Psychometric Properties of the Persian Structured Clinical Interview for Personality Disorders of DSM-5-Personality Version (SCID-5-PD)


 Background: The aim of this study was to investigate the psychometric properties of the Persian version of a structured clinical interview for personality disorders based on DSM-5 (R) (SCID-5-PD) in the population of patients with psychiatric disorders in Tehran.Method: The study population includes all outpatients and inpatients referred to three psychiatric centers in Tehran, including Iran Psychiatric Hospital, Rasoul Akram Hospital and Clinic of Behavioral Sciences and Mental Health (Tehran Psychiatric Institute). Inclusion criteria were age between 16 and 70 years, written Informed consent and the ability to understand and speak in Persian and no special physical problems that interfere with the interview process. Sampling in this study was done by Convenience sampling. In this study, in addition to the demographic questionnaire, the Persian version of SCID-5-PD was used. Finally, in order to evaluate the information, the methods of face and content validity and diagnostic validity, test-retest reliability and inter-rater reliability were used.Results: the diagnoses related to obsessive-compulsive personality disorder, paranoid, schizotypal, schizoid, histrionic, narcissistic, borderline and antisocial kappa were higher than 0.4 and the diagnoses related to avoidant personality disorder were dependent and in other certain disorders are below 0.4. Regarding borderline personality disorder with kappa 0.839, the highest agreement was reported between the two reports of the psychiatrist and the SCID interviewer. Also, the specificity results were mostly better than the sensitivity results, and in all diagnoses except obsessive-compulsive and paranoid personality disorder, the specificity was higher than 0.9 and in these two diagnoses, the specificity was higher than 0.85, which indicates the desired characteristic. SCID-5-PD. The sensitivity of all diagnoses except avoidant and dependent personality disorder was also reported to be higher than 0.8; But the susceptibility of avoidant and dependent personality disorder was 0.66. Also, the study of LR + / LR- ratio showed that this tool has the best diagnosis for histrionic, antisocial and schizotypal personality disorder. It is also suitable for other personality disorders except schizoid personality disorder and certain other disorders.Conclusion: According to the findings of the present study, SCID-5-PD can be used in psychiatric clinics and hospitals as a diagnostic tool. In general, this version is suitable for most diagnoses; but with regard to diagnoses of schizoid personality disorder and certain other disorders, this should be done with more caution.


Background
Approximately a quarter of patients in mental health centers and 45% of outpatients meet the minimum criteria for a personality disorders (PDs) (1,2). PDs is associated with a wide range of psychological disorders(3), higher mortality(4), poor therapeutic outcome (5), high burden Disease(6), high economic burden (7). On the other hand, patients with PDs who are not diagnosed clinically may receive ineffective or harmful treatments. Therefore, accurate and timely evaluation of PDs is very important. To evaluate PDs and Psychiatric diagnosis, Structured Clinical Interview for DSM-IV Axis I Disorders (SCID-I) and Axis II Disorders (SCID-II) are acceptable validity and reliability scales (8,9).
Prior to the structured clinical interview for the DSM-5 (SCID-5), the structured clinical interview for the DSM-IV (SCID-IV) demonstrated acceptable internal reliability(8) and acceptable test-retest reliability (10). However, psychometric studies for SCID-5 have not been published. Following the publication of the fourth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) (11), several structured diagnostic interviews were conducted and conducted over the years by researchers and clinicians. In Iran, the validity and reliability of SCID based on DSM-IV were evaluated in a study. In this study, 299 patients aged 18 to 65 years referred to outpatient clinics and inpatient wards of psychiatric centers in Tehran participated. In the test of reliability by test-retest method, 104 of the clients were evaluated with SCID independently (with an interval of three to seven days) and the agreement between the two rounds of diagnosis was measured. Overall, the ndings showed that the diagnostic agreement was moderate to good for most speci c and general diagnoses (kappa above 0.6). The overall agreement (total kappa) was 0.52 for all current diagnoses and 0.55 for all lifetime diagnoses (11).
