A risk-analysis jurisprudence concedes that there are special cases in which the need to protect society outweighs an individual's right to traditional constitutional protections.[409] Although courts have always recognized that there are exceptions to constitutional protections,[410] there has not been a rigorous effort to define the conditions that trigger an exception. This tension is very apparent in Schall and Barefoot, in which it was argued that the methodologies used to predict future dangerousness were so flawed as to be meaningless.[411] The Supreme Court only increased this tension by rejecting substantive due process arguments out of hand, relying on the adversarial process to "sort the wheat from the chaff"'.[412]

If the prevention cases share a common jurisprudential basis with traditional public health cases, they also share an analogous risk calculus. The parameters used to determine the risk of a communicable disease may be transformed into a formulation for ascertaining a detainee's risk of future dangerousness. If we call the risk factor for a communicable disease (R), then (R) is determined by the severity (S) of the disease, its transmissibility (T), and the number of persons at risk (P): (R) = (S)(T)(P). Severity is a complex term that integrates the duration *388 of the disease with the mortality and morbidity.[413] Transmissibility is a measure of the ease with which the disease may spread from individual to individual. Transmissibility varies greatly for different diseases.[414] The population at risk is determined by individual susceptibility, patterns of high risk behavior, and demographic[415] and geographic constraints on transmission.

The higher (R), the greater the threat to citizens and to the stability of the community. In a constitutional context, the balance between individual liberty interests and societal protection will shift away from individual rights with increasing (R). When (R) is very low, such as for the common cold, there is no rational relationship between detention or other intrusive measures and disease control. When (R) is very high, as for pan-drug- resistant tuberculosis, then the rationale for intrusive state action,[416] including permanent isolation, is compelling.[417]

*389 A parallel risk analysis in prevention cases might allow a rigorous evaluation of individual detentions, focusing on factors similar to those relevant to traditional public health risk analysis:

1) the threat to society posed by the behavior in question;[418]

2) the probability that the defendant would engage in the dangerous behavior;[419]

3) the number of persons at risk of engaging in the dangerous behavior.[420]

The resulting risk calculus might have factors of (R) for the risk to society, (T) for the threat posed by the activity under consideration, (D) for the probability of dangerousness, and (N) for the number of possible offenders: (R) = (T)(D)(N).

The greater (R), the greater the state's right to protect itself from the activity in question. Protective measures would range from minimal regulation,[421] to systematic regulation,[422] to prohibition,[423] to prospective actions to prevent possible future activity.[424] More significantly, as (R) decreases, the importance of substantive due process increases.[425]

For example, in Salerno the Appeals Court postulated the threatened destruction of an airliner by terrorists.[426] Intuitively, (R) would be very high for this scenario. The high risk involved would provide the state with a sufficient basis to detain the suspected terrorist with only summary proceedings. Given these factual assumptions, the risk of *390 an improper detention would be outweighed by the risk of the airliner's destruction.

Modifying the facts shifts the risks dramatically. If the concern is that a student organization has been infiltrated by extremists who might incite the students to violent action, then (T) and (D) will be critical. Because the nature of the violent action is unknown, it will be difficult to determine its threat and thus (T) will be low. Since there is also great uncertainty about the probability of any action, (D) will be small. Thus, the risk associated with this activity is insignificant, (T)(D)(N) is small, and detention would require a heavy burden of proof.

