Participants

Between June 2015 and January 2018, a total of 115 patients were referred to the San Juan Mental Health Day Hospital from several Mental Health Units. 69 patients belong to health area 19 (Alicante Province - General Hospital), and 46 belong to area 17 (Alicante Province - Sant Joan d'Alacant).

Out of the 115 patients, 79 of the patients are women and 36 are men. 19 dismissed the application (7 women and 12 men), and 11 did not meet the criteria required to enter the program. Therefore, the incidental sample (not randomized), is composed of a total of N = 85 individuals.

With regards to the Clinical Diagnosis for which they were referred to the day hospital, out of 85 individuals, 51 % are diagnosed with Borderline Personality Disorder, 47 % with Personality Disorder Not Otherwise Specified or Mixed Personality (PD-NOS/MP), and only 2 % with Other Personality Disorders (Obsessive-Compulsive Personality Disorder and Histrionic Disorder).

The socio-demographic profile of these patients would be as follows. By gender, 75 % of individuals are women, compared to 25 % of men. The average age is 34.5 years old (range between 18 and 54). 40 % are less than 30 years old, 31 % are between 30-40 years old, 21 % are between 40-50 years old, and 8 % are over 50 years old.

Regarding their living situation, 63 % of the people seen continue to live in the family home, and 24 % live alone or with their partner. There is no data for the remaining subjects. When considering educational levels, 24 % of individuals haven't obtained a basic education and 13 % have. 32 % have studied at intermediate level (BUP, baccalaureate, vocational training or training courses), and 8 % have university studies pending completion. Only 5 % have finished university studies. There is no data available for the remaining 18 %.

In relation to their economic situation, 58 % lack their own economic income. Regarding their employment situation, 68 % are unemployed, 18 % have jobs, of these, 15 % work occasionally and 3 % have a temporary occupational disability status. 8 % of patients have a National Social Security Institute acknowledged permanent occupational disability, and 26 % have an acknowledged disability.

Evaluation and Instrument Design

The structure of the evaluation in the Day Hospital is more complex, as it considers both individual assessment and assessment of the program itself. However, for the purpose of this piece of research, the main focus is specifically on the instruments that allow for a profile of users cared for in the centre to be determined. In order to select the instruments, certain factors were taken into consideration. The instruments needed to be precise, adequate, up to date, useful, reliable and valid. In the data collection referred to below, the patients consent was obtained. The patients were assured that their personal information was confidential and were informed of how this information would be used.

The following assessment instruments were used in the present investigation:

Initial assessment, Semi-structured interview. The most relevant information for the assistance in each case was collected. The Referral Sheets were taken as a reference point. Socio-demographic and clinical data was extracted from these interviews along with data related to admissions and visits to the emergency department in the last year.

International Personality Disorder Examination, IPDE Module DSM-IV, WHO (^{Loranger, Janca, & Sartorius, 1995}). There are two versions of this structured interview. The Spanish version, based on the DSM-IV (^{López-Ibor, Pérez, & Rubio, 1996}), was used. This version allows for the assessment of different disorders from the different functioning areas of the personality, a key aspect of the dimensional model in Section III of the DSM-5. In addition, the interview includes a self-administered screening questionnaire, which selects the most relevant items to be assessed in the interview.

The Personality Inventory of the DSM-5, PID-5-Adults (^{Krueger, Derringer, Markon, Watson, & Skodol, 2012}, ²⁰¹⁴). The Personality Inventory for DSM-5 (PID-5), Full version Adult, was used. It adopts the dimensional model, is a 220 item self-rated personality trait assessment scale for adults age 18 and older. It evaluates 25 facets of personality traits: Anhedonia, Anxiousness, Attention seeking, Callousness, Deceitfulness, Depressitivity, Distractibility, Eccentricity, Emotional Lability, Grandiosity, Hostility, Impulsivity, Intimacy Avoidance, Irresponsibility, Manipulativeness, Perceptual Dysregulation, Perseveration, Restricted Affectivity, Rigid Perfectionism, Risk taking, Separation insecurity, Submissiveness, Suspiciousness, Unusual Beliefs and Experiences, and Withdrawal. Certain triplets of facets (groups of 3), can be combined to assess the five domains of traits: Negative affect, Detachment, Antagonism, Disinhibition and Psychoticism.

