UNITED STATES DISTRICT COURT SOUTHERN DISTRICT OF NEW YORK
May 14, 1985
PETER PERANZO, ISADORE FELIX, OSCAR ROMAN, FERDINAND FRILANDO, ROBERT LAWRENCE, MARCELLA PHIPPS, JAMES BOYD, on bahalf of themselves and all others similarly situated, Plaintiffs, against THOMAS COUGHLIN, Commissioner, New York State Department of Correctional Services, and RAMON RODRIGUEZ, Chairman of the New York State Board of Parole, GERALD M. BURKE, JOSEPH V. SALO, WILLIAM J. BARNWELL, MAURICE DEAN, THEODORE KIRKLAND, MANUEL PERRON, IRVING GREEENBERG, MARIA BUCHANAN, SAMUEL D. SHERRID, JOSEPH MULHOLLAND, BARBARA TREEN, and J. KEVIN McNIFF, Commissioners of New York State Board of Parole, in their Official Capacities, Defendants
The opinion of the court was delivered by: SAND
As part of continuing effort to deal with the troublesome problem of narcotics use within penal institutions,
New York, as well as a number of other states, has adopted a program of randomly subjecting inmates' urine to the Syva EMIT
-st drug detecting test.
The New York State Department of Correctional Services has promulgated Departmental Directive # 4937,
which sets forth the rules governing the state's operation of its urine testing program. According to the Directive, an inmate may be required to provide a urine sample to prison authorities for purposes of determining whether the inmate has used drugs or alcohol.
The urine sample is then subjected to EMIT testing in order to make this determination. The record indicates that the EMIT testing apparatus is a purely mechanical, "idiot proof" device, requiring the operator to exercise no discretion, read no graphs and make no subjective interpretations.
Tr. 171. Among other procedural requirements, the Directive provides that "If a positive result is obtained, the apparatus shall be re-calibrated and a second test shall be performed on the same sample, by a different trained individual if available." Directive # 4937, Section D(5)(a)(4). If the second EMIT test yields a negative result, the initial positive result is disregarded and no disciplinary action results. If the second test yields a positive result, "the individual performing the urinalysis testing shall cause a 'Misbehavior Report' to be written. The Misbehavior Report shall be accompanied by the 'Request for Urinalysis Test' form, the 'Urinalysis Procedure' form, and any printed documents produced by the urinalysis testing apparatus." Id.
As a consequence of a Misbehavior Report being written, a disciplinary hearing is held. At the hearing, "a positive urinalysis result may be used as evidence of the illicit use by the inmate of the drug or alcohol indicated by the result." Id., Section E. While both the state and the inmate are generally permitted to introduce certain additional inculpatory or exculpatory evidence, disciplinary sanctions may be imposed against an inmate found to have engaged in the illicit use of narcotics include confinement to one's cell ("keep lock"), transfer to a Special Housing Unit, or loss of good behavior allowance ("good time") and/or other institutional privileges. Tr. 282.
Inmates may also be denied parole release based in part on their use of narcotics as established by the results of EMIT testing.
Plaintiffs, inmates of New York prisons, have moved to preliminary enjoin defendants
from taking any disciplinary action against them based solely upon confirmed results of urinalysis tests performed with the Syva EMIT-st Drug Detection System or any other testing system utilizing the same testing method and to enjoin them further to release from disciplinary confinement any inmate being held in any Special Housing Unit or under keeplock pursuant to any disciplinary proceeding disposition which was based solely upon unconfirmed results of urinalysis tests performed with the Syva EMIT-st Drug Detection System or any other testing system utilizing the same testing method and to enjoin them further from denying parole release to any inmate based solely upon unconfirmed results of urinalysis tests performed with the Syva EMIT-st Drug Detection System or any other testing system utilizing the same testing method.
Plaintiffs' Motion for a Preliminary Injunction, at 1-2. Defendants have cross-moved to dismiss the complaint.