Although risk calculus is not meant to provide an algorithmic model for judging individual liberty deprivations, it is a useful tool for discussing prevention questions in a more rigorous manner than is allowed by traditional constitutional formulations. Because risk analysis splits complex problems into simple components, a risk analysis may also help in the recognition of polycentric problems that defy single component solutions.[427]

Courts and legislatures have often based their actions on an implicit risk analysis.[428] The problem with these implicit analyses is that their underlying assumptions are not open to scrutiny and challenge.[429] An explicit analytical risk analysis is most sorely needed when courts and legislature attempt to make scientifically valid legal policy, because although scientific research can provide valuable insights on the nature of social problems, it seldom provides unambiguous answers that can be executed through statutory rules. A science-based legal policy must rely on expert decision-makers to apply knowledge to specific factual situations; a legislatively or judicially imposed risk analysis would serve to *391 limit the discretion of these experts.[430] The alternative is to have jurors evaluate the expert's decisions based on no more than the personal credibility of the expert. As illustrated in Barefoot, a "beyond a reasonable doubt"' standard is little protection if the underlying substantive issue is improperly framed.[431]

Explicit risk analysis may seem to undermine constitutional protections. Yet the Supreme Court has always balanced the individual's right to liberty against the state's right to protect itself. From the "clear and present danger"' limitation on free speech, to restrictions on free exercise of religion by military personnel, the Court has shown time and again that there are no absolutely protected freedoms. Explicit risk analysis would give greater rigor to vague, legalistic formulations such as "strict scrutiny"' and "reasonable relationship"'.

The analogy between detention to prevent crime and isolation to control disease is not perfect, but the parallel is more than a surface resemblance. A systematic attempt should be made to apply risk analysis techniques that have been developed for disease spreading behavior to the larger universe of socially destructive behavior. Risk analysis would provide a basis for more humane imposition of individual restriction than is practiced by the Supreme Court today.

[409] The notion of "traditional"' constitutional protection is, of course, a fairly slippery concept. Most criminal law constitutional protections are of recent origin, as are civil rights protections. It might be argued that the line between traditional protections and the historical lack of constitutional protections from state imposed restrictions lies somewhere between 1945 and 1960.

[410] See, e.g., Ex parte McCardle, 74 U.S. 506 (1868) (war and insurrection); Carlson v. Landson, 342 U.S. 524 (1952) (immigration and naturalization); Korematsu v. United States, 323 U.S. 214 (1944) (detention of American citizens of Japanese ancestry during wartime upheld).

[411] See also Schall, 513 F. Supp. at 707; Barefoot, 463 U.S. at 896-97. Subsequent research on the 31 named defendants in Schall determined that the 24 who could be located had continued to engage in criminal activities. See KRAMER, AT A TENDER AGE 242 (1988).

[412] Barefoot, 463 U.S. at 901 n.7.

[413] For example, the common cold is a disease with a very low severity because it has a short duration, causes only moderate discomfort and has a very low mortality. HIV infection is a very severe disease because it is untreatable, its duration is not self-limited and it causes significant morbidity and ultimately death in all infected persons. The severity of tuberculosis varies depending on the susceptibility of the organism to the available treatments and the resistance of the patient to the disease. Most cases of tuberculosis are treatable and cause only limited symptoms. There are strains of tuberculosis that are resistant to all available antitubercular drugs. These strains cause a disease that is disabling and frequently fatal.

[414] For example, the common cold is highly infectious and is spread through respiratory secretions. This results in a highly transmissible disease. Gonorrhea is highly infectious, but is only spread through sexual contact. Thus, gonorrhea is less transmissible than a cold but is still sufficiently transmissible to have an incidence of 3 million cases a year. Tuberculosis is spread through respiratory secretions but, unlike cold and gonorrhea, is not highly contagious. Leprosy is extremely difficult to transmit, requiring both close personal contact with an infected person and a defective immune system.

[415] Gonorrhea is a demographically determined disease because it is confined to persons who are sexually active.

[416] This is implicitly limited to the universe of actions that have medical rationales. It would not include punitive measures such as forced labor or torture.

