Field Test of On-Site
Drug Detection Devices
Final Report -- October 2000

IV. Results

 

Study Participants

The research team collected demographic data for 783 of the 800 individuals who participated in the study. These data are presented in Table 3. The overwhelming majority (93%) of the individuals who agreed to participate were male. This percentage differed slightly by site with females accounting for 4% of the participants in Nassau County and 6% in Houston. The percentage of males in the sample was comparable to the percentage of males in the arrest data from both Nassau County and Houston, although the percentage of males was somewhat higher in the study sample. The average age of the participants was 32 years and ranged from 17 years of age to 72 years of age. The mean age and age range were nearly identical in Nassau County and Houston and matched the 1999 arrest data in Nassau County. (Arrest data from Houston were not available). Caucasian participants comprised 45% of the study participants with Hispanic participants comprising 40%, African Americans 11% and Asians 2%. For the remaining 2% of participants, race information was not obtained. Expected differences were found between Nassau County and Houston with regard to Hispanic participants. In Houston, Hispanic participants comprised 57% of the sample (with Caucasians comprising 26%). In Nassau County, Hispanics accounted for 24% of the participants, with Caucasians comprising 64%.

Table 3
Demographic Characteristics of Study Participants

  Nassau County Houston Total
Mean Age 32 32 32
Gender (%*)
Male 94% 92% 93%
Female 6% 4% 5%
Unknown 0% 4% 2%
Ethnicity (%*)
White 64% 26% 45%
African American 11% 11% 11%
Latino/Hispanic 24% 57% 40%
Asian 1% 3% 2%
Other <1% 0% <1%
Unknown 0% 5% 2%

* Total may not add to 100% due to rounding

On-Site Device Results and MS Confirmations

In total, 288 of the 800 cases (36%) tested positive for at least one of the five drugs on at least one of the on-site devices. MS confirmations were conducted on 322 cases (40%). These cases included the random sample of thirty-four cases that tested negative on all five devices. Data are summarized by drug and reported as false positive results, false negative results, and unconfirmed positive results. There are a number of ways to calculate the percentages of false positive and false negative results. We have adopted the conventional approach of calculating the percentages by using the entire sample of 800 cases as the denominator for each drug by each device. Percents are truncated to two decimal places.

Definitions. A false positive was recorded when the device indicated a positive result, but no drug(s) or metabolites were detected in the MS confirmation at a sufficient concentration to explain the result (for example, the result from Device A was positive for PCP, but the MS result for PCP was negative).

However, defining this category for the drug classes other than PCP required several considerations. The on-site devices are immunoassay-based tests with antibodies designed to react to a specific drug metabolite (e.g., morphine) within a drug class (e.g., opiates). The target metabolite for PCP is PCP, for cocaine is benzoylecgonine (BZE) and the target metabolite for marijuana is THC-COOH. With the analysis of opiates and amphetamines, there are other drugs within these drug classes that have similar chemical structures and the antibodies in the on-site devices have cross reactivity to these analogs. This reactivity varies with the antibody, analog, and with the devices. The extent to which these related drugs contribute to the response of the on-site device depends on the specific cross reactivity of the antibody in that device. The results presented below describe the false positive percentages for those cases in which the devices indicated a positive result for a drug class (e.g., amphetamines), but no drugs or metabolites were detected in the confirmation.

False negative results were assigned to those results where the device tested negative, but the sample contained drug concentrations greater than or equal to the device screening cutoff (for example, Device B result was negative for cocaine, but the MS result was greater than or equal to 300 ng/mL). We categorized the false negative samples into two groups: 1) those in which the MS confirmation concentration was greater than the screening cutoff and 2) those in which the MS confirmation was greater than the DOT/DHHS confirmation cutoff. Category 1 is the most important. As noted earlier, all of the devices are designed to screen for the presence of the target drugs at the DOT/DHHS screening cutoff criteria. The second criterion, the DOT/DHHS MS confirmation, is the cutoff at which the MS confirms the presence of specific drug analytes. For some drugs, these concentrations are lower than the screening concentrations. These cutoff concentrations can be found in Table 4.

The term Unconfirmed Positive is used to describe those cases where the device result was positive, but the concentration of target drug(s) or metabolites in the urine, as determined by MS, was below the DOT/DHHS confirmation cutoff. For example, the result for Device C was positive for PCP, but the MS result for PCP was 10 ng/mL. The drug was present in the urine, but the concentration was less than the DOT/DHHS MS confirmation cutoff of 25 ng/mL.

