate was below 0.5 percent for two consecutive years, the FHWA is announcing it is lowering the random alcohol testing rate for calendar year 1998 to 10 percent, in accordance with the provisions of the testing regulations. DATES: January 1, 1998. FOR FURTHER INFORMATION CONTACT: Mr. Charles Rombro, Office of Motor Carrier Information Analysis (HIA-20), (202) 366-5615; Federal Highway Administration, 400 Seventh Street, SW., Washington, DC 20590. SUPPLEMENTARY INFORMATION: On December 23, 1993 (58 FR 68220), the FHWA announced it would require motor carriers subject to 49 CFR part 391, later replaced by part 382, to implement and maintain specific controlled substance testing data, and submit an appropriate annual report when requested. All motor carriers must maintain this information. The FHWA randomly selects a sample of motor carriers annually and asks those selected to submit their data. On February 15, 1994 (59 FR 7484), the FHWA promulgated new controlled substances and alcohol testing rules in 49 CFR part 382. These rules combined the controlled substances annual report with a similar alcohol rule ``violation'' annual report. An alcohol rule violation for purposes of the annual report are alcohol concentrations of 0.04 or greater and refusals to submit to alcohol testing. On March 13, 1995, the FHWA amended the rule to reduce the information collection burden on all respondents, including small entities (60 FR 13369). The current rule at Sec. 382.403, formerly at 49 CFR 391.87(h), is essential for the FHWA to accomplish the following four goals. 1. Collect controlled substance and alcohol testing statistical data. 2. Use the data to analyze its current approach to deterring and detecting illegal controlled substance abuse and alcohol misuse in the motor carrier industry. 3. Determine each calendar year's random selection rates for alcohol and controlled substance testing under the rule. 4. Provide for a more efficient and effective regulatory program. In 1995, the FHWA requested a sample of motor carriers report to the FHWA data collected in 1994. The FHWA determined the random positive controlled substance testing rate for commercial motor vehicle (CMV) drivers subject to 49 CFR part 391, subpart H, for the period of January 1, 1994, through December 31, 1994, was 2.6 percent. In 1996, the FHWA requested a sample of motor carriers report to the FHWA data collected in 1995. The FHWA calculated a random positive controlled substance testing rate--``the positive rate''--for 1995 of 2.8 percent. The 1995 random alcohol violation testing rate--``the violation rate''--for CDL drivers of motor carriers with 50 or more CDL drivers was 0.14 percent. The FHWA performed similar calculations on the 1996 data, based on forms received from a random sample of carriers in early 1997. The ``positive rate'' for controlled substances was 2.2 percent. For alcohol, the ``violation rate'' was 0.18 percent. The estimated rates, with their associated 95 percent confidence intervals, are presented in Table 1. The estimation procedures for the 1995 and 1996 rates incorporate carrier data on drivers refusing to test. Specific definitions for violation rate and positive rate were added to the regulations during calendar year 1994 for testing in 1995. Based on these rule changes, refusals to take a random test are counted as ``positive.'' Adding refusals slightly increases the rates for alcohol concentrations of 0.04 or greater and verified positive controlled substances tests, because the number of refusals each year were very small. See the definitions for ``violation rate'' added on February 15, 1994 (59 FR 7484) and ``positive rate'' added on December 2, 1994 (59 FR 62218). This results in a higher rate than would be the case if the FHWA excluded refusals. Table 1.--Random Testing ---------------------------------------------------------------------------------------------------------------- Calendar Survey rate 95% confidence year (%) (i nterval (%)) ---------------------------------------------------------------------------------------------------------------- Controlled substances.............................................. 1994 2.6 1.9 *1995 2.8 0.9 *1996 2.2 0.8 Alcohol............................................................ *1995 0.14 0.04 [[Page 2173]] *1996 0.18 0.06 ---------------------------------------------------------------------------------------------------------------- * Includes drivers refusing to submit to testing. The FHWA based the calendar year 1994 rates upon U.S. domiciled motor carriers subject to 49 CFR part 391, subpart H operating in interstate commerce. The FHWA based the calendar year 1995 rates upon a sample from the following population: 1. Small U.S. domiciled interstate motor carriers subject to 49 CFR part 391, subpart H; and 2. Large U.S. domiciled intrastate or interstate motor carriers (also known as employers) subject to 49 CFR part 382. All U.S. domiciled motor carriers became subject to 49 CFR part 382 on January 1, 1996. As a result, the 1996 rates are based on a sample