On the other hand, with the introduction of the fth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5) and the signi cant changes that were made in the DSM-5 compared to the DSM-IV, the need for a diagnostic interview based on DSM-5 diagnostic criteria became necessary. SCID-5 has 5 versions: Clinician Version (12), Personality Disorders version (13), Research Version (14) and Clinical Trials Version (15) and Alternative model for Personality Disorders(16). The purpose of this study is speci cally to examine the psychometric properties of structured interviews based on the Structured Clinical Interview for personality disorders of DSM-5-personality Version (SCID-5-PD)(SCID-5-PD). SCID-5-PD is used to assist therapists and researchers in assessing 10 DSM-5 personality disorders in clusters A, B, and C and otherwise identi ed personality disorders. SCID-5-PD can be used for dimensional or categorical diagnostics. This structured interview is an updated version of the DSM-IV structured interview for Axis II disorders. Although the criteria for personality disorders in the DSM-5 have not changed, the interview questions have been carefully reviewed to optimally identify the underlying construct in the diagnostic criteria. In addition, a dimensional scoring has been added. The questions assess 10 personality disorders: avoidant, dependent, obsessive-compulsive, paranoid, schizotypal, schizoid, hysterical, narcissistic, and antisocial personality disorders. SCPD-5-PD can be used in a variety of research studies, just like Structured Clinical Interview for Axis II Disorders (SCID II) (13).
The only study that has been done on the standardization of SCID-5 versions so far is a study that has examined the clinical version of SCID-5 in Iran (17). The results of this study showed that considering the kappa criterion, for all diagnoses except anxiety disorders, kappa was higher than 0.4 as a result of agreement above average, but in anxiety disorders with kappa 0.34 indicates moderate agreement between the two reports. Also, according to the diagnosis of psychiatrists as a gold standard, in most diagnoses except anxiety disorders, it was higher than 0.80, which indicates the desirable characteristics of this tool in the diagnosis of disorders. The sensitivity of all diagnoses was higher than 0.80. In fact, it has the right diagnostic capability for all disorders except depressive disorders. Also, the ndings showed that Structured Clinical Interview for personality disorders of DSM-5-clinical Version (SCID-5-CV) has testretest reliability in all diagnostic disorders except depressive disorders. According to the ndings, SCID-5-CV can be used for screening purposes and in psychiatric clinics and hospitals and to evaluate the treatment process of patients. In general, this version is suitable for most diagnoses, especially schizophrenia and other psychotic disorders; but with regard to diagnoses related to depressive and anxiety disorders, this should be done with more caution.
Therefore, considering what was mentioned and considering the need for accurate evaluation of PDs and the fact that the SCID 5-PD version has not yet been standardized in Iranian society, and since structured interviews and accurate evaluation of patients, the basis of most medical and research work in The purpose of this study is to investigate the psychometric properties of the Persian version of SCID 5-PD in Iranian society.

Method
The study population included all outpatients and inpatients referred to three psychiatric centers in Tehran, including Iran Psychiatric Hospital, Rasoul Akram Hospital and Clinic of Behavioral Sciences and Mental Health (Tehran Psychiatric Institute).Among them, 250 people according to the inclusion criteria in this study, which include age between 18 to 65 years, written Informed consent and the ability to speak Persian, as well as exclusion criteria, including severe mental retardation or dementia, symptoms severe psychosis. Sampling in this study was done by Convenience sampling. After the required coordination with the mentioned centers, the researchers referred to the centers and prepared a list of all hospitalized and outpatient patients every day. Then, patients who met the criteria for entering the study were invited to enter the study with their written Informed consent. It should be noted that interviews with outpatients were performed on the same day of referral and interviews with hospitalized patients during the rst week of hospitalization. Interviewers included PhD students in clinical psychology who had undergone a special training course. Interviews were conducted with the presence of two evaluators without access to patients' records. One of the evaluators conducted the interview, but both lled out the checklist independently. Also, in order to evaluate the reliability of the test-retest among the sample members, 109 patients were evaluated independently in the second round at intervals of 7-10 days.
Also, in the present study, in order to evaluate the psychometrics of the Persian version of SCID-5-PD, the translated version of SCID-5-PD by Amini et al.(18) Was used. The following methods of validity and reliability in the present study will be brie y explained: 1. Face validity: To determine the face validity, 5 clinical specialists, all of whom had at least 2 years of clinical experience, were asked to answer the questions raised in the forms regarding the evaluation of face validity and content. Formal validity was achieved through two methods, qualitative and quantitative.