By Order of Reference dated February 6, 1985, these motions were referred to Magistrate Buchwald to hear and report proposed findings and recommendations. Pursuant to that reference, the Magistrate held a three-day hearing on the reliability of the EMIT test. On April 2, 1985, the Magistrate issued a report recommending the denial of defendants' motion to dismiss; the denial of plaintiffs' request that inmates being held in Special Housing Units or keeplock pursuant to any disciplinary proceeding based solely on EMIT test results be released from such confinement; and the granting of a preliminary injunction enjoining defendants from taking any future disciplinary action against inmates based solely on the results of EMIT tests which have not been confirmed by other tests, and ordering defendants to perform confirming tests (using another reliable analytical method) on positive EMIT tests before using the results as the sole basis for disciplinary action.
The matter is presently before this Court on objections filed by both parties. Pursuant to 28 U.S.C. § 636, we have reviewed de novo those portions of the report to which objection has been made. For the reasons discussed below, we adopt so much of the Magistrate's Report as denied defendnats' motion to dismiss and plaintiffs' motion for a preliminary injunction ordering the release of those inmates now in disciplinary confinement as a result of disciplinary proceedings based solely on EMIT test results. We believe, however, that plaintiffs have not sustained their burden of demonstrating an entitlement to preliminary injunctive relief as to the use of EMIT test results in future disciplinary proceedings or parole determinations and therefore deny plaintiffs' motion for such relief.
As the Magistrate correctly recognized, the critical issue in this case relates to the reliability of the EMIT test as a basis for imposing disciplinary sanctions against inmates. Before this Court can properly set about the task of resolving this issue, the precise nature, context and scope of our inquiry must be defined. The question, "How reliable is the test?" is one for scientists to resolve; the question, "Is the test sufficiently reliable such that its use as the basis for imposing disciplinary sanctions against prisoners does not offend constitutional standards?" is a legal matter to which different standards apply.
Obviously, the possibility that any person may unjustly suffer any deprivation because of a fallible scientific procedure is a cause for serious concern. We recognize that plaintiffs' interest in the accuracy of the drug testing procedures utilized by the defendants in this case is substantial. The courts have also recognized, however, that due process is not synonymous with a requirement of scientific exactitude or error-free procedures. See Greenholtz v. Nebraska Penal Inmates, 442 U.S. 1, 7, 60 L. Ed. 2d 668, 99 S. Ct. 2100 (1979) ("there simply is no constitutional guarantee that all executive decision-making must comply with standards that assure error-free determinations"); see also Mackey v. Montrym, 443 U.S. 1, 13, 61 L. Ed. 2d 321, 99 S. Ct. 2612 (1979) (same). Even with respect to one of the most fundamental of individual rights -- the right to an accurate determination of guilt at a criminal trial -- optimal precision is not the consitutional norm. See United States v. Ivic, 700 F.2d 51, 69 n.11 (2d Cir. 1983) ("It is not required that the government prove guilt beyond all possible doubt. The test is one of reasonable doubt."). Similarly, we recognize that concepts of "reliability", "fairness" and "procedural adequacy" cannot be evaluated in the abstract, but must be examined in light of the particular factual setting with which the state's action is concerned. Unlike the private interest in the accuracy of a criminal proceeding involving individuals possessed of the full panoply of liberty rights, see Ake v. Oklahoma, 470 U.S. 68, 105 S. Ct. 1087, 84 L. Ed. 2d 53 (1985), the liberty interest of prisoners, particularly in light of the already-existing restrictions on their liberty and the indisputable responsibility of prison authorities for maintaining order and discipline and deterring prisoner misconduct, is unquestionably of a different character.
The procedural safeguards to be accorded prisoners need not be coextensive with those provided for defendants in criminal proceedings. See Wolff v. McDonnell, 418 U.S. 539, 94 S. Ct. 2963, 41 L. Ed. 2d 935 (1974), ("Prison disciplinary proceedings are not part of a criminal prosecution, and the full panoply of rights due a defendant in such proceedings does not apply"). Accordingly, if due process does not require error-free determinations in matters affecting the liberty interests of ordinary citizens, then mandating such a level of accuracy would certainly overstate the requirements of due process in the context of prisoner drug-testing.