of all domestic carriers. All foreign domiciled motor carriers (e.g., Canadian and Mexican motor carriers) became subject to 49 CFR part 382 on July 1, 1997. Based upon 49 CFR 382.305 and the results of the survey, the controlled substances selection and testing rate must remain at 50 percent of the average number of CDL driver positions for calendar year 1998. The controlled substances positive rate must be below 1.0 percent each year for two consecutive years before the regulations would allow the FHWA to reduce the selection and testing rate to 25 percent. The alcohol selection and testing rate will be lowered to 10 percent of the average number of CDL driver positions for calendar year 1998, in accordance with the provisions of 49 CFR 382.305(d)(1). The FHWA may increase the alcohol selection and testing rate for calendar year 1999 if the MIS reports received for calendar year 1997 indicate the alcohol testing violation rate for 1997 is over 0.5 percent. The provisions of 49 CFR 382.305(d)(1) are as follows: Sec. 382.305 Random testing. (d)(1) When the minimum annual percentage rate for random alcohol testing is 25 percent or more, the FHWA Administrator may lower this rate to 10 percent of all driver positions if the FHWA Administrator determines the data received under the reporting requirements of Sec. 382.403 for two consecutive calendar years indicate the violation rate is less than 0.5 percent. Petition The Truckload Carriers Association (TCA, formerly named the Interstate Truckload Carriers Conference) petitioned the FHWA Administrator on October 23, 1996, to reduce the random testing rate applicable for alcohol and controlled substances testing on behalf of itself and four of its member motor carriers: C.R. England & Sons, Inc., Salt Lake City, Utah; Fortune Transportation, Windom, Minnesota; Roberson Transportation Services, Farmer City, Illinois; and Weinrich Truckline, Inc., Hinton, Iowa. The FHWA, as stated above, is reducing the random alcohol testing rate for calendar year 1998 for all motor carriers subject to 49 CFR part 382. Thus, this notice partially responds to the TCA petition. The other part requests a reduction in the random controlled substances testing rate for only the TCA's named and unnamed members. As the DOT and the FHWA explained on December 2, 1994 (59 FR 62218, at 62222), the FHWA will determine changes to the random testing rate based upon the motor carrier industry as a whole as the triggering group, not individual motor carriers or industry segments. The DOT and the FHWA believed then, and continue to believe, this is the fairest and most effective approach. Basing the testing rate upon the industry- wide positive rate provides an incentive for motor carriers and groups of motor carriers with successful programs to pressure problem carriers and subgroups to improve their performance. The FHWA, therefore, denies the TCA's petition with respect to reducing the controlled substances testing for TCA members only. How Did FHWA Determine These Rates? The appendices to this notice contain the methods used to analyze the data and calculate the positive testing rates. The FHWA has attempted to make the discussion as straightforward as possible while not slighting any substantive issues or formulas. The FHWA believes a discussion of the process must be available to the widest available audience; therefore the statistical methodology and the rates are provided below. Will the Reduction of the Random Alcohol Testing Rate Reduce Highway Safety? The FHWA does not believe reducing the random alcohol testing rate to 10 percent will diminish the safe operation of commercial motor vehicles. First, the rates show the motor carrier industry already has a very low alcohol violation rate. The FHWA's MIS data indicates the motor carrier industry is achieving about 99.8 percent alcohol free drivers. Second, recent analysis of fatal crash results show CMV drivers involved in fatal crashes do not have a significant problem with alcohol while driving CMVs. According to the National Highway Traffic Safety Administration's (NHTSA) publication, Traffic Safety Facts 1996: Large Trucks, 1.4 percent of truck drivers driving large trucks involved in fatal crashes in 1996 were intoxicated. Intoxication rates for drivers of passenger cars, light trucks, and motorcycles were 18.8 percent, 21.9 percent, and 30.3 percent, respectively. Rates for these other drivers are 13 to 21 times higher than the intoxication rate for drivers of large trucks. Third, the percentage of intoxicated truck drivers involved in fatal crashes has fallen over the last ten years. The NHTSA's data for 1987 showed 2.7 percent of truck drivers involved in fatal crashes were intoxicated, compared to 1.4 percent in 1996, a 52 percent decline. This suggests to the FHWA that truck drivers use of alcohol may be falling. See Appendix F for a graph showing the data. Finally, the industry is promoting alcohol free driving through educational material distributed by its associations. For example, please refer to material such as the