A) Determining qualitative face validity: In qualitative assessment of face validity, observance of grammar, use of appropriate words, importance of items, placement of items in their proper place, time of completing the designed tool were considered. After collecting the opinions of experts, in consultation with the members of the research team, the required changes were made in the translation of SCID-5-PD. B) Determining the quantitative face validity: The face validity of the measures was measured quantitatively using the item effect method. For this purpose, the specialists completed a questionnaire that based on each item SCID-5-PD in terms of importance. It is quite important (score 5), it is important (score 4), it is moderately important (score 3), it is slightly important (score 2) and it is not important at all (score 1). After completing the questionnaires face validity was calculated by experts using the formula of item effect method. If the impact score of the item was more than 1.5, the item was considered suitable for the next analysis.
2. Content validity: To determine the content validity, two qualitative and quantitative methods were used: A) Determining the validity of qualitative content: In the qualitative method, the Persian translation version was provided to the experts to make them desirable in terms of clarity in terms of clarity (use of simple and understandable words), use of common language (avoidance of using Technical and specialized terms) to examine. If necessary, changes were made to the translated text of SCID-5-PD for simpli cation and greater comprehensibility. Experts also examined the text in terms of di culty in understanding phrases and words, the appropriate t and relevance of items, the possibility of ambiguity and incorrect interpretations of phrases or the existence of inadequacies in meanings, and if there are problems, their comments as minor changes in the questionnaire applied. B) Determining the validity of Quantitative content: Content validity was also quantitatively calculated based on the opinions of experts and by calculating two content validity ratio (CVR) and content validity index (CVI). The content validity ratio index was used to ensure that the most important and correct content (item necessity) was selected, and the content validity index was used to ensure that the tool items were best designed to measure content. To determine the content validity ratio, experts responded to each of the SCID-5-PD items in a range of essential, useful, but not necessary and unnecessary. Then the content validity ratio (CVR) was calculated. Finally, if the calculated value is greater than 0.99 (minimum validity value for a panel of 5 experts), the validity of the content of that item will be accepted.
3. Diagnostic validity: Psychiatrists diagnosis was considered the gold standard. The gold standard of diagnosis was the records in the hospital/clinic les according to the routine standards of this universitya liated hospitals/clinic. This routine include 1) early interview with the patient by a resident of Psychiatry, 2) gathering the history data of patient including lifetime course of the disorder, and any previous treatment and recorded Psychiatric diagnosis in outpatient and inpatient settings, 3) interview with accessible family members, and 4) recording the nal diagnosis by a supervisor Psychiatrist based on an independent interview with the patient and all the gathered data, according to the DSM-5. Then, after a short break, the client was invited to participate in the research and his rights, including the freedom not to continue the research at any stage, etc. were explained and a written consent form was completed. patients who met the research requirements and agreed to take the test were referred to a SCID interviewer. The interviews were conducted in a separate room, privately, without access to patients' records. The SCID interviewer completed the interview without any knowledge of the patient's diagnosis. Considering the psychiatrist's diagnosis as a gold standard, the agreement between SCID diagnoses and psychiatrist's diagnosis through kappa criteria as well as sensitivity, speci city, positive and negative Likelihood ratios(LR+/LR-) were investigated. 4. Reliability of the test-retest: To check the reliability of the test-retest, after the rst SCID interview, the interviewer asked the client to come for the next 7-10 days for the next interview. In the second visit, the second interviewer, who had no knowledge of the diagnostic results and previous interviews, conducted the SCID interview in full. Correlation between the rst and second scores of the SCID interview was used to determine the test-retest reliability. The test-retest reliability between the rst and second rounds of SCID-5-PD was evaluated using the intra-cluster correlation index.
5. Inter-rater reliability: To evaluate the inter-experimental reliability, the kappa value obtained from SCID-5-PD by two interviewers was used.
In the present study, in addition to the demographic questionnaire, the Persian version of SCID-5-PD was used: Demographic Characteristics Questionnaire: Personal information questionnaire includes gender, age, level of education, marital status and occupation and number of children, history of psychiatric disorder and substance or alcohol abuse, history of suicide and legal problems, and history of medication and hospitalization. SCID-5-PD can be used for dimensional or categorical diagnostics. This structured interview is an updated version of the structured interview based on the DSM-IV for Axis II. The questions assess 10 personality disorders: avoidant, dependent, obsessive-compulsive, paranoid, schizotypal, schizoid, hysterical, narcissistic, and antisocial personality disorders (10). Finally, in order to describe and analyze the research data, descriptive and inferential indicators and statistical methods such as sensitivity, speci city, positive and negative likelihood and phi coe cient were used.