Table 4
Screening and MS Cutoff Concentrations

  On-Site Device and
DOT/DHHS Screening Cutoff

ng/mL
DOT/DHHS
MS Confirmation criteria

ng/mL

CANNABINOIDS

   
THC-COOH 50 15

COCAINE

300  
Cocaine
BZE
  150
150
AMPHETAMINES 1000  
Amphetamine
Methamphetamine
  500
500
OPIATES* 300  
Morphine
Codeine
  300
300
PHENCYLIDINE 25  
PCP   25

*Recently changed to 2,000 ng/mL screen and confirmation (DHHS, 1997).

False Positive Results

Table 5 presents the false positive percentages by device based on the total number of tests conducted (n=800) for that drug with each device. As indicated above, these data include only those cases in which the device's positive result was explained by the presence of a sufficient concentration of one or more of the drugs or metabolites described in the MS confirmation methods (see Appendix B: Confirmation Testing Procedures).

Table 5
On-Site Device False Positive Results* as % of All Samples Analyzed (n=800)

  THC-COOH Cocaine/BZE Amphetamines
(**)
Opiates
(***)
PCP
Triage® 0.00 0.00 1.75 0.12 0.62
TesTcup® 0.00 0.00 0.25 0.25 0.75
AccuSign® 0.25 0.12 0.25 0.25 1.50
Rapid Drug Screen® 0.12 0.12 0.25 0.25 0.75
TesTstik® 0.00 0.00 0.12 0.25 0.62

*Analytes not detected by MS
** % Adjusted for the presence of
methylenedioxymethamphetamine (MDMA/"Ecstasy")
*** % Adjusted for the presence of hydromorphone/hydrocodone

THC-COOH. Two hundred and eight samples were submitted for confirmation THC-COOH testing. One hundred and seventy-two of these samples tested positive on one or more of the on-site devices for THC-COOH. The remaining samples were negative samples. Only two of the 172 samples that tested positive on one or more on-site devices were negative by MS. Both samples were false positive results with AccuSign® (0.25%). One of the two samples (0.12%) was a false positive result with Rapid Drug Screen®.

Cocaine. One hundred and sixty-three samples were submitted for MS cocaine and metabolite (BZE) testing. One hundred and twenty-four of the samples tested positive on one or more of the on-site devices for BZE and the remaining were negative samples. Only one of the 163 samples that tested positive by the devices was negative by MS. That sample produced a false positive result with both AccuSign® (0.12%) and Rapid Drug Screen® (0.12%).

Amphetamines. Seventy-seven samples were submitted for MS analysis of amphetamines. Thirty-nine of the samples tested positive on one or more of the on-site devices for amphetamines and the remaining samples were negative samples. Of the thirty-nine samples that tested positive using the on-site devices, only six had MS measurable concentrations of amphetamine, methamphetamine or phentermine (the target analytes). The false positive rates, adjusted for the potential effects of MDMA (see Unconfirmed Positives below), were as follows: Triage® 1.75%; TesTcup® 0.25%; AccuSign® 0.25%; Rapid Drug Screen® 0.25%; and TesTstik® 0.12%.

Opiates. Seventy-seven samples were submitted for MS analysis of opiates. Thirty-eight of the samples tested positive on one or more of the on-site devices for opiates and the remaining were negative sample challenges. Of the thirty-eight samples that tested positive, only nineteen had measurable concentrations of total morphine or codeine by MS. However, hydrocodone and hydromorphone (see Unconfirmed Positives below) were found in nearly all the samples. The opiate false positive rates (adjusted for the presence of hydromorphone and hydrocodone) were as follows: Triage® 0.12%; TesTcup® 0.25%; AccuSign® 0.25%; Rapid Drug Screen® 0.25%; and TesTstik® 0.25%.

PCP. Seventy-five samples were submitted for MS analysis of PCP. Thirty-eight of the samples tested positive on one or more of the on-site devices and the remaining samples were negative samples. Of the thirty-eight samples that tested positive, twenty-three contained measurable concentrations of PCP by MS. The false positive rates were as follows: Triage® 0.62% (n = 5); TesTcup® 0.75% (n = 6); AccuSign® 1.50% (n = 12); Rapid Drug Screen® 0.75% (n = 6); and TesTstik® 0.62% (n = 5). Four samples tested positive with all of the devices and a fifth sample tested positive with four of the five devices.