[417] As diseases such as HIV bring public health issues before the courts again, it should be possible systematically to evaluate proposed restrictions by following this risk calculus. Following this form, a public health risk brief should include:

1) an analysis of the severity of the disease, including its treatability, mortality, morbidity, and special factors (e.g., whether the disease causes a congenital syndrome in the children of infected pregnant women) that would influence a risk assessment;

2) an analysis of the transmissibility of the disease, including the component of voluntary action involved in transmission and the probability of an effective contact (i.e., a contact that transmits the disease) during various activities;

3) a description of the population at risk, including both its demographics (i.e., the number of persons at risk and their identifying characteristics, such as age, sex, ethnicity, etc.) and special characteristics of subpopulations at risk.

This format would aid public health officials in evaluating the constitutionality of proposed disease control measures. It would also allow courts to better separate factual medical issues from political and policy issues. For example, Utah and California might agree on the medical risk of gonorrhea, but disagree on the extent of individual restriction that the risk would justify.

[418] For example, a violent sexual offender would pose a greater risk than a chronic writer of hot checks.

[419] While this is certainly a difficult determination, there is evidence that for certain crimes, such as sexual abuse, offenders tend to be recidivists. See the excellent discussion of preventive detention in Alschuler, Preventive Pretrial Dentention and the Failure of Interest- Balancing Approach to Due Process, 85 MICH. L. REV. 510, 537-51 (1986).

[420] This is the most constitutionally troubling criterion. But it is necessary if we are to pose rational solutions to problems that affect grossly different sized populations: one might "detain"' every suspected psychopathic serial killer, but that remedy that would be unworkable for suspected illegal drug users.

[421] Publishing would fall into this category. The First Amendment embodies the idea that publishing is a benefit to society. The state is limited to prohibiting narrowly described activities such as child pornography.

[422] For example, the right to fly an airplane is heavily regulated because of the risk to society of incompetent pilots.

[423] This is the domain of criminal law.

[424] This is the traditional domain of public health jurisprudence.

[425] As Barefoot illustrates, due process and strict standards of proof offer scant protection when the underlying issue to be proved is as insubstantial as an arbitrary probability of future dangerousness.

[426] Salerno, 794 F.2d at 74.

[427] The best example may be drug use. Drug laws treat opiates (heroin, morphine, codeine, etc.) the same way as stimulants (cocaine, amphetamine, etc.). Pharmacologically and psychologically they are profoundly different. Even a heavy opiate user can lead a productive life and not pose a threat to society. Opiate-associated crime is driven by the unavailability of the drug, rather than the pharmacology of the drug. Conversely, heavy stimulant usage leads to an agitated, paranoid psychosis. This psychosis is incompatible with everyday responsibilities and leads to antisocial actions, irrespective of access to the drug. By ignoring these differing risk profiles, the United States has allowed illegal drug usage to shift from potentially manageable substances to unmanageable stimulants.

[428] See American Textile Mfrs. Inst. v. Donovan, 452 U.S. 490, 514 (1980).

[429] The Supreme Court used an implicit risk analysis in Florida v. Riley, 109 S. Ct. 693, 697 (1989) to extend the "plain view"' exception of the Fourth Amendment to anything visible to a police officer in a helicopter. In his dissent, Justice Brennan criticized this implicit analysis because it ignores the value of privacy to non-criminals: "By casting its "risk analysis"' solely in terms of the expectations and risks that 'wrongdoers' or 'one contemplating illegal activities' ought to bear, the plurality opinion, I think, misses the point entirely. . . . " Id. at 703 n.6.

[430] In Industrial Union Department, AFL-CIO v. American Petroleum Institute, 448 U.S. 607 (1980), the Court invalidated OSHA's benzene regulations because they were not supported by a proper risk analysis as required by Congress. The Court used the requirement of a risk analysis to check the otherwise unlimited power of OSHA to promulgate workplace exposure standards. 448 U.S. at 653.

[431] See supra note 231 and accompanying text. Arguably, Texas capital murder defendants would be better protected by a law that requires the expert to make specific findings on the probability of dangerousness (D) and the nature of that dangerousness (T). The jury would then determine if the product of these factors outweighed the value of the defendant's life.