Preliminary Thoughts on Evaluation Research for the "Expanded Addiction Treatment" Policy

Regarding the Executive Yuan's planned "expanded addiction treatment" policy, the author previously published two articles pointing out that existing research is incomplete and recommending the immediate initiation of larger-scale follow-up studies. However, these only outline the general direction; in practice, designing a methodologically sound recidivism study is no easy task. Taking comparative studies between compulsory drug rehabilitation and addiction treatment for drug offenders as an example, this article proposes several key dimensions in the hope of eliciting further discussion.

1. Interdisciplinary Research Teams

Fields to include: medicine, psychological counseling, social work, correctional practice personnel, and criminal policy researchers. Overcoming drug and alcohol addiction often requires medical personnel, but these behaviors may also have psychological roots requiring psychological counselors. For example, domestic experiments have found that psychotherapy is superior to medication alone in treating alcoholism (Ko et al., 2001, Discussion on Addiction Treatment Models for Alcoholism). Furthermore, criminological research shows that adjusting an offender's family, social, and workplace environments significantly helps reduce recidivism, which requires the intervention of social workers.

Taiwan's correctional system already possesses considerable talent with psychological counseling or social work backgrounds, and civil society organizations are not lacking in social workers who have long interacted with offenders. Inviting these individuals to join the team will undoubtedly enhance research capacity. What criminal justice is concerned with is whether the fruits of the efforts made by these medical, psychological, and social work personnel can be translated into a reduction in recidivism, which can be analyzed by criminal policy researchers using crime and judicial statistics.

2. Prior Notice, Consent, and the Right to Withdraw at Any Time

Any correctional measure involves "changing the offender," which may include altering their personality, mindset, behavioral habits, etc. This falls under deep intervention in the core domain of the right to the free development of personality; the state must not impose this coercively, nor may it force offenders to participate in correctional research. Since the various treatments discussed here—whether imprisonment, observation and rehabilitation, compulsory rehabilitation, or addiction treatment—are part of criminal proceedings and inevitably carry varying degrees of coercive implications or concerns, prior consent from research participants (offenders) is a necessary condition.

1. Referring to standard practices in criminological and psychological research, the summary of the research should be explained to offenders beforehand, informing them that: participation in research is not mandatory, refusal to participate will not result in any disadvantages, and agreeing to participate will not confer special benefits (though the possibility of earlier release from prison or facilities due to a lowered individual risk of recidivism is a reasonable benefit permitted by law). Participants may also withdraw at any time after agreeing to join, without suffering any disadvantages (other than the reasonable consequence of failing to achieve an earlier release due to an unreduced risk of recidivism).

2. Written explanations should be distributed for offenders to take back, read, and ponder for several days. If they still agree, they should be asked to sign and return the document before the research can begin.

3. Researchers should inform participants each time they meet, prior to commencing the session's measures, that they can withdraw from the research at any time.

4. If juvenile offenders are the research subjects, it must be noted that consent from their statutory representatives may be required.

3. Measurement of Recidivism Risk

Measurements should be taken at least once before and once after treatment (intervention) to determine whether the measures genuinely reduce recidivism risk. It is naturally better to take multiple measurements at different time points: first, to reduce potential errors that occur when only a few measurements are taken; second, much like a doctor tracking a patient's condition through follow-up visits, this allows for timely adjustments to the treatment course and interventions.

1. Medicine possesses certain tools for assessing the risk of drug relapse. Although not designed for predicting crime, because the use of Category 1 and Category 2 drugs is legally defined as a crime, it can be inferred that these tools have good predictive validity for the recidivism of Category 1 and Category 2 drug users. These tools can be operated by physicians on the team.

2. Additionally, criminological tools for assessing recidivism risk can be utilized as a supplement. The Kaohsiung Juvenile and Family Court already has experience using the "Youth Level of Service/Case Management Inventory (YLS/CMI)" (Li Feng-sheng, 2011, Level of Service and Case Management Inventory, Kaohsiung Juvenile, 19, 53-61; Ch'en Chen-su and Kao Min-bang, 2009, Report on the Trial Implementation of the Level of Service and Case Management Assessment Inventory by the Taiwan Kaohsiung Juvenile Court, Kaohsiung Juvenile, 15, 19–26), although this is a general crime prediction tool primarily designed for juvenile offenders—meaning it is neither targeted specifically at drug recidivism nor at adult offenders.

However, the YLS/CMI was actually developed from the LS/CMI designed for adult offenders, and domestic research shows that risk factors for drug recidivism largely correspond to three scales within the YLS/CMI and LS/CMI: "criminal history," "procurement/employment," and "peer relations" (Ch'en Ch'ao-fan and Lin Chien-yang, 2015, Research on the Degree of Recidivism Risk and Related Factors Affecting Recidivism Among Persons Undergoing Rehabilitation).

Furthermore, Minister without Portfolio Lo Ping-cheng pointed out during the "Judicial Reform National Conference Review Seminar" on December 24 that due to limited resources, expanded addiction treatment may be prioritized for juvenile offenders. Synthesizing these tool characteristics and the new policy content, the experience of the Kaohsiung Juvenile and Family Court is invaluable, and juvenile probation officers from the court could perhaps be invited to join the research.

4. Attempting Experiments

The Ministry of Justice's policy brief on April 20 (page 5) cited the speculation of Hsu Chun-chin et al. (2013, Research on the Feasibility of Treatment and Decriminalization for Drug Users, page III): the custodial nature of rehabilitation centers and the abundance or scarcity of medical and counseling resources may affect correctional efficacy. Comparative studies between addiction treatment and compulsory rehabilitation should indeed account for these impacts, which can be controlled and verified to a considerable extent through experimental design. Broadly speaking:

1. If permitted by law, pre-intervention recidivism risk scores can be used for matching, and offenders can be randomly assigned to compulsory rehabilitation and addiction treatment groups.