False Negative Results

False negative results were assigned to those results where the device tested negative, but the sample contained the target drug at a concentration above the device screening cutoff. This is an important category of errant results because it represents the DUI arrestees who had significant concentrations of drugs in their urine, but who would not have been identified using the device results. In addition, these are cases that may not have been prosecuted for driving under the influence of drugs because of the results. Table 6 presents two sets of the false negative results. The first set of results is based on the DOT/DHHS screening cutoff concentrations and the second set is based on the DOT/DHHS MS confirmation cutoff. There were no false negatives using any device for the thirty-four samples that tested negative on all five tests for all five drugs. False negative results were obtained, however, on those samples that tested negative on some, but not all, of the devices for a given drug.

Table 6

On-Site Device False Negative Results as % of All Samples Analyzed (n=800)
Drug Present in Concentrations Greater than the Screening Cutoff
  THC-COOH Cocaine/BZE Amphetamines Opiates PCP
Triage® 0.25 0.25 0.00 0.00 0.00
TesTcup-5® 0.25 0.00 0.00 0.12 0.25
AccuSign® 0.12 0.12 0.00 0.00 0.00
Rapid Drug Screen® 0.37 0.12 0.00 0.00 0.00
TesTstik® 0.25 0.12 0.00 0.00 0.00
On-Site Device False Negative Results as % of All Samples Analyzed (n=800)
Drug Present in Concentrations Greater than the MS Confirmation Cutoff
  THC-COOH Cocaine/BZE Amphetamines Opiates PCP
Triage® 0.50 0.37 0.00 0.00 0.00
TesTcup-5® 0.87 0.25 0.00 0.12 0.25
AccuSign® 0.25 0.12 0.00 0.00 0.00
Rapid Drug Screen® 0.37 0.12 0.00 0.00 0.12
TesTstik® 0.75 0.25 0.00 0.00 0.00

THC-COOH. A number of samples that tested negative on one or more devices contained more than 50 ng/mL of THC-COOH by MS. Using this screening criterion, the false negative rates were as follows: Triage® 0.25% (n = 2); TesTcup® 0.25% (n = 2); AccuSign® 0.12% (n = 1); Rapid Drug Screen® 0.37% (n = 3); and TesTstik® 0.25% (n = 2). Additional samples tested negative on one or more devices that contained greater than 15 ng/mL of THC-COOH by MS. Using this DOT/DHHS MS confirmation criterion, the false negative rates were as follows: Triage® 0.50% (n = 4); TesTcup® 0.87% (n = 7); AccuSign® 0.25% (n = 2); Rapid Drug Screen® 0.37% (n = 3); and TesTstik® 0.75% (n = 6). There did not appear to be a pattern for the samples that were not detected by the devices. Concentrations varied from 17 ng/mL to 130 ng/mL and no sample in the false negative category tested negative on all five of the devices.

Cocaine. Using the criterion that a false negative is any sample that tested negative with a device, but contained more than 300 ng/mL (the screening cutoff) of BZE (the target metabolite) by MS, the false negative rates were as follows: Triage® 0.25% (n = 2); TesTcup® 0.00% (n = 0); AccuSign® 0.12% (n = 1); Rapid Drug Screen® 0.12% (n = 1); and TesTstik® 0.12% (n = 1). These rates were adjusted to reflect the following cross reactivities to cocaine (that is, there were cases in which the sample contained sufficient concentrations of the parent drug cocaine in addition to BZE to trigger a positive response, but did not): Triage® 40%; TesTcup® 4%; AccuSign® 60%; Rapid Drug Screen® 28%; and TesTstik® 4%. Additional samples tested negative on one or more devices that contained more than 150 ng/mL (the MS confirmation cutoff) of BZE (or equivalent) by MS. Using this DOT/DHHS criterion, the false negative rates were as follows: Triage® 0.37% (n = 3); TesTcup® 0.25% (n = 2); AccuSign® 0.12% (n = 1); Rapid Drug Screen® 0.12% (n = 1); and TesTstik® 0.25% (n = 2). There were too few false negative results to determine if there was a pattern in the discrepancies.

Amphetamines. There were no false negative amphetamine results among the thirty-four samples that tested negative on all five devices. Nor were there any false negative results using the DOT/DHHS screening or MS confirmation criteria for samples that obtained mixed results from the devices for amphetamines.

Opiates. One sample tested negative using TesTcup® (0.12%) that was a false negative result using either the DOT/DHHS screening or MS confirmation criteria for reporting a sample positive for opiates. No false negative errors were found with the other four devices.

PCP. Two samples tested negative using TesTcup® (0.25%) and one with Rapid Drug Screen® (0.12%) that were false negative results using either the DOT/DHHS screening or MS confirmation criteria for reporting a sample positive for PCP. No false negative errors were reported from the other devices.