The clinician is asked to calculate and use average scores for each facet and domain, to a 4-point scale. Higher average scores indicate greater "dysfunction" in a specific personality trait facet or domain, and lower scores, point towards an adaptive and resilient personality.

The five domains are maladaptive variants of the personality model known as the Big Five or Five Factors. They are also similar to the five domains of the psychopathological personality, Personality Psychopathology Five, PSY-5 (^{Anderson et al., 2014}).

In several studies the instrument showed good reliability and validity qualities (^{Al-Dajani, Gralnick, & Bagby, 2016}; ^{Bastiaens et al., 2016}; ^{Quilty, Ayearst, Chmielewski, Pollock, & Bagby, 2013}). In its Spanish validation, the facets showed good internal consistency (α = .86 and .79), and discriminant validity between clinical and non-clinical subjects. It is noted that the upper poles indicate pathology and the lower poles normality, as well as the hierarchical structure of the instrument (^{Echeburúa, Salaberría, & Cruz-Sáez, 2014}; ^{Gutiérrez et al., 2015}).

Borderline Evaluation of Severity Over Time, BEST (^{Pfohl et al., 2009}), developed by the authors of the STEPPS Treatment Program. A self-rated scale to measure severity and change over time in persons with BPD. It is composed of 15 items evaluated on a Likert type scale (1 to 5), which are grouped into three subscales. Both A Thoughts and Feelings (items 1 to 8), and B Negative Behaviour subscales (items 9 to 12), correspond to the DSM-IV-TR criteria for BPD. Each item is rated from 1 (none / slight), to 5 (extreme), in a selected time interval (30 days, 7 days, other). Subscale C evaluates Positive Behaviours (items 13 to 15), on a scale of 5 (almost always) to 1 (almost never). The instrument demonstrated moderate test-retest reliability (ρ = 0.62, n = 130, p < .001), high internal consistency (Cronbach's α .90 and .92), high discriminant validity and sensitivity to clinical change (^{Blum et al., 2002}, 2009). For analysis, the direct scores of the three subscales were converted into a mean score, all of them oscillated between 1 and 5, making them comparable to each other.

COPE-28, the Spanish version of Carver's Brief-COPE (1997; ^{Carver, 1989};^{Crespo & Cruzado, 1997}) by ^{Morán, Landero y González (2009}). This is a 28-item inventory, in which answers are obtained based on a Likert type ordinal scale. It is comprised of 4 alternative answers from 0 (I never do this), to 3 (I always do this). These items are grouped into 14 subscales of strategies or coping styles used by people in stressful situations: 1. Active coping, 2. Planning, 3. Use of emotional support, 4. Use of instrumental/social support, 5. Religion, 6. Positive reframing, 7. Acceptance, 8. Denial, 9. Humor, 10. Self-distraction, 11. Self-blame, 12. Behavioural disengagement, 13. Venting/venting of emotions and 14. Substance use. The authors found reliability with Cronbach's α coefficient test between .45 and .92, and a test-retest procedure of .46 to .86.

Design and statistical analysis

This is a two-stage study. In the first stage, which this article focuses on, between June 2015 and January 2018, a descriptive, cross-sectional study was carried out at the San Juan Mental Health Day Hospital, Alicante. In the second stage, a study was conducted to assess the results of the program.

The statistical analysis was performed with the SPSS package v15., using the Excel spread sheet as a supplement. Socio-demographic variables (age, gender, marital status, living situation, level of studies, employment status), and some clinical variables (use of health resources and clinical diagnosis according to DSM-5), were defined through descriptive statistics. These later define different profiles of the individuals, PD-NOS/MP and BPD.