Result
This study is a descriptive and correlational study. As shown in Table 1, the demographic characteristics of the present study sample; in total, the data of 250 patients were analyzed and the age range of patients was between 17 and 74 years with a mean of 32.56 and a standard deviation of 10.68. Among the patients, 123 (49.2%) con rmed the existence of thoughts or history of suicide attempt. Also, 74 people reported a history of substance or alcohol abuse. Also, 241 patients (96.4%) had a history of drug treatment of psychiatric disorders in the past and 212 patients (85.8%) had been hospitalized in a psychiatric hospital at least once in the past for more than one day (Table 1). The data of 250 people were evaluated and nalized, and to evaluate the test -retest reliability, 109 (43.6%) patients were evaluated again by SCID-5-PD after 7-10 days. As the Fig. 1 shows, according to the patients records, 80 patients were without a diagnosis of a psychiatric diagnosis. The highest frequency of disorders in the studied patients was related to depressive disorders (n = 74) and bipolar disorder (n = 40). Schizophrenia with 15 people, obsessive-compulsive disorder with 13 people and substance abuse with 12 people were and other disorders were less frequent with 4 people. Figure 1 shows Frequency of psychiatric disorders based on DSM-5 In order to evaluate the qualitative face validity collected opinions of experts in a joint meeting and then required changes were made in the translation of SCID-5-PD. Also, item impact scores were used to evaluate the quantitative face validity. As the results in Table 2 show, the impact factor of all questions is higher than 1.5, which is the acceptance score, and the minimum impact factor is 3. Therefore, the face validity of all questions was con rmed.
To determine the content validity, two methods were used: qualitative and quantitative. Qualitative validation was performed by 5 experts in the eld of psychology. In order to quantitatively evaluate the content validity based on the opinions of these 5 experts (necessity of the question) was calculated by CVR method. Each specialist responded to each item in a range of essentials, useful but not necessary and unnecessary. Responses were calculated based on the CVR formula and adapted to the Lawshe (19). Numbers higher than 0.99 were accepted. Based on this, and considering that all members considered all items necessary, all items were approved.
The Waltz and Basal indices (20) were used to calculate the CVI. For this purpose, experts were asked to rate each interview item on the three criteria of relevance or speci city, simplicity and uency, and clarity or transparency, based on a 4-part Likert scale. A score of 0.79 and above is recommended for accepting items based on the CVI score. If the content validity index score is between 0.70 and 0.79, the phrase needs to be revised and revised, and if the content validity index score is lower If it is 0.70, the phrase should be deleted (21). The content validity index score was calculated by summing the agreeing scores for each item that ranked 3rd and 4th (highest score) on the total number of voters. Due to the fact that the value obtained for all items was higher than 0.8, which was within the acceptable range, all items were approved (Table 2). Table 2 shows the agreement between SCID and psychiatrists diagnoses(kappa) and also the sensitivity, speci city, positive and negative likelihood ratio and LR + / LR-ratio of SCID-5-PD, if the psychiatrist's diagnosis is the gold standard. As can be seen in this table, the diagnoses for obsessive-compulsive, paranoid, schizotypal, schizoid, histrionic, narcissistic, borderline, and antisocial personality disorder kappa are higher than 0.4 and the diagnoses for avoidant, Dependent personality disorder and other certain disorders are below 0.4. Borderline personality disorder with kappa 0.839 had the highest agreement between the psychiatrist and the SCID interviewer. Therefore, it can be concluded that there was a low agreement between the psychiatrist and the SCID interviewer regarding the diagnosis of obsessive-compulsive and paranoid personality disorder and certain other disorders, and a high agreement on the diagnosis of other disorders.