article ``Road Rage,'' in the Owner Operator Independent Driver Association's ``Land Line'' Magazine, November/December 1997 issue. The article suggests ten ``simple rules of safe driving'' including the following rule. ``Don't drive when fatigued or under the influence of alcohol or drugs. Many over the counter medicines can contain alcohol or other ingredients that can affect your driving skills.'' This article is also available from OOIDA's internet universal resource locator at http:\\www.ooida.com and http:\\www.landlinemag.com [[Page 2174]] Also refer to material available from the American Trucking Associations, Inc. (ATA). ATA states one of its safety initatives is alcohol and drug testing regulations. Some of ATA's material is available on the internet, at http:\\www.truckline.com. Will the Reduction of the Random Alcohol Testing Rate Reduce Any Burdens in Collecting Information? Yes, the FHWA will be sending a revised burden estimate to the Office of Management and Budget (OMB) reflecting the savings. On September 22, 1997, the OMB reapproved the collections of information contained in the FHWA's controlled substances and alcohol testing regulations. See FHWA Docket FHWA-97-2313. The OMB extended the expiration date of its new approval until September 30, 2000. The OMB approved an annual time burden to collect information of 57,479,400 hours and an annual financial burden to collect information of $3,004,913,000. The costs to actually conduct the testing are much higher than these information collection burdens. Based upon the FHWA's reduction of the random alcohol testing rate to ten percent of the average driver positions, the FHWA will submit to OMB a request to lower the burden estimates to 57,169,400 hours and $2,974,170,000. This is a savings of 310,000 burden hours and $14,743,000 of financial burden each year to collect this information. Appendix A--FHWA Calculations for the 1994 Positive Testing Rate In 1995, the FHWA selected a stratified random sample of 1,048 motor carriers. Each selected motor carrier was mailed a request to prepare and submit to the FHWA its calendar year 1994 controlled substance testing data (the FHWA's alcohol testing regulations had not yet taken effect). Five hundred forty-seven motor carriers responded to the request. The strata were defined by the number of power units owned by each motor carrier reported on the MCMIS. The sample size and response rates, by strata, are shown in Table 2. Table 2.--1994 Initial Responses Received ------------------------------------------------------------------------ No. Strata (No. of power units) randomly No. sampled responding ------------------------------------------------------------------------ 1-19.............................................. 204 105 20-99............................................. 153 127 100-999........................................... 102 86 1000 or more...................................... 126 106 Unspecified *..................................... 459 123 --------------------- Total:.......................................... 1048 * * 547 ------------------------------------------------------------------------ * The unspecified stratum includes motor carriers that have not provided the FHWA's MCMIS with information about the number of CMV drivers who work for them. * * 551 responses were recevied, but 4 motor carrier responses failed to report sufficient information to classify the motor carriers in any stratum upon receipt of the report. Nonrespondents Four hundred ninety-seven of the 1,048 motor carriers contacted did not respond. The FHWA made a number of efforts to contact these nonrespondents. A second mailing and telephone contacts revealed the vast majority of nonrespondents had legitimate reasons for not returning the MIS forms. Many carriers were no longer in business, many did not employ CDL drivers, some were exempt farmers, and others only operated intrastate. Those nonresponding carriers that were determined to be in business and within the scope of the survey were recontacted. Drug and alcohol testing data from these carriers were then used to adjust the rates in each sampling stratum. The adjustments are described in Appendix D. Estimation of the Random Positive Rate The estimate of the rate of controlled substance use within the motor carrier industry was carried out using two steps. First, the initial estimates of the rate of controlled substance use were made for each sampling stratum. Second, the estimates were adjusted for nonresponse. All information in the sample of nonrespondents showed zero positive rates in that group. This means the estimates in the respondent groups were adjusted downward by the rate of response. (See Appendix D.) Results Estimates of the rate of controlled substance use were made for the industry as a whole and for the individual strata. The estimated rate of controlled substance use in the motor carrier industry is (with a 95 percent confidence interval (CI)): r=.026 (2.6%) 95% CI: (0.7%, 4.5%) Appendix B--FHWA Calculations for the 1995 and 1996 Controlled Substances Positive and Alcohol Violation Rates I. Disposition of the 1995 Sample For the 1995 survey, MIS forms were mailed to 1,747 motor carriers, selected from a stratified random sample. Two hundred sixty-five carriers required to conduct alcohol tests provided usable alcohol data and 648 carriers provided usable controlled substance data. Information on the disposition of the sample, by sampling stratum, is provided in Table 3. Table 3.