Unconfirmed Positive Results -- Based On Analytical Cutoff

A major consideration in evaluating the accuracy and reliability of the devices was the selection of a comparison standard. Numerous studies have been published to assess the accuracy and reliability of on-site drug screening test devices (Buchan, 1997; Crouch, 1997; 1998; Ferrara, 1994; Hwang, 1994; Koch, 1994; Towt, 1995). The basic design of many of these studies was similar to the study reported here. On-site test results were compared to test results obtained from one or more alternate methods. However, usually these caparisons were made based on laboratory results using the DOT/DHHS testing guidelines.

In this section we present the results of this study based on those guidelines - specifically the MS confirmation concentration guidelines. Unconfirmed positive results were those results in which the device(s) tested positive, but compared to the DOT and DHHS confirmation guidelines, would have confirmed negative because the MS drug concentration was less than the higher confirmation standard (DOT and DHHS). This is an important category of samples because it represents those DUI arrestees who had drugs in their urine, but who would have been reported as drug free using these widely accepted standards. It also is the category to which many of the "False Positive" results in previous studies should have been assigned. Table 7 presents the percentage of unconfirmed positive results by device based on the DOT/DHHS confirmation criteria.

Table 7
Unconfirmed Positive Results as % of All Samples Analyzed (n=800)

Drug Present by MS Below the DOT/DHHS Confirmation Criteria

  THC-COOH Cocaine/BZE Amphetamines Opiates PCP
Triage® 1.12 0.50 0.12 0.37 0.25
TesTcup® 1.12 0.50 0.12 0.25 0.25
AccuSign® 1.00 1.37 0.12 0.37 0.37
Rapid Drug Screen® 1.00 1.37 0.12 0.37 0.37
TesTstik® 1.12 0.50 0.12 0.37 0.25

THC-COOH. Of the hundred and seventy-two samples that tested positive on one or more of the on-site devices for THC-COOH, only two were negative by MS. However, several had MS concentrations of THC-COOH that were less than the 15 ng/mL DOT/DHHS MS confirmation standard. The unconfirmed positive results rates were as follows: Triage® 1.12% (n = 9); TesTcup® 1.12% (n = 9); AccuSign® 1.00% (n = 8); Rapid Drug Screen® 1.00% (n = 8); and TesTstik® 1.12% (n = 9). There were seven samples that produced a positive result with all of the devices. The mean THC-COOH concentration of these seven samples was 10.33 ng/mL. Only one contained less than 10 ng/mL of THC-COOH. Lowering the current cutoff by 1/2 to 7.5 ng/mL for THC-COOH would result in the following unconfirmed positive results: Triage® 0.25% (n = 2); TesTcup® 0.25% (n = 2); AccuSign® 0.25% (n = 2); Rapid Drug Screen® 0.12% (n = 1); and TesTstik® 0.12% (n = 1).

Cocaine. Of the hundred and twenty-four samples that tested positive on one or more of the on-site devices for BZE, only one was negative by MS. However, fourteen had MS concentrations of BZE less than the 150 ng/mL DOT/DHHS MS confirmation standard. The unconfirmed positive rates were as follows: Triage® 0.50% (n = 4); TesTcup® 0.50% (n = 4); AccuSign® 1.37% (n = 11); Rapid Drug Screen® 1.37% (n = 11); and TesTstik® 0.50% (n = 4). There were three samples that produced positive results on five devices. The mean BZE concentration of these unconfirmed positive samples was 86.57 ng/mL. Lowering the current cutoff by 1/2 to 75 ng/mL for BZE would result in the following unconfirmed positive rates: Triage® 0.00% (n = 0); TesTcup® 0.00% (n = 0); AccuSign® 0.75% (n = 6); Rapid Drug Screen® 0.75% (n = 6); and TesTstik® 0.00% (n = 0).

Amphetamines. Of the thirty-nine samples that tested positive on one or more of the on-site devices for amphetamines, only six had MS measurable concentrations of amphetamine, methamphetamine or phentermine. One sample produced a positive result on all of the devices, had an MS concentration of amphetamine that was less than the 500 ng/mL DOT/DHHS MS standard, and could be classified as an unconfirmed positive result. The unconfirmed positive rate for all of the devices was 0.12% (n = 1). An additional sample contained more than 1,000 ng/mL of phentermine, tested positive on all of the devices, and could be classified as an unconfirmed positive result because DOT and DHHS do not include phentermine as a drug for confirmation. Only Rapid Drug Screen® publishes a device cross reactivity to phentermine (1%).