In the self-reports and quantitative instruments (PID-5, BEST, COPE-28), along with the descriptive analysis for the global sample, the means of the subgroups were compared according to the diagnosis (PD-NOS/MP vs. BPD). The Student's t-test of comparison of means for independent samples was used. To check if the equality of variances is assumed, the Levene test for Equality of Variances was applied in each comparison of means. The effect size for each pair of comparisons between groups was obtained using ^{Cohen's δ, following the interpretation proposed by Cohen (1988}). δ between .50 and .79 would, indicate a moderate effect, above .80 would be considered a large effect.

Clinical Profile of the Sample

As previously mentioned in the participants section, the main clinical diagnoses found for the 85 individuals admitted are; Borderline Personality Disorder (51 %), PD-NOS or Mixed Personality (47 %) and only 2 % Other Personality Disorders (Obsessive-Compulsive Personality Disorder and Histrionic Disorder).

The Referral Sheet shows that amongst the 85 individuals, 65 % present at least one comorbid disorder. 17 % have Depressive Disorders including Dysthymia, 9 % have Anxiety Disorders, 8 % exhibit Obsessive-Compulsive Disorder (OCD), 12 % have Substance Use Disorder or Dual pathology (with Non Specific Anxiety Disorder), 5 % have Adaptive Disorder, 6 % have Psychotic or Schizoaffective Disorder, 4 % have Bipolar Disorder, 2 % have an Eating Disorder and 2 % have Attention Deficit Hyperactivity Disorder (ADHD).

Upon reviewing individual records from the year prior to accessing the program, it can be seen that 100 % of the participants were being treated at their Mental Health Unit. There were a total of 23 admissions in the Psychiatric Hospitalization Unit, and 99 visits to emergency service. In addition, 10 of the individuals had been in other outpatients or long-stay departments (Addictive Behaviour Unit, Eating Disorders Unit, Dr Esquerdo, amongst others).

If profile differences for BPD (n = 43) and PD-NOS/Mixed Personality (n = 40) are considered:

Traits profile in PID-5

70 patients completed the PID-5, although in order to obtain more representative results, two cases with different specific diagnoses, Obsessive-Compulsive Personality Disorder and Histrionic Disorder were excluded. Therefore, for this test, the results of N = 68 were analysed, 37 of which belong to the BPD subgroup, and 31 belong to the PD-NOS/MP subgroup.

The scores in the PID-5 are expressed in a Mean Score (M), for each domain and facet, ranging between 0 and 4. Higher mean scores (M = 4), indicate greater "dysfunction" in the facet or domain.

At the start, the group as a whole obtained the highest score in the Negative Affect domain (M = 1.95, SD = .57), followed by Disinhibition (M = 1.69, SD = .62), Detachment (M = 1.42, SD = .57) and Psychoticism (M = 1.19, SD = .65). The lowest score obtained is Antagonism (M = .64; SD = .51). It follows that the patients included in the program stand out for being high Negative Affect, which, as defined in the DSM-5, implies having "frequent and intense experiences of high levels of varied negative emotions and their behavioural and interpersonal manifestations, as well as irrational ideas and difficulty facing stressful situations".

With regard to Facets, the means and standard deviations can be seen in the following Table 1.

Table 1. Scores in the Facets of PID-5 for the total group. Means and Standard Deviations. 

As can be seen in this table, at the beginning the following Facets stand out: Lability (M = 2.20, SD = .71), Anxiousness (M = 2.11, SD = .59), Depression (M = 2.04, SD = .72), Anhedonia (M = 1.90, SD = .61), and Impulsivity (M = 1.91; SD = .88). They have means that are higher than or close to 2.

It is clear in the PID-5 PRE that patients included in the program need treatment focused on Negative Affect (lability, depression, anxiousness and anhedonia) and Impulsivity. It suggests there is a need to intervene both in emotional and behavioral regulation, in irrational beliefs, as well as in coping with stress.

Differences between PD-NOS/MP and BPD in the PID-5

If the group that returned the questionnaire in the PID-5 is divided into subgroups according to the diagnosis at admission, it can be seen that from “N = 68”, 37 are diagnosed with BPD, and 31 are diagnosed with PD-NOS/MP. There are noticeable differences between these two groups in the PID-5.