-- Sample Sizes and Level of Response by Size Class ---------------------------------------------------------------------------------------------------------------- Respondents Number of Viable Respondents providing Size (CDL drivers) carriers on Total mail- sample providing controlled sampling outs units alcohol substance frame data data ---------------------------------------------------------------------------------------------------------------- Extra Large (1000 +)........................... 87 87 83 77 78 Large (100-999)................................ 953 150 145 126 137 Medium (50-99)................................. 1,029 100 95 62 82 Small (20-49).................................. 2,982 100 90 (NA) 80 Very Small (1-19).............................. 49,859 410 325 (NA) 152 Size Unknown................................... 99,069 900 592 (NA) 119 ---------------------------------------------------------------- Total...................................... 153,979 1,747 1,330 265 648 ---------------------------------------------------------------------------------------------------------------- [[Page 2175]] Also listed in this table is response rate information by size class stratum. The term Viable Sample Units refers to the number of solicited sample units found to be in-business and in-scope (e.g., a motor carrier with trucks over 11,794 kilograms gross vehicle weight rating and at least one CDL driver). As can be seen from the table, the FHWA found a large number of carriers selected for sample solicitation from the MCMIS sampling frame to be non-viable. This was particularly problematic for carriers selected from the ``size unknown'' size class stratum, where only 66 percent of the original sample units selected were found to be viable. For the 1995 survey, intrastate carriers in the small size classes (fewer than 50 CDL drivers) were not required to submit alcohol testing information. Hence, the sample estimate for the alcohol usage rate is based upon data from the three largest size classes only (50 or more CDL drivers). In addition, data were excluded from the ``size unknown'' size class stratum for this estimate, since the FHWA believes most carriers in this stratum are small companies. Dividing the total number of sample units providing useable data by the total number of viable units in the sample gives the response rate for the survey. Using this approach for the sampling strata where data were required, the response rate for the 1995 survey is 82 percent for alcohol testing (based upon the three largest size classes where reporting was required) and 49 percent for controlled substance testing (based upon all size classes). II. 1995 Survey Results The survey estimate for the percentage of CDL drivers testing positive for controlled substances is 2.8 percent, with an estimated standard error of 0.46 percent. Based upon these results, a 95 percent confidence interval for this estimate ranges from 1.9 percent to 3.7 percent (0.028, plus or minus the quantity 0.0046 times 2). Based on statistical theory, if the survey were to be replicated, it would be expected that the estimate for the percentage of CDL drivers testing positive would fall within this range in 95 out of 100 surveys. For the alcohol use, the survey estimate for the percentage of CDL drivers testing 0.04 or greater alcohol concentration is 0.14 percent. The estimated standard error for this estimate is 0.00018, thus giving a 95 percent confidence interval for the estimate of 0.10 percent to 0.18 percent. III. Disposition of the 1996 Sample For the 1996 survey, MIS forms were mailed to almost 3,400 motor carriers, selected from a stratified random sample. Two thousand seven hundred carriers responded, of which 1,140 provided usable data. The alcohol use rate from the 1996 survey includes data from both small and large carriers, as all carriers were required to conduct random tests in 1996. Information on the disposition of the sample, by sampling stratum, is provided in Table 4. Table 4.--Sample Sizes and Level of Response by Size Class ---------------------------------------------------------------------------------------------------------------- Number of carriers on Total mail Viable Respondents Size (CDL drivers) sampling outs sample providing frame units data ---------------------------------------------------------------------------------------------------------------- Extra Large (1000 +)........................................ 96 96 93 81 Large (100-999)............................................. 1,324 300 293 258 Medium (50-99).............................................. 