Opiates. Of the thirty-eight samples that tested positive for opiates using the on-site devices, nineteen contained measurable concentrations of morphine or codeine by MS. However, three of these nineteen samples had MS concentrations of morphine and/or codeine that were lower than the 300 ng/mL DOT and DHHS MS standard and were classified as unconfirmed positive results. The unconfirmed positive rates were as follows: Triage® 0.37% (n = 3); TesTcup® 0.25% (n = 2); AccuSign® 0.37% (n = 3); Rapid Drug Screen® 0.37% (n = 3); and TesTstik® 0.37% (n = 3). One sample tested positive on all of the devices and contained 115 ng/mL of codeine. Two additional samples did not test uniformly positive on the devices, contained no codeine, and had 32 and 277 ng/mL, respectively, of morphine.

PCP. Of the thirty-eight samples that tested positive on the on-site devices, twenty-three had measurable concentration of PCP by MS and three had concentrations of PCP that were less than the 25 ng/mL DOT and DHHS standard. The unconfirmed positive rates were as follows: Triage® 0.25% (n=2); TesTcup® 0.25% (n=2); AccuSign® 0.37% (n=3); Rapid Drug Screen® 0.37 (n=3); and TesTstik® 0.25% (n=2). Because the screen and MS confirmation concentrations are both 25 ng/mL for PCP, there is no obvious explanation for the errors.

Unconfirmed Positive Results -- Based On Tested Drugs

As discussed above, a major consideration in evaluating the accuracy and reliability of the devices was the selection of a comparison standard. Usually, on-site test results have been compared to laboratory results using the DOT/DHHS testing guidelines. In this section, we present the results of the current study based on the drugs tested in those guidelines. These unconfirmed positive results were those results in which the device(s) tested positive, but compared to the DOT and DHHS confirmation guidelines, would have been reported negative because the sample did not contain detectable concentrations of a DOT/DHHS MS target drug. Again, this is an important category of samples because these are arrestees who had drugs in their urine, but would have been reported as drug free using these widely accepted standards. This is the category to which many of the "False Positive" results in previous studies should have been assigned. Table 8 presents the percentage of unconfirmed positive results based on tested drugs by device.

Table 8
Unconfirmed Positive Results as % of All Samples Analyzed (n=800)
Based on Target Drugs

  THC-COOH Cocaine/BZE Amphetamines Opiates PCP
Triage® 0.00 0.00 3.75 2.25 0.62
TesTcup® 0.00 0.00 2.25 2.37 0.75
AccuSign® 0.25 0.12 2.25 2.37 1.50
Rapid Drug Screen® 0.12 0.12 2.50 2.37 0.75
TesTstik® 0.00 0.00 2.12 2.25 0.62

THC-COOH, Cocaine, and PCP. Because the target analyte for the devices and the MS confirmation for these drugs are essentially the same, no difference was observed between the unconfirmed positive results based on tested drugs and the false positive rate presented earlier.

Amphetamines. Thirty-nine of the samples tested positive on one or more of the on-site devices for amphetamines and the remaining samples were negative samples. As indicated above, of the thirty-nine samples that tested positive using the on-site devices, only six had MS measurable concentrations of amphetamine, methamphetamine or phentermine (the target analytes). However, sixteen of the thirty samples that resulted in drug positive finding with the Triage® test contained methylenedioxymethamphetamine (MDMA), an illicit drug known by the street name "Ecstasy." Sixteen of eighteen TesTcup® and AccuSign® amphetamine positive samples contained MDMA, seventeen of nineteen Rapid Drug Screen® positive samples contained MDMA, and sixteen of seventeen TesTstik® positive samples contained MDMA. The cross reactivity of the devices to MDMA was as follows: Triage®, 30%; TesTcup®, 50%; AccuSign®, 14%; Rapid Drug Screen®, 30%; and TesTstik® 25%. (This cross reactivity indicates the concentration of MDMA necessary in the sample for each device to cause a positive response). For example, the cross reactivity of TesTcup® to MDMA was 50%. The screening cutoff concentration for amphetamine/ methamphetamine for all devices is 1000 ng/mL. For MDMA to cause a positive response with the TesTcup® device, a concentration of at least 2000 ng/mL of MDMA (twice that of the amphetamine cutoff) would be necessary. That is, TesTcup® is 50% as sensitive to MDMA as it is to the target analyte (amphetamine or methamphetamine). MDMA was only qualitatively identified in the samples. Therefore, it is not possible to definitively predict whether the MDMA concentrations were sufficient to produce a positive result with each device. However, the sixteen samples that contained MDMA tested positive on all of the devices and, in many of the urine samples where it was identified, the MS response of MDMA exceeded the response of the 1,000 ng/mL methamphetamine standard. Therefore, the presence of MDMA was the likely explanation for the positive result. Without taking into consideration these cross-reactivities (that is, positive on-site results which would not have been confirmed by MS using the standard confirmation methods), the unconfirmed positive rates for the target analytes (amphetamine, methamphetamine, and phentermine) were as follows: Triage® 3.75% (n = 30); TesTcup® 2.25% (n = 18); AccuSign® 2.25% (n = 18); Rapid Drug Screen® 2.50% (n = 20); and TesTstik® 2.12% (n = 17). As discussed above, the false positive rates reported earlier, adjusted for the potential effects of MDMA, were as follows: Triage® 1.75%; TesTcup® 0.25%; AccuSign® 0.25%; Rapid Drug Screen® 0.25%; and TesTstik® 0.12%.