When comparing in the PID-5, BPD subgroup to the PD-NOS/MP subgroup, noticeably higher scores can be observed in all domains (greater dysfunction), except in antagonism. The Negative Affect Domain stands out (M = 2.09, SD = .50), followed by Disinhibition (M = 1.85, SD = .56). Regarding Facets, Lability (M = 2.40, SD = .55), Depression (M = 2.27, ST = .50), Anxiousness (M = 2.21, ST = .61), Impulsivity (M = 2.13, SD =. 79), Anhedonia (M = 2.03, SD = .49), and Distractability (M = 1.96, SD = .60), all stand out.

It is worth highlighting that according to the DSM-5, Section III, in order to diagnose BPD, four or more of seven pathological features should be present. At least one must be (5), (6) or (7). These include: 1. Emotional lability 2. Anxiousness, 3. Separation Insecurity, 4. Depression, 5. Impulsivity, 6. Risk-taking 7. Hostility. As can be seen, these conditions are met in this group, meaning, therefore, that they are correctly classified in this diagnostic category.

On the other hand two prominent facets can be found, Anhedonia and Distractability. Distractability is part of the disinhibition domain, and is defined as "difficulty concentrating and paying attention to tasks; easily distracted by external stimuli; problems with fulfilling goal-oriented behaviours, such as planning and performing tasks". These two facets can have an effect on difficulty with following a therapeutic program, so they should be taken into account when considering the type and mode of intervention.

The PD-NOS/MP subgroup, obtains lower scores in all the domains compared to the BPD subgroup, except in antagonism, although there aren't any scores close to 2 in any domain. However, when observing the facets, scores in lability (M = 1.97, SD = .81), and anxiousness (M = 1.99, SD = .55), stand out. Up to this point, when using the PID-5, it can be deduced that the patients referred, who are diagnosed with PD-NOS/MP essentially need an approach that tackles lability and anxiousness.

Having seen the profiles, it is necessary to determine whether the differences between the BPD group and the PD-NOS/MP group, in mean scores of the Domains and the Facets of the PID-5 are statistically significant. For this, the Student's t-test for independent samples was used. This data is collected in Table 2.

Table 2. Scores in PID-5. Differences between BPD and PD-NOS/MP, Student´s t-test comparison of means for independent groups. 

Legend: 1: Perceptual Dysregulation, 2: Restricted Affectivity, 3: Separation Insecurity, 4: Unusual Beliefs and Experiences; M: Means; S.D.: standard devia-tions; t: Student for independent groups; d.f.: degrees of freedom; p: probability obtained; C.I: confidence interval at 95 %; Inf .: lower limit; Sup .: upper limit; δ: size effect of the Cohen delta; ^* Significant difference p < .05; ^**Significant difference p < 0.1.

As Table 2 demonstrates, the Domains in which significant differences have been found (p < .05), with an Medium effect size for ^{Cohen's delta (Cohen, 1988}), are Negative Affect (t ₍₆₆₎ = 2.23, p < .05), and Disinhibition (t ₍₆₆₎ = 2.37, p < .05).

With regard to Facets, significant differences have been found between both groups. In Depression (t ₍₄₇₎ = 2,965, p = .005), the difference is significant of p < .01, and the effect size would be Large (^{Cohen, 1988}). Differences can be found at .05, with Medium effect sizes, in Lability (t _(51.6) = 2.476, p = .017), Impulsivity (t ₍₆₆₎ = 2.376, p = .020), Distractability (t ₍₆₆₎ = 2.026, p = .047), and Eccentricity (t ₍₆₆₎ = 2.006, p = .049). As can be seen, the BPD not only obtains higher scores in these facets, but the differences between both groups are significant. It is noted that this group also scores high in Anhedonia, despite the difference not being significant.

These profiles point towards a global need to address Negative affect, especially Lability and Anxiousness. For the specific case of BPD, it is also necessary to address depression, and there is a particular need to address the disinhibition domain (impulsivity and distractibility facets), and eccentricity.