Courts have often been called upon to the role which specific tests may be permitted to play in legal proceedings. Some testing procedures have been found to be so unreliable that courts hav held the results inadmissible for most purposes. See, e.g., Barrel of Fun, Inc. v. State Farm Fire & Casualty Co., 739 F.2d 1028 (5th Cir. 1984) (psychological stress evaluation tests); United States v. Hunter, 672 F.2d 815 (10th Cir. 1982) (polygraph tests); United States v. Clark, 598 F.2d 994 (5th Cir. 1979) (same), cert. denied, 449 U.S. 1128, 67 L. Ed. 2d 116, 101 S. Ct. 949 (1981); United States v. Brown, 557 F.2d 541 (6th Cir. 1977) (ion microprobic analysis of human hair). But other tests, although not foolproof, are deemed of sufficient reliability that they may be received in evidence not only in administrative proceedings but also in criminal trials as well. See, e.g., People v. Gower, 42 N.Y.2d 117, 397 N.Y.S.2d 368, 366 N.E.2d 69 (1977) (breathlyzer test to detect use of alcohol by motorists); United States v. Distler, 671 F.2d 954 (6th Cir. 1981) (gas chromatograph analysis for matching oil samples); United States v. Stifel, 433 F.2d 431 (6th Cir. 1970) (neutron activation analysis of bomb package fragments), cert. denied, 401 U.S. 994, 28 L. Ed. 2d 531, 91 S. Ct. 1232 (1971).
The Supreme Court's recent decision in California v. Trombetta, 467 U.S. 479, 104 S. Ct. 2528, 81 L. Ed. 2d 413 (1984), is an indication of the willingness of the legal system not to mandate procedures which will assure absolute scientific accuracy in determinations of criminal culpability. In Trombetta, the Court held that law enforcement authorities are not constitutionally required under the Due Process Clause to preserve the breach sample of a suspected drunk driver in order to introduce the result of a breath-analysis test a criminal trial. The Court concluded that preservation of the breath sample would ordinarily be of little exculpatory value in light of the testing procedures used by California law enforcement authorities. The Court observed that in order to
protect suspects against machine malfunctions, the Department [of Health] has developed test procedures that include two independent measurements (which must be closely correlated for the results to be admissible) bracketed by blank runs designed to ensure that the machine is purged of alcohol traces from previous tests ... In all but a tiny fraction of cases, preserved breath samples would simply confirm the Intoxilyzer's determination that the defendant had a high level of blood-alcohol concentration at the time of the test.
Id. at 2534. The Court also observed that defendants had other means of establishing their innocence, including raising the possibility of machine malfunction, having an opportunity to inspect the machine and its weekly results as a means of impeaching its reliability, producing evidence that radio waves or chemicals in their blood interfered with the test results, and cross-examining the law enforcement officer who administered the test. Although the Court noted the possibility that an unreliable breath test might give rise to a consitutional sufficiency-of-the-evidence challenge; it stated only that "if the Intoxilyzer were truly prone to erroneous readings, then Intoxilyzer results without more might be insufficient to establish guilt beyond a reasonable doubt." Id. at 2534 n.10. Thus, while the Court did not directly address the reliability of the breath test under due process standards, its analysis nevertheless reflects a general recognition that states need not implement all possible procedural safeguards against erroneous deprivations of liberty when utilizing the results of scientific testing devices in accusatory proceedings.
We therefore turn to the basic issue of whether defendants' EMIT test procedure has been shown by plaintiffs to be so unreliable that it cannot be the "sole basis" for demonstrating narcotics use at inmate disciplinary proceedings.
As the Magistrate noted:
Unfortunately, there have been no studies subjecting the EMIT tests from the New York DOCS on-site testing program to reliability testing using an alternative confirming method. Therefore, we must look to results from other studies, which range widely.
Report, at 15.
To this end, a number of experts testified at the hearing. Plaintiffs called Dr. Z. Joseph Mule, director of the New York State Substance Abuse Services Testing and Research Laboratory of Brooklyn, New York, and the holder of a Ph.D. in biochemical pharmacology. Dr. Mule testified about a procedure followed in his lab whereby samples were subjected to EMIT testing for marijuana. If the results of the EMIT test were positive, a second EMIT test was performed on a duplicate sample. If the results again proved positive, a third test was performed using a different, newly developed test. Tr. 77. According to Dr. Mule, as a result of this confirmatory assay "we can confirm 97 percent of all EMIT positives. We have now analyzed in excess of 800 samples using this technique." Tr. 78.