1,402 200 194 162 Small (20-49)............................................... 3,857 200 189 155 Very Small (1-19)........................................... 71,074 800 636 240 Size Unknown................................................ 103,036 1,800 1,296 246 --------------------------------------------------- Total................................................... 180,789 3,396 2,701 1,142 ---------------------------------------------------------------------------------------------------------------- IV. 1996 Survey Results The survey estimate for the percentage of CDL drivers testing positive for controlled substances is 2.2 percent, with an estimated standard error of 0.4 percent. Based upon these results, a 95 percent confidence interval for this estimate ranges from 1.5 percent to 3.1 percent (0.022, plus or minus the quantity 0.004 times 2). Thus, if the survey were to be replicated, statistical theory would suggest that the estimate for the percentage of CDL drivers testing positive would fall within this range in 95 out of 100 surveys. For alcohol use, the survey estimate for the percentage of CDL drivers testing 0.04 or greater alcohol concentration is 0.18 percent. The estimated standard error for this estimate is 0.0003, thus giving a 95 percent confidence interval for the estimate of 0.12 percent to 0.24 percent. Appendix C--Methodology for Estimating the Controlled Substances Positive and Alcohol Violation Rates The annual positive rate and its sampling error must be estimated for the motor carrier industry. The estimation of the positive rate and the sampling error must be based upon the samples of annual test results. These estimates are based on a two-stage sampling design in which the primary sampling units (motor carriers) are stratified on the number of secondaries (i.e., CDL drivers) in each primary unit. This stratification procedure helps to increase the precision of the survey estimates. When stratification is combined with simple subsampling and each stratum is subsampled independently, an unbiased estimate of the overall mean per secondary is given by a weighted mean (for L strata) of the individual rates obtained for each stratum: [GRAPHIC] [TIFF OMITTED] TR14JA98.000 where the summation is taken over the L strata and p = overall mean per secondary (positive rate) Mi = number of secondaries (i.e., CDL drivers) in the ith stratum M = total number of CDL drivers in the motor carrier industry pi = the positive rate in the ith stratum Estimation of the Sampling Variance [GRAPHIC] [TIFF OMITTED] TR14JA98.001 where Var(Pi) = is the sampling variance of the positive rates in each stratum Mi = number of secondaries (i.e., CDL drivers) in the ith stratum [[Page 2176]] M = total number of CDL drivers in the motor carrier industry The estimation of the sampling variance term, var (Pi), is given as: [GRAPHIC] [TIFF OMITTED] TR14JA98.002 where N=the total number of motor carriers in a stratum n=the number of motor carriers in the sample in the stratum f1=n/N fi=the fraction of CDL drivers sampled in motor carrier i M=average number of CDL drivers for all motor carriers in a stratum Mi=the number of CDL drivers in the ith motor carrier selected pi=the positive rate in the ith motor carrier Appendix D--Estimating Stratum Rates With a Sample of Nonrespondents When a sample of nonrespondents is taken, an unbiased estimate of the positive rate is given as: p=tpr+(1-t)pnr where pr=the positive rate among the respondents pnr=the positive rate among the nonrespondents t=the estimated rate of response in the population (1-t=the estimated rate of nonresponse) The variance of this estimator is: Var=(1-)[S\2\+(g-1)(1-t)So2]/n where f=the sampling fraction of the original sample S\2\=the variance of p in the population So2=the variance of p among the nonrespondents g=the inverse of the fraction on nonrespondent sampled t=response rate n=size of the original sample Appendix E--References for the Testing MIS Calculations Cochran, William G. (1977). ``Sampling Theory,'' Third edition, John Wiley & Sons, Inc., New York. Jessen, R.J. (1978). ``Statistical Survey Techniques,'' John Wiley & Sons, Inc., New York. Johnson, N.L. and S. Kotz (1969). ``Discrete Distributions,'' John Wiley & Sons, Inc., New York. Lessler, J.T. and W.D. Kalsbeek (1992). ``Nonsampling Error in Surveys,'' Wiley-Interscience, New York. Szameitat, K. and K. Schaeffer (1963). ``Imperfect Frames in Statistics and the Consequences for Their Use in Sampling,'' Bulletin of the International Statistical Institute (40) 517-544. Appendix F This appendix appears at the end of this document. (Authority: 49 U.S.C. 322, 31136, 31301 et seq.; 49 CFR 1.48) Issued on: December 29, 1997. Kenneth R. Wykle, Federal Highway Administrator. The following Appendix F contains a chart illustrating: Blood Alcohol Concentration (grams per decileter) of Drivers of Large Trucks Involved in Fatal Crashes, 1982-1996. BILLING CODE 4910-22-P [[Page 2177]] [GRAPHIC] [TIFF OMITTED] TR14JA98.003 [FR Doc. 98-566 Filed 1-13-98; 8:45 am] BILLING CODE 4910-22-C