Profile and differences in BEST

As previously mentioned in the methodology, BEST includes 15 items that are grouped into three subscales. Subscale A, Thoughts and Emotions (items 1 to 8), Subscale B, Negative Behaviours (items 9 to 12), and Subscale C that evaluates Positive Behaviours (items 13 to 15).

With a view to making the scores from different sections comparable, the direct scores from each subscale A, B and C, were converted into mean scores, oscillating between a minimum score of 1 and a maximum score of 5. The first month's mean score was used as the baseline for each subscale and item, that is, the mean obtained in the first 4 sessions of treatment.

The main interest is not only in knowing the profile of the global group, but also in finding out if there are significant differences between the two BPD and PD-NOS/MP groups analysed in subscales and items.

Given that not all the patients admitted to the DH stay in the program for the entirety of the first month, a n = 65 was found, in which 34 patients belong to the BPD subgroup and 31 belong to the PD-NOS/MP subgroup. For subscales A-B-C and items, the means obtained in the first month by the group are shown in Table 3 below.

Table 3. Global group scores in BEST. Means and Standard Deviations in the subscales and items. 

^*1 The contents of the items correspond to the evaluated concept, not the original text.

It can be observed that the items with the highest score in the first month are mood instability (M = 3.539, SD = 1.18), and chronic feelings of emptiness (M = 3.43, SD = 1.263), followed by changes in how patients see themselves (M = 3.159; SD = 1.180). In a positive sense, the item that by far scores the highest in the questionnaires is following therapeutic instructions (M = 4.044, SD = .969).

If the differences between the BPD and PD-NOS/MP group are analysed, with Student's t-test for independent samples, significant differences in all subscales with a Large effect size (p < .01; δ > .80), in subscales A (t ₍₆₃₎ = 3.221, p = .002, 95 % IC = (.254, 1.088), δ Cohen = .79), and B (t ₍₆₃₎ = 3.855, p <.000, 95 % CI = (.390, 1.232), δ Cohen = .95). And a Medium effect size in subscale C (t ₍₆₃₎ = -2.414, p = .019, 95 % IC = (-.889, - .083), δ Cohen = .59).

Therefore, the BPD group scores higher in subscales A and B, whilst the PD-NOS/MP group does so in subscale C. This would indicate that not only does the BPD group have significantly more intense symptomatology (cognitive, emotional and behavioural), but also in comparison to the PD-NOS/MP group, they use coping behaviour less frequently for their symptoms.

When the BEST items are compared in the first month, the BPD group score higher on all items except those related to coping behaviours (conversely for PD-NOS). It is necessary to point out that although the groups coincide in the items that receive the highest scores (chronic feelings of emptiness and extreme changes in how they see themselves); there are significant differences between them. In fact, the difference in means in the items between the BPD group (n = 34), and the PD-NOS/MP group (n = 31), is significant in most items, as can be seen in Table 4.

Table 4. BEST ratings. Differences between BPD and PD-NOS/MP, Student´s t-test comparison of means for independent groups. 

Legend: ^* Significant difference p < .05; ^** Significant difference p < .01; d.f.: Degrees of Freedom. C.I: confidence interval. δ: Cohen´s delta.

Thus a significant difference of p < .01 and a Large effect size (Cohen's δ > .80), for items: 3-changes in identity, 4-changes in mood, 8-ideas of suicide, and 11-impulsive behaviours are found. There is a significant difference of p < .05, and a Medium effect size (Cohen's δ > .50), for the items: 2- opinion of others, 7-feelings of emptiness, 10-getting hurt or attempting to commit suicide, and 12- fits of anger or problems controlling it.

When the differences in items are observed, the BPD group has significantly more intense symptomatology (cognitive, emotional and behavioural), when compared to the PD-NOS/MP group. Ideas and suicidal behaviours (items 8 and 10), as well as the impulsive behaviours (item 11), are of particular importance, which imply that this group needs to be monitored more closely during the therapeutic process.