Despite his conclusion that confirmatory testing demonstrated that the EMIT test technique was "extremely reliable and valid," Tr. 78, Dr. Mule suggested that he would not "go to court" based solely on the results of a single testing procedure but would require a confirmation of the results by a different testing procedure. Tr. 80, 86. Dr. Mule summed up his opinion as follows:
When you are above 95 percent or 96 percent or 97 percent, you can feel very, very confident with the technique. But, again, it you are really going to do something with that person in some form that is serious, you know, being incarcerated or have some serious freedoms being taken away from him, you eliminate that 2, 3, 4 percent whatever the case may be, by backing that test up, because the chances of seeing it here with a very good assay and then not seeing it on confirmation, are very small, if, in fact, it is truly there. So you are eliminating almost all of that.
If you did a third test or a fourth test, you could reach the sublime or the ridiculous with how many particular analytical tests you are going to subject a specimen to. But the more you do, you can reduce the probabilities to almost zero. It depends upon how far you wish to go and how deeply concerned you are with what you are going to do with your results.
Defendants called as their expert Dr. Karl Verebey, assisting director at the lab headed by Dr. Mule. Dr. Verebey testified to the virtue of performing a second EMIT test on urine samples, Tr. 172-73, 193-94, and to the recognition accorded the EMIT drug detection system by the scientific community as a valid testing procedure. Tr. 165-66. With respect to the value of performing an alternative confirming test, Dr. Verebey essentially agreed with Dr. Mule that:
[I]t's simply the question of what is at issue. The test itself is rather reliable and a duplicate run is reliable, but it depends on what are -- there are punitive measures. I definitely believe that there should be a confirmation, so it is really a judgment -- it's a judgment question, it's not a scientific question.
Tr. 193-94. Dr. Verebey was of the opinion tht (that) the EMIT test for marijuana compares favorably with the alcohol Breathalyzer test with which he is also familiar. According to Dr. Verebey, "I would rely more on the EMIT test for a positive or negative than on the actual numbers I would get from a Breathalyzer." Tr. 167.
Counterbalancing this strong showing of the reliability of double EMIT tests performed in the New York State Substance Abuse Services Testing and Research Laboratory was the testimony of Dr. David J. Greenblatt, chief of the division of Clinical Pharmacology at Tufts New England Medical Center at Boston and a professor of psychiatry and medicine at Tufts. Dr. David j. Greenblatt, chief of the division of Clinical Pharmacology at Tufts New England Medical Center at Boston and a professor of psychiatry and medicine at Tufts. Dr. Greenblatt's familiarity with EMIT testing was based on his review of the literature and some studies which his lab and conducted in collaberation with other laboratories. Tr. 217. It was his opinion that the EMIT test for marijuana "is sufficiently inaccurate and nonspecific that the results of any given test cannot be as such considered to be valid." Tr. 218. A good deal of his testimony consisted of a review of the literature together with that of the parties' other experts,
the reliability of the EMIT drug detection procedure has been established in various studies and settings as follows:
Test Location Error Sample Explanation
New Jersey Dept. of
Corrections (Tr. 238) 25% 400 Single EMIT testing only
Center for Disease survey of Source of error (analytical
Control (Tr. 222) 4% 64 labs or clerical) not determined
--- (Tr. 230) 20% 30 80% confirmation rate
attributed to urine
--- (Tr. 230) 15% 20 85% confirmation rate
attributed to borderline
EMIT values n*
Substance Abuse Ser-
vices Testing and Testing performed under
Research Laboratory 2.3% 802 laboratory conditions
n* The record also suggests that these studies involved confirmation of a single, uncorroborated EMIT positive result. Tr. 230.
Whether or not a lesser showing of reliability is acceptable under the Due Process Clause with respect to inmates than non-inmates, we conclude that plaintiffs have failed to establish a basis for preliminary injunctive relief even if the defendants are held to the more stringent standard. The worst that can be said at this point is that under some circumstances, where EMIT test results were analyzed without performing double EMIT testing, an error rate as high as 25% has been found. As Dr. Verebey recognized, the utility of double EMIT testing is that it gives the inmate the "benefit of the doubt" in situations where the positive result is close to the cut-off point.
