**Table 1.**

*The list of screening and severity diagnosis assessments.*

test [45], SAT had excellence in test-retest [43]. Eight items of TOP reached 0.75, and eight items are below 0.6 [45]. ATOP was proposed by Australia researchers to assess alcohol or drug use and its risk profile, general health, and well-being. ATOP contains 22 items and averagely scored more than 0.7 in test-retest [46]. In test-retest, ATOP had 19 items excellent, 1 item good, and 2 poor.

Zilm and Sellers (1978) proposed a quantitative technique to assess the level of physical dependence of narcotics, with administering naloxone [47]. They gave an equation of objective severity scoring index (OSSI). However, this method has not been tested in reliability or validity, and Zilm and Sellers claim it relies on the experience of executors.

## **2.4 Assessments selecting**

All assessments are listed in **Table 1**. It concludes the target substance, number of questions, assessment approach, and reliability. The reliability is from test-retest, and the reliability coefficient below 0.40 is Poor; 0.40 to 0.59 is Fair, 0.60–0.74 is Good, and 0.75–1.00 is Excellent [48]. There are other assessments, such as Antisocial Personality Disorder, CIDI, General Health Questionnaire, Primary Care Posttraumatic Stress Disorder Screen, Health of the Nation Outcome Scales, and Michigan Alcoholism Screening Test, designed for psychological or alcoholic diagnosis and are not discussed in detail in this review.

Two main approaches of drug use disorder severity assessments are interview and self-report. In terms of reliability, there is no significant difference between interview and self-report. Several studies have proved that self-report assessments are as reliable as interview ones [49–51]. Compared with interview, self-report is more cost-effective and convenient, but the understanding of questions might affect the accuracy of selfreport. Moreover, self-report instrument is more likely to collect honest answers and face-to-face interview might be unsuccessful to, because the questions would make the interviewees uncomfortable [52]. In interview assessments, there are two types, semi-structured and fully structured. Both of them have advantages and disadvantages. Fully structured interview does not need clinical judgment, and as a result, it does not need experienced clinicians. Semi-structured interview, in contrast, can obtain more detailed information of patients' status, but more human cost and time cost [53].

Specific to each instrument, the reliability has been listed above, and all assessments are generally reliable. Some studies compared different assessments and found no significant difference in general, but disagreement in specific field [54, 55]. For example, the reliabilities of SDSS for alcohol, cocaine, heroin, and sedatives were excellent, but for cannabis, it was just fair [24]. SSADDA is more sensitive to cocaine and opioid [27]. In addition, the validity of assessments may not vary between different races. Taking DSM-IV as an example, Horton et al. reported that there is no significant difference between African-Americans and Caucasians, when using this assessment [55]. Taken together, when screening instruments or severity assessments were selected, factors, including genders, different stages of drug use or withdrawal, reliability in different drugs, time, human resource and economic cost, and the condition of patients, should be considered. It is important to choose one or more assessments, based on patients' conditions to get accurate results.