Two additional Triage® false positive samples contained phenylpropanolamine, pseudoephedrine, or ephedrine. The cross reactivities of the devices to these over-the-counter drugs was quite low and not a likely explanation for a positive test result. However, this explanation cannot be ruled out as only qualitative MS analyses were performed for these drugs.

Opiates. Seventy-seven samples were submitted for MS analysis of opiates. Thirty-eight of the samples tested positive on one or more of the on-site devices for opiates and the remaining were negative sample challenges. Of the thirty-eight samples that tested positive, only nineteen had measurable concentrations of total morphine or codeine by MS. However, all but one of the Triage® and all but two of the TesTcup®, AccuSign®, Rapid Drug Screen®, and TesTstik® positive opiate samples contained hydromorphone and hydrocodone at concentrations greater than 200 ng/mL. Hydromorphone (available as a prescription pain medication Dilaudid®) is a metabolite of hydrocodone. Hydrocodone is also a prescription pain medication available as Vicadin®, Lortab®, Lorset®, and others. The cross reactivity of the devices to hydromorphone was: Triage®, 75%; TesTcup®, 43%; AccuSign®, 50%; Rapid Drug Screen®, not published; and TesTstik®, 43%. The cross reactivity of the devices to hydrocodone was as follows: Triage®, not published; TesTcup®, 60%; AccuSign®, 60%; Rapid Drug Screen®, 6%; and TesTstik®, 38%. These cross reactivities indicate the concentration of hydrocodone or hydromorphone needed to produce a positive response for each device. The screening cutoff for morphine (the target analyte for opiates) for all devices was 300 ng/mL. For hydromorphone to cause a positive response from the AccuSign® device, for example, a concentration of at least 600 ng/mL would be necessary because AccuSign® had a 50% cross reactivity to hydrocodone. Both hydrocodone and hydromorphone were present in all of the samples. Therefore, the immunoresponse of each sample was equal to the combined effects of both drugs. Also, in many of the urine samples the estimated drug concentrations, based on their MS responses relative to morphine and codeine, were quite high. Further evidence that the positive on-site device results were attributable to hydromorphone and hydrocodone is found in the data. All the positive opiate samples that contained hydromorphone and/or hydrocodone were positive on all of the devices - with only two exceptions. If the presence of hydrocodone and hydromorphone are not taken into consideration, the unconfirmed positive rates based on tested drugs (target analytes) rates were as follows: Triage® 2.25% (n = 18); TesTcup® 2.37% (n = 19); AccuSign® 2.37% (n = 19); Rapid Drug Screen® 2.37% (n = 19); and TesTstik® 2.25% (n = 18).

Comparison of Officer Test Results and Research Analyst Test Results

Results from the tests conducted by the officers were compared to results from the tests conducted by the research analysts to determine whether the devices performed differently when handled by the officers. Table 9 presents the results of those comparisons. Of the 4,000 tests performed with the devices used (5 devices x 800 samples), there were forty-seven tests in which one of the devices resulted in an error (i.e., a result that was not confirmed by MS) and the other four devices performed accurately. Twenty-seven of those errors were the result of the research analysts performing the test and twenty of the errors were the result of the officers performing the tests. When the number of tests performed by the officers and the research analysts is taken into consideration, the error rate for the research analysts was 0.8% (27/3200). The error for the officers was 2.5% (20/800). One sample accounted for four of the errors made by the officers (i.e., the officer incorrectly recorded the results for four of the five drugs for one device). If that sample is removed from the analysis of the errors, the error rate for the officers drops to 2.0%. More dramatic differences were found in comparisons across individual devices. As Table 9 indicates, officers had considerably higher percentages of errors for AccuSign®, Rapid Drug Screen®, and Triage® than the research analysts.