2. To minimize confounding factors, efforts should be made to create equal conditions between compulsory rehabilitation and addiction treatment. For example: providing identical resources, with the same group of doctors, psychologists, and social workers performing duties across both systems. If significant differences in recidivism rates are still obtained after minimizing confounding factors, they can more likely be attributed to structural or institutional differences between the two systems.

3. The seemingly unalterable custodial nature can actually have its intensity adjusted to observe the degree of impact. For example, also matching by recidivism risk scores, persons undergoing compulsory rehabilitation could be randomly assigned to groups with more or fewer visits from family and friends, or granted unequal opportunities for outings.

4. If the law does not permit random assignment as in (1) and (3) (as it might violate the constitutional principle of equality), item (2) can still be executed to create conditions that are as equal as possible, and post-hoc statistical controls can be applied to recidivism risk scores and the intervention characteristics mentioned in the next point, thereby enhancing the inferential validity of non-experimental research.

5. Measuring the Characteristics of Treatment (Correctional) Measures

1. Based on existing knowledge from correctional recidivism research, it can be inferred that there may be many differences between compulsory rehabilitation and addiction treatment that could affect treatment efficacy, such as: the custodial nature, opportunities for contact with family and friends, the average resources allocated to each patient (offender), whether improvements are targeted at dynamic risk factors, whether treatment (correctional) density is designed according to the level of risk, whether life counseling, family counseling, or employment counseling is conducted simultaneously with addiction treatment, the degree of closeness in the counselor-offender (or doctor-patient) relationship, and whether individuals are treated with dignity, etc.

2. Regardless of whether experiments are conducted, these characteristics should be measured. If experiments are feasible, measuring these characteristics helps in-depth analysis to search for moderating or mediating variables. If experiments are not feasible, statistically controlling for recidivism risk and these characteristics—as mentioned in the previous point—can reduce the influence of confounding variables and increase the probability of correct inference.

6. Tracking Drug Relapse (Recidivism) Phenomena

Two commonly used methods are self-reports by offenders and official records.

1. Self-reports have the potential to be closer to the truth, but offenders may report falsely due to a lack of trust in researchers or fear of prosecution; therefore, establishing a relationship of trust between researchers and offenders is key. The advantage of official records is that they usually have objective evidence for verification, while the disadvantages are the problem of dark figures of crime and susceptibility to judicial resources and policy fluctuations. Therefore, it is recommended to use both methods.

2. It is currently known that drug recidivism peaks within 2 to 3 years after treatment; the closer the time is to the end of treatment, the higher the risk of recidivism. Although lower starting from the 4th year, it cannot be ignored even up to the 5th year. Taking persons undergoing compulsory rehabilitation as an example, the marginal recidivism rate only drops below 5% in the 4th year after release, and is still 3% in the 5th year. Therefore, it is recommended to track for at least 3 years post-treatment, with 5 years or more being even better. As for frequency, it is recommended to track at least once every 6 months, obtaining self-reports of the number of recidivism incidents and dates of recidivism for each period; if manpower and resources permit, more intensive tracking is naturally better.

3. Although official records are easier to obtain—files are created whenever an offense enters the judicial system—it must be noted that in current judicial statistics, the "elapsed period before recidivism" is calculated "from the date of release from the facility to the date of new investigative case registration and assignment" (see page 61 of the annual "Judicial Statistics Yearbook"). This is not the true date of recidivism and will cause errors in analysis (underestimating the speed of recidivism). It is recommended that when using official criminal records, information on the date of recidivism should be taken from the crime date recorded in the case file or obtained by questioning the defendant.

7. Privilege Against Self-Incrimination and Confidentiality Obligations

During the research process of criminology and criminal policy, research teams frequently learn of criminal acts undiscovered by official authorities. Since addiction treatment inherently possesses the nature of medical and psychological treatment, physicians, psychologists, and their assistants on the team inherently possess a privilege against testifying and bear a confidentiality obligation. However, there may still be individuals on the team to whom relevant regulations do not apply, and other types of treatment do not necessarily involve therapy; thus, this advocacy is raised here:

The Code of Criminal Procedure and the Criminal Code should be amended as soon as possible to explicitly grant criminologists and criminal policy researchers (the entire team) a privilege against testifying and confidentiality obligations equivalent to those in the doctor-patient relationship. This prevents offenders from cooperating with research only to be betrayed, or prevents the state from abusing the obligation to testify, which would ultimately destroy the trust between researchers and subjects. Prior to legislative amendments, constitutionally guaranteed academic freedom may perhaps serve as the basis for such rights and obligations.

Conclusion

Drug use and drunk driving are two major challenges in constructing Taiwan's social safety net. Researching and proving the efficacy of various related criminal policies carries high practical, social, and national fiscal significance. Expanded addiction treatment for drug-using offenders is still in the planning stage, while alcohol addiction treatment—planned concurrently—will be brand new. This presents a prime opportunity for comparative research between old and new systems.

It is recommended that relevant authorities such as the Ministry of Justice, the Ministry of Health and Welfare, and the Judicial Yuan allocate resources for more, larger-scale, and longer-term addiction treatment research, and begin carefully clarifying the efficacy of the old systems about to be replaced so they can be compared with addiction treatment. Through solid research, relevant policies and discussions can be made more scientifically grounded, enabling targeted solutions, avoiding wasted efforts, dispelling public misunderstandings, and securing nationwide support for the policies.