Tr. 172-73. Furthermore, the two studies with the large sample sizes and with test procedures more analogous to those followed by the defendant indicate a margin of error 2.3% and 4%. We also note that these error, or "false positive", rates or not synonymous with -- indeed, they overstate -- the level of EMIT test unreliability. As both Dr. Mule and Dr. Greenblatt recognized, a certain portion of these percentages is attributable to conditions or circumstances largely unrelated to the accuracy of the EMIT test itself. (e.g., operator error), as well as to the inability to confirm the EMIT test's positive result in all cases despite the fact that the positive result may in fact be accurate.
Tr. 111, 114, 234-35. While little evidence was introduced by either party concerning the error rate experienced by New York's Department of Correctional Services,
the burden of showing entitlement to preliminary injunctive relief rests with plaintiffs. We believe that the above evidence falls well short of demonstrating that the EMIT testing procedure as employed by defendants is so unreliable that plaintiffs are entitled to the extraordinary relief which they presently request.
Although the parameters of "due process" are not generally articulated in mathematical terms, we note that at least one court has endeavored to assign numerical benchmarks to the various standards of proof which govern judicial proceedings. In United States v. Fatico, 458 F. Supp. 388 (E.D.N.Y. 1978), aff'd, 603 F.2d 1053 (2d Cir. 1979), cert. denied, 444 U.S. 1073, 62 L. Ed. 2d 755, 100 S. Ct. 1018 (1980), Judge Weinstein opined that the probabilities associated with the various standards of proof may be fairly estimated as 50% for preponderance of the evidence, 70% for clear and convincing evidence, above 80% for clear, unequivocal and convincing evidence, and 95% for evidence beyond a reasonable doubt. Id. at 403-06. While the comparison of burden of proof percentages to drug test reliability rates is not an analytically precise one, application of the Fatico guidelines to the instant case suggests that the EMIT testing system produces results whose level of accuracy is somewhere between the levels of certainty achieved under the clear, unequivocal and convincing standard and the beyond a reasonable doubt standard. Indeed, the two studies involving the largest sample sizes place the EMIT test at a level of certainty even higher than the reasonable doubt standard. This rough comparison confirms the conclusion we have reached under more traditional legal analysis that the reliability of the EMIT system as operated by defendants has not, at this stage of the proceedings, been shown to be so inadequate that preliminary injunctive relief is appropriate.
Other courts have recognized that the use of the EMIT system as a basis for imposing disciplinary sanctions for inmate drug use is not necessarily impermissible. In Jensen v. Lick, 589 F. Supp. 35 (D.N.D. 1984), the court held that the North Dakota prison authorities' use of the EMIT system to detect inmate drug use was constitutionally permissible. Under the testing procedures followed by the state's prison authorities, inmates were referred for disciplinary action on the basis of a positive EMIT test; no alternative testing procedure was used to confirm the EMIT test result. The court noted that the Center for Disease Control in Atlanta had found that EMIT testing procedure to be 97% to 99% accurate, and concluded that the test's "almost complete certainty" constituted an adequate level of reliability for purposes of imposing disciplinary sanctions on prisoners. Id. at 38-39.
In Smith v. State, 250 Ga. 438, 298 S.E.2d 482 (Ga. 1983), the court held that the results of the EMIT test were admissible at a probation revocation hearing involved a probationer's alleged use of controlled substance. The court stated that as a general evidentiary matter, it was "proper for the trial judge to decide whether the procedure or technique in question has reached a scientific stage of verifiable certainty" such that it was competent evidence in a judicial proceeding. The court upheld the trial court's determination, made after hearing expert testimony concerning the operation and accuracy of the EMIT test system, that the EMIT test results were admissible as evidence of narcotics use.
In Orr v. Kuhlman, Mot. No. 46, slip op. (N.Y. County Ct. Oct. 19, 1984) (annexed as Ex. 9 to Defendants' Cross-Motions to Dismiss the Complaint), the court rejected a claim that the state's failure to provide an inmate with an opportunity to have a confirming test performed on his urine same at his own expense violated the Fourteenth Amendment. The court found that the results yielded by New York prison authorities' urine test procedure provided "a reasonable basis for [the state's] determination that [the inmate's] urine sample contained traces of marijuana." Id. at 4. In so holding, the court placed emphasis on the fact that double testing was performed in determining whether an inmate had used narcotics. Id.