Table 9
Comparison of Research Analyst and Officer Errors by Device

  Research Analysts
(n=3200)
Officers
(n=800)
Triage® 1.7%
(11/640)
3.1%
(5/160)
TesTcup-5® 0.9%
(6/640)
0.6%
(1/160)
AccuSign® 0.8%
(5/640)
3.8%
(6/160)
Rapid Drug Screen® 0.5%
(3/640)
4.4%
(7/160)
TesTstik® 0.3%
(2/640)
0.6%
(1/160)
TOTAL 0.8%
(27/3200)
2.5%
(20/800)

Drug Positive Rate

Although this study was not a prevalence study (such a study would have required much more stringent sampling procedures), a substantial number (36%) of the 800 drivers who were arrested for suspected DUI screened positive for one or more drugs in their urine using the on-site devices. This rate is considerably higher than rates found in previous studies. The results of the on-site devices by drug are presented in Table 10.

Table 10
Percentage of DUI Suspects Found Positive by On-Site Devices*

  THC COC PCP AMP MOR
Nassau County 27.3%
(n=109)
15.3%
(n=61)
3.5%
(n=14)
8.8%
(n=35)
4.3%
(n=17)
Houston 16.0%
(n=64)
15.8%
(n=63)
5.8%
(n=23)
1.8%
(n=7)
5.5%
(n=22)
Total 21.6%
(n=173)
15.5%
(n=124)
4.6%
(n=37)
5.25
(n=42)
4.9
(n=39)

* Not a prevalence rate. The drug positive counts present a drug positive rate greater than the overall rate reported above, but some of the samples tested positive for more than one drug - these are only counted once for the purposes of the total drug positive rate.

Although THC showed the highest drug positive rate, drug use patterns differed in each site. In Nassau County, the drug positive rate for THC (27.3%) was substantially higher than the drug positive rate for any other drug - cocaine had the next highest positive rate with 15.3%. By contrast, in Houston, the drug positive rates for THC (16.0%) and cocaine (15.3%) were similar. In addition, the drug positive rate for amphetamines in Nassau County (8.8%), as indicated by the devices, was nearly five times as high as in Houston (1.8).

The overall drug positive rate was reduced when the MS confirmation data were considered. Of the 288 samples that tested positive on one or more of the devices, 233 were confirmed by MS using the DOT/DHHS guidelines for confirmations (standard confirmation cutoff concentrations, restricted drug class, etc.) for a confirmed drug positive rate of 29%. In many cases, the samples tested positive for more than one drug. However, as described above, many of the samples that tested positive on the drug testing devices obtained negative confirmations because the drug concentrations were below the DOT/DHHS cutoffs or the samples tested positive for drugs not included in the standard MS panel (e.g., MDMA, hydrocodone, hydromorphone, etc.). In the cases of THC and cocaine for example, 24 of the 28 samples that were negative by MS contained measurable drug concentrations. The drug positive rate that includes samples testing positive below the DOT/DHHS cutoff concentrations or for drugs not commonly tested during MS confirmations was 33%.

DRE Evaluations

In addition to the data from the on-site devices and the MS confirmations, we also collected data, when available, from the DRE evaluations. Officers conducted DRE evaluations in forty-one cases. Results from the DRE evaluations and on-site devices were compared to assess how the devices might be used in conjunction with the DRE evaluations. The results are based on a limited number of DRE evaluations and are not designed to assess the accuracy of the DRE evaluations. We expected to find discrepancies between the DRE evaluations and the on-site devices for a number of reasons. First, the DRE evaluations are designed to assess driving impairment from drugs while the on-site devices detect the presence of the drug or metabolite in urine. Given the number of days a drug metabolite may be detected in urine, it is entirely possible that the on-site devices could detect the presence of a drug and DRE evaluation indicate no evidence of impairment from that drug at the time of arrest because drug metabolites may be detected in urine days or weeks after use. In addition, decisions made by law enforcement officers to charge a suspect with driving under the influence of a drug are not made casually and officers are likely to be extremely cautious about making those decisions. Finally, it is also possible that a suspect may show impairment from a drug prior to the drug metabolite appearing in the urine and being detected by the devices.