Profile and differences in COPE-28

In relation to stress coping strategies, 61 subjects completed the questionnaire COPE-28. The scores in each coping strategy range from 0 to 6, depending on the frequency of use of the strategy.

The most frequently used strategies (M > 3), by this group are: Use of instrumental/social support, Active coping, Use of emotional support, Self-distraction, Venting/venting of emotions and Self-blame. Those that are used with less frequency (M < 3), are: Planning, Positive reframing, resorting to Religion, Substance use, Behavioural disengagement, Denial of having problems and Humour.

When PD-NOS/MP and BPD groups are compared, with Student's t -test for independent samples, they score similarly in all strategies, with the exception of BPD scoring significantly higher in Self-blame (t ₍₅₉₎ = 2.40, p = .019, 95 % CI = (.161, 1.763), Cohen δ = .86), and Substance use (t ₍₅₉₎ = 2.062, p = .044, 95 % IC = (.030, 2.011), Cohen δ = .687).

These results point to the need to address strategies such as positive re-framing, self-blame and venting/venting of emotions (appropriate expression of emotions). When referring to BPD, it is also necessary to develop alternative strategies to drug use to cope with stress.

Profile and differences in the IPDE

Prior to conducting a structured IPDE interview, a questionnaire is administered that allows for the selection of the most relevant items to be asked in the interview.

It is necessary to point out that this is not a diagnostic instrument; however, the study's intention was to consider to what extent it could be used to help provide a diagnosis and establish whether the results conform to the initial diagnostic. On the other hand, to compare any possible differences in the IPDE between the two groups (BPD and PD-NOS/MP).

The analysis of the two cases with diagnoses other than BPD or PD-NOS/MP is excluded from the analysis. In the total group (N = 68), 76 % of cases (n = 52), meet the necessary number of criteria to consider BPD as a probable diagnosis. 53 % of the subjects (n = 36), meet the necessary number of criteria to consider Avoidant Personality Disorder as the diagnosis.

However, of the 68 cases, only 20 % meet the necessary number of criteria to consider a single diagnosis, compared to 67 % that meets the necessary number of criteria for two or more disorders. There is even 13 % that do not meet the necessary number of criteria to be diagnosed with any type of disorder. These data in the IPDE disagree with the initial diagnoses by which they have been derived to the resource, contributing confusion in the diagnosis.

If we observe by diagnosis, in the PD-NOS/Personality Mixed subgroup, in analysing the frequencies, out of the 31 cases, the 55 % cases meet the criteria for at least two disorders, and 5 cases (17 %), do not meet criteria for any specific disorder. Nevertheless, 8 cases (28 %), can be seen to meet the necessary number of criteria for a single diagnosis, out of these, 5 cases (17 %), exclusively meet criteria for BPD. Therefore, with the IPDE data, at least 28 % would be incorrectly classified in the PD-NOS/MP category.

The subgroup diagnosed with BPD (n = 37), taking into consideration the means and in the frequency of cases for each one, scores more features in all of the subscales of the IPDE. Borderline and evasive subscales stand out. Precisely in these scales they have a significant difference between both groups (BPD vs PD-NOS/MP), with the Student´s t-test for independent samples, in the Limit scale (t _{(63 )}= -2.99, p = .004, 95 % IC = (-2.858, -.561), δ de Cohen = .78), and the Avoidant scale (t ₍₆₃₎ = -.546, p = .014, 95 % IC = (-2.144, -.25685), δ de Cohen = .658).

In analysing the frequencies, the subgroup BPD, it can be observed that of the 37 subjects, only 5 meet the criteria for a single disorder, and only 2 of these subjects (6 %), meet the criteria for BPD. The other 32 (94 %), meet the criteria for the diagnosis of 2 or more disorders. Therefore, for the BPD group, the IPDE suggests that 94 % of the cases would be considered as having Mixed Personality Disorder, so they would be incorrectly classified as BPD. This result contrasts with the results seen in the PID-5, where this group obtained a profile of traits that complied with the BPD criteria according to the dimensional-hybrid model.