If we consider the diagnoses presented by psychiatrists as a gold standard, the results of speci city were mostly better than the results of sensitivity, as in all diagnoses except obsessive-compulsive and paranoid personality disorder, the characteristic was above 0.9 and in these two disorders also had a speci city higher than 0.85, which indicates the favorable characteristics of SCID-5-PD. The sensitivity of all diagnoses except avoidant and dependent personality disorder was higher than 0.8. But the susceptibility of avoidant and dependent personality disorder was 0.66. Also, the study of LR + / LR-ratio showed that this tool has the best diagnosis for histrionic, antisocial and schizotypal personality disorder. The tool is also suitable for other personality disorders except schizoid personality disorder and certain other disorders To test the inter-rater reliability, two testers separately completed the interview scores. As can be seen in Table 3, the results obtained from the phi coe cient of the rst and second interviewers showed that in the diagnoses of paranoid, histrionic, narcissistic, borderline and antisocial personality disorder, there is a very strong relationship in Diagnoses of avoidant, dependent, obsessive-compulsive, and schizoid personality disorder were strongly correlated, and in the diagnosis of schizotypal personality disorder, there was moderate association, all of which were signi cant at the level of α = 0.000. But in the other personality disorders, there was a very weak and insigni cant relationship between certain other disorders at the signi cance level of α = 0.912. Therefore, SCID-5-PD has a good inter-rater reliability for all personality disorders except the other personality disorders and schizotypal disorder.
Also, again 109 patients were interviewed to evaluate the time reliability of the interview. The results obtained from the phi coe cient of the rst and second interviews (Table 3) showed There is signi cance a very strong correlation between the diagnoses of dependent, schizotypal, narcissistic, borderline and antisocial personality disorder in the rst and second interviews at the level of α = 0.000. In the diagnoses of avoidant, obsessive-compulsive, paranoid, histrionic personality disorder and personality disorder category of certain other disorders, there is a strong correlation between the rst and second interviews at the level of α = 0.000 In the diagnosis of schizoid personality disorder, there is a moderate relationship between the rst and second interviews at the level of α = 0.000. Therefore, SCID-5-PD has a good testretest reliability for all personality disorders except schizoid personality disorder.

Discussion
The present study was designed and conducted to investigate the psychometric properties of the Persian version of the SCID-5-PD. For this purpose, 250 hospitalized and outpatient patients referred to three psychiatric centers in Tehran, including Iran Psychiatric Hospital, Rasoul Akram Hospital and Clinic of Behavioral Sciences and Mental Health (Tehran Psychiatric Institute)., were included in the study and structured clinical interviews were conducted with them for personality disorders. 109 patients were interviewed again after 7 to 10 days. Then, the data were analyzed to evaluate the content validity, face validity, diagnostic validity, as well as the reliability between the evaluators and the reliability of the testretest.
Face and content validity studies showed that all items are essential; they are speci c, simple and clearly stated and relevant to the subject matter. According to searches in scienti c databases, so far only one study has examined the psychometric properties of SCID-5-PD in Italian (22), in which only the agreement between the evaluators in the diagnosis and validity has been examined. Face, content and diagnostic have not been studied in that study.
In a study conducted to evaluate the psychometric properties of the Persian version of a SCID-IV for personality Disorders for Axis II, similar results were reported with the ndings of the present study on face and content validity (23). On the other hand, since the SCID-5-PD interview questions have been revised, the diagnostic criteria for personality disorders in the DSM-IV and DSM-5 have not changed, and in the SCID-5-PD interview for each diagnostic criterion. There is at least one question(13), so con rmation of face and content validity in the new version was also expected.
The results of the present study showed that the degree of agreement between SCID diagnoses and psychiatrists' diagnosis (as a gold standard) in diagnoses related to obsessive-compulsive disorder, paranoid, schizotypal, schizoid, histrionic, narcissistic, borderline and antisocial kappa is higher than 0.4 and Diagnosis of avoidant, dependent, and other personality disorders was reported to be less than 0.4. Borderline personality disorder with kappa 0.839 had the highest agreement between the psychiatrist and the SCID interviewer. Characteristics and sensitivity studies also indicate that this tool has the best diagnosis for histrionic, antisocial and schizotypal personality disorder. The tool is also suitable for other personality disorders except schizoid personality disorder and certain other disorders. To explain these ndings, there is no study that can compare the results with the present study, but about the low agreement between the evaluators about dependent, avoidance and other personality disorders, as well as the low speci city and sensitivity of the interview. To diagnose schizoid personality disorder and other personality disorders, it is necessary to pay attention to the frequency of these disorders in the sample of the present study (dependent disorder 3 people, avoidance disorder 3 people, schizoid disorder 2 people and other certain disorders zero). Probably because of their very low base rate in the sample under study, which makes statistical comparisons di cult and meaningless. It seems that the very high base rate of borderline personality disorder has resulted in the highest level of agreement among evaluators.