THC-COOH. In twenty-two of the forty-one cases in which the DRE evaluation was conducted, the officer indicated that the arrestee was driving under the influence of THC. In twenty of those cases (91%), the on-site devices and MS tests confirmed those findings. In two cases, the DRE findings were not consistent with the on-site devices or MS confirmation. However, in one of those cases, the DRE evaluation also indicated that the arrestee was driving under the influence of cocaine. Both the devices and MS test confirmed this finding. There were also six additional cases in which the devices and MS confirmation indicated the presence of THC, which was not indicated on the DRE evaluation. In one of those cases, the DRE evaluation indicated the presence of PCP, which was confirmed by both the devices and the MS confirmation. These results suggest that, although the THC metabolite was present in urine at the time of arrest, no driving impairment was apparent from that drug. This finding is due largely to the fact that drug metabolites may be found in urine days (or sometimes weeks) after the drug is taken (when it is no longer a factor in impairment).

Cocaine. Officers noted the presence of cocaine in nine cases, eight (89%) of which were consistent with the on-site devices and confirmed by MS tests. In the one case that was not confirmed by MS, the presence of cocaine below the DHHS cutoff concentration was indicated and Triage® also indicated a drug positive for cocaine. An additional six cases resulted in drug positive findings on both the devices and MS confirmation but negative DRE evaluations. However, in four of those cases the DRE evaluation indicated the presence of other drugs, which were also identified by the devices and MS confirmation.

Amphetamines. Two cases resulted in positive findings for amphetamines on the DRE evaluation. In neither case, however, did the devices or MS confirmation indicate the presence of amphetamines. In one of those cases, the DRE evaluation indicated the presence of PCP as well, which was confirmed by the devices and the MS test. In the second case, the devices and MS confirmation also noted the presence of cocaine, although the DRE evaluation for that drug was negative.

Opiates. There were five cases that resulted in positive DRE evaluations for opiates. Two of those cases were confirmed by the devices and MS tests. In two additional cases, the devices indicated the presence of opiates and the MS tests confirmed the presence of hydrocodone and hydromorphone. DRE evaluations resulted in negative findings for opiates in three cases in which the devices and the MS confirmations indicated the presence of opiates. In two of those cases, the DRE evaluation resulted in positive findings for THC, which were also confirmed by the devices and MS confirmation.

PCP. DRE evaluations noted the presence of PCP in five cases, all of which were confirmed by the devices and MS confirmation. There were five additional cases in which the DRE indicated a negative finding for PCP and the devices and MS confirmations indicated positive results. In three of those cases, the DRE evaluation indicated the presence of THC, which was confirmed by the devices and MS.

Officer Ratings and Interviews

As indicated in the Methods section, the law enforcement officers who participated in the field test rated each device. These ratings were not intended to make comparative assessments of the specific devices, but rather to provide information on the officer's subjective judgment of the device as it was being used. The ratings were not meant to indicate that one or more devices were superior. Indeed, as indicated below, one of the devices rated most favorably by the officers was also one of the devices that produced the greatest percentage of errors for the officers.

The results of the ratings can be found in Table 11. AccuSign® received the highest rating across each element followed by Rapid Drug Screen®, TesTcup®, TesTstik®, and Triage®. During the follow-up interviews with a sample of the officers, they indicated that the devices that required the least time and urine handling received more favorable ratings.

Table 11
Officer Ratings
(5 Point Scale: 1=Least Favorable, 5=Most Favorable)

  Ease of
Use
Time
Needed
Need for
Specimen
Handling
Readability
of Results
Usefulness
of Devices
for Testing
Total
Triage® 2.34 1.72 2.42 3.01 2.43 2.38
TesTcup-5® 3.05 2.91 2.70 3.28 3.14 3.02
AccuSign® 3.86 3.73 3.61 3.72 3.77 3.74
Rapid Drug Screen® 3.34 3.08 2.71 3.25 3.22 3.12
TesTstik® 3.04 2.59 2.57 3.16 2.89 2.85

The seventeen officers who participated in the follow-up interviews were also asked to comment generally on the use of the drug testing devices. The majority of officers believed that the devices could be used routinely in the traffic division and all of the officers interviewed believed there were benefits to using the devices. The most frequently noted benefits were that the devices could be used to provide quick results so that the officer has some indication if there were drugs in the arrestee's system and that the results could help reinforce the decisions the officers make in the field. Many of the officers, including DRE-trained and non-DRE-trained officers, felt the devices were ideal supplemental tools for the DRE program. One of the possible drawbacks mentioned by the officers was that some officers might rely too much on the device results and not enough on the examination designed to determine impairment. Other drawbacks included: 1) officers' discomfort with handling urine; 2) the limited number of drug classes detected by the devices; and 3) questions about the accuracy of the device.

Back -- Table of Contents -- Forward