Other courts have exhibited considerably less tolerance for the imprecision inherent in EMIT testing. In Johnson v. Walton, No. S61-84 Rm (Vt. Super. Ct. Feb. 14, 1985), the court was confronted with a constitutional challenge to Vermont's inmate drug testing program. The court noted that the thin-layer chromatography (TLC) testing procedure used by Vermont prison authorities had an 86% to 94% rate of reliability, and that the EMIT procedure also used by state prison authorities had a 97% to 100% reliability rate. The court found that when one test was used to confirm the results of the other, "a positive result indicates presence of the drug under investigation with scientific certainty that goes beyond a reasonable doubt." Id. at 8. The court then held that the state's failure to confirm positive test results violated due process. Id. at 9. As noted above, we believe there is a considerable gap between a finding that a particular testing procedure assures scientifically verifiable exactitude beyond that which is necessary to convict an individual of criminal charges, and a determination that the failure to provide this level of accuracy in the context of prisoner drug testing violates due process.
In light of the insufficiency of the present showing of unreliability of double EMIT testing and the dearth of information regarding New York's experience with its EMIT testing procedures, preliminary injunctive relief is unwarranted.
This is not to say that the use of additional safeguards would not serve a worthwhile purpose. While some states, such as North Dakota and Georgia, allow EMIT test results alone to be utilized in inmate disciplinary or probation revocation hearings, other states, such as New Jersey, Massachusetts and Vermont, now utilize (or have been required to use) a confirming test in such circumstances. See Tr. 237-38; see also 28 C.F.R. § 550.30 (1984) (Federal Bureau of Prisons requirement that positive urine test be "validated to substantiate the positive result"). Indeed, even Syva, the EMIT test manufacturer, recommends that positive results be confirmed by any of several chromatographic procedures. See Defendants' Ex. 6, at 4.
The inclusion of this recommendation, however, by no means renders the unconfirmed use of the EMIT test an unreliable procedure to follow. One can well appreciate why a manufacturer in Syva's position -- aware of the desirable but elusive goal of 100% reliability in test performance, unaware of the particular use to be made of, or the procedures in fact to be followed in using, the equipment it sells, and perhaps concerned with product liability exposure even where the margin of testing error is relatively small -- would temper its otherwise enthusiastic endorsement of the test's accuracy with such a precautionary recommendation. In fact, the manufacturer's operating instructions themselves expressly contemplate conducting two EMIT tests, as defendants have done, as a means of achieving an added degree of assurance from the EMIT test system itself. We recognize and respect the belief of the scientific community that additional confirming tests should be performed in order to achieve the highest possible degree of reliability. Where the risk of erroneous determinations of drug use has not been shown to be significant, however, the benefit of using a confirmatory testing procedure is necessarily more limited. On the record before us, the risk that double EMIT testing will result in erroneous deprivations of inmates' liberty interest has not been shown to be so significant that additional testing procedures can be required at this stage of the proceedings as a matter of due process, regardless of the incremental benefits that additional procedures might yield or the burdens which such procedures may impose.
We agree with the Magistrate, however, that the complaint should not be dismissed but that the matter should proceed to a trial on the merits. Cf. Storms v. Coughlin, 600 F. Supp. 1214 (S.D.N.Y. 1984) (evidence concerning lack of reliability of EMIT testing for inmate drug use not strong enough to justify preliminary injunctive relief but raised issue of substance sufficient to withstand motion to dismiss). While the present showing of EMIT test unreliability is insufficient for preliminary injunction purposes, the allegations of plaintiffs' complaint, if substantiated at trial, give rise to a cognizable constitutional claim and, at a minimum, raise issues of real concern. We would hope that before such s trial takes place, our knowledge of the degree of reliability of EMIT testing as conducted by the defendants would be increased.
We would hope as well that the defendants would give serious consideration to whether, as a matter of policy and prudence, confirmatory testing should be performed at least in those circumstances where there is reason to believe that a misbehavior report will impact severely on an inmate.
Insofar as the Magistrate's Report denies defendants' motion to dismiss and denies plaintiffs' application for retroactive injunctive relief, it is adopted and approved. Plaintiffs' motion for preliminary injunctive relief in all other respect is denied.
The parties are directed to appear before this Court on Thursday, May 23, 1985, at 2:15 P.M., for oral argument on plaintiffs' motion for class certification.