The low agreement on schizoid personality disorders and certain other disorders can also be due to the lack of a comprehensive examination of personality disorders by psychiatrists in hospitalized patients. Because in the medical environment and the hospital, priority is given to controlling the obvious and more severe symptoms of patients. Studies have also shown that people hospitalized for psychiatric disorders receive only Axis 1 diagnoses, unless their clinical picture is directly related to personality disorder, such as suicidal behaviors, or if personality disorder is very prominent (24). In fact, this may be due to a lack of time for a comprehensive review of personality disorders or a lack of awareness of the importance of personality in the management and prognosis of psychological disorders. Therefore, the comorbidity rate in studies that use unstructured clinical evaluations is much lower than in studies that use structured evaluation tools. Also, patients with Axis I disorder are treated similarly, and the speci c vulnerabilities of patients with personality disorder are ignored (25,26). It has also been reported that clinicians consider the possibility of a comorbid personality disorder primarily when the patient has not responded to treatment (27). The results of a study in line with the present study have shown that people with cluster C personality disorders remain generally unknown and generally have manifestations such as mood disorders and physical symptoms(28).
All of these factors can affect a psychiatrist's diagnosis and cause the psychiatrist's diagnosis to be inconsistent with a structured clinical interview for personality disorders, which in turn reduces the rate of agreement between assessors on speci c disorders. As predicted from previous SCID-II reliability studies (13,23), the Persian translation of SCID-5-PD is capable of diagnosing personality disorders classi ed in the second part of DSM-5 personality disorders with good to excellent reliability (29,30).
Although in our sample the baseline rates of dependent, avoidance, schizoid, and other personality disorders were too limited to perform agreement assessments between evaluators, so that "other personality disorders" were never evaluated by a psychiatrist, the results showed reliability between evaluators. The Persian version of SCID-5-P has good inter-rater reliability for all personality disorders except the personality disorder category of other personality disorders.
Also, the phi coe cient between the rst and second interviews showed that SCID-5-PD has good testretest reliability for all personality disorders except schizoid personality disorder and other personality disorders. Again, it was not found to compare and explain the results of a study that speci cally assessed the test-retest reliability of the SCID-5-PD. However, in studying the psychometric properties of SCID-II(23), similar results were obtained regarding the test-retest reliability of structured clinical interviews for personality disorders. Moreover, since personality disorders are persistent and pervasive in nature and do not change much over days and weeks(31), it is not unreasonable to expect similar diagnoses to be made in time-lapse assessments. This research, like other researches in this eld, has limitations, so the data should be considered according to several limitations. Pairwise design generally provides a more optimistic estimate of the actual reliability of the instrument(32); However, it should be noted that paired interview design is the most commonly used method for assessing the reliability between testers due to its simplicity and most similarity to real situations. The samples of the present study were limited to a speci c geographical area, with a limited number, and were voluntary and purpose-based. These conditions make the generalization of results cautious. Also, limited sample and lack of access to all disorders in equal numbers and different prevalence of disorders can create limitations in the generalizability of results. The unavailability of an empty bed in a hospital also generally only allows patients with very severe symptoms to be admitted to the hospital. It should be noted that this population of patients with severe symptoms do not represent the general public. Even with these limitations, the results of the present study con rm the hypothesis that the Persian translation of SCID-5-PD has good inter-rater reliability, at least among clinical clients who voluntarily seek psychotherapy. Therefore, it is suggested that a similar study be conducted on samples with a variety of indigenous and demographic characteristics and, taking into account the results, to compare and examine the processes involved in different ecosystems of Iran. According to the results obtained in this study, it can be suggested that clinical specialists can use this version well in their research and treatment work, but caution should be observed regarding schizoid and schizotypal personality disorders.

Conclusion
The ndings of the present study indicate that the Persian translation of SCID-5-PD is likely to provide an acceptable assessment in the classi cation of personality disorders. It can be used as a reliable and valid tool in research and treatment.