Self-Harm Inventory

Participants involuntarily hospitalized in a psychiatric facility: (Sansone et. al, 1998; N = 32; 50% female; mean (SD) age = 36.2 (13.33) years)

•  
  • Scores greater than or equal to 5 on the SHI are indicative of Borderline Personality Disorder. The SHI was 78.1% correct in classifying all research participants (Kappa = .33).
    • The SHI was 60% correct in classifying research participants without Borderline Personality Disorder.
    • The SHI was 81.5% correct in classifying research participants with Borderline Personality Disorder according to the DSM-IV criteria.

Participants with Borderline Personality Disorder: (Sansone et. al, 2008; N = 120; 61% female; mean age (SD) = 38.69 (11.74) years; 81.5% White, 15.1% Black, 1.7% Native American, 0.8% Asian, 0.8% race not otherwise specified)

• Mean (SD) score: 7.41 (4.82)

Participants involuntarily hospitalized in a psychiatric facility: (Sansone et. al, 1998; N = 32; 50% female; mean (SD) age = 36.2 (13.33) years)

• Mean (SD) scores of participants with Borderline Personality Disorder: 8.48 (4.11)
• Mean (SD) scores of participants without Borderline Personality Disorder: 3.80 (2.17)

Inpatient Psychiatric Sample: (Selbom et al. 2017, N = 145; 61% female; mean (SD) age = 38.06 (13.06) years; 73.6% White participants, 16% Black, 10.1% race not otherwise specified; 64.3% of participants reported at least one previous psychiatric hospitalization)

• Mean (SD) score: 7.31 (4.80)

Inpatient Psychiatric Sample: (Selbom et al. 2017, N = 145; 61% female; mean (SD) age = 38.06 (13.06) years; 73.6% White participants, 16% Black, 10.1% race not otherwise specified; 64.3% of participants reported at least one previous psychiatric hospitalization)

Excellent internal consistency (Cronbach’s alpha = 0.84)

Concurrent validity:

Participants with Borderline Personality Disorder: (Sansone et. al, 2008; N = 120; 61% female; mean age (SD) = 38.69 (11.74) years; 81.5% White, 15.1% Black, 1.7% Native American, 0.8% Asian, 0.8% race not otherwise specified)

• Adequate correlation between SHI scores and self-reported numbers of medically self-sabotaging behaviors (r = .55)

Participants involuntarily hospitalized in a psychiatric facility: (Sansone et. al, 1998; N = 32; 50% female; mean (SD) age = 36.2 (13.33) years)

Excellent correlation between SHI and PDQ-R scores (r = 0.71)

Convergent validity:

Inpatient Psychiatric Sample: (Selbom et al. 2017, N = 145; 61% female; mean (SD) age = 38.06 (13.06) years; 73.6% White participants, 16% Black, 10.1% race not otherwise specified; 64.3% of participants reported at least one previous psychiatric hospitalization)

• Excellent correlation between SHI and PDQ-4 scores (r = 0.62)
• Excellent correlation between SHI and SCID-2-PQ scores (r = 0.70)
• Excellent correlation between SHI and MSI-BPD scores (r = 0.68)

Development Study: (Sansone et. al, 1998; N = 221; primary care setting; 28% consulted for obesity, 51% for eating disorders or substance abuse, 19% for other non-emergency concerns)

•  
  • “Using a cutoff score of 5 on the SHI appears to best balance diagnostic sensitivity with false-identification of BPD.”
  • “Using a cutoff score of 5 on each measure resulted in… a Kappa of .61 for the SHI.”

Development Study: (Sansone et. al, 1998; N = 221; primary care setting; 28% consulted for obesity, 51% for eating disorders or substance abuse, 19% for other non-emergency concerns)

• Mean (SD) for participants with Borderline Personality Disorder: 10.15 (4.57)
• Mean (SD) for participants without Borderline Personality Disorder: 2.30 (2.51)

Development Study: (Sansone et. al, 1998; N = 221; primary care setting; 28% consulted for obesity, 51% for eating disorders or substance abuse, 19% for other non-emergency concerns)

• Excellent internal consistency (Cronbach’s alpha = 0.80)

Concurrent validity:

Development Study: (Sansone et. al, 1998; N = 221; primary care setting; 28% consulted for obesity, 51% for eating disorders or substance abuse, 19% for other non-emergency concerns)

• Excellent correlation between SHI and PDQ-R scores (r = .74)
• Excellent correlation between SHI and DIB scores (r = .76)
  • Distribution of SHI Scores and their relationship to a Borderline Personality Disorder diagnosis by the DIB:

SHI Score

BPD Diagnosis

n (%)

No BPD diagnosis

n (%)

0

0 (0)

62 (36.9)

1

1 (1.9)

17 (10.1)

2

1 (1.9)

27 (16.1)

3

2 (3.8)

12 (7.1)

4

2 (3.8)

20 (11.9)

5

5 (9.4)

7 (4.2)

6

1 (1.9)

9 (5.4)

7

4 (7.5)

6 (3.6)

8

1 (1.9)

4 (2.4)

9

7 (13.2)

3 (1.8)

10 or greater

29 (54.7)

1 (0.06)

Total N

53 (100)

168 (100)

Development Study: (Sansone et. al, 1998; N = 221; primary care setting; 28% consulted for obesity, 51% for eating disorders or substance abuse, 19% for other non-emergency concerns)

• “The proportion of respondents who endorsed each of the 22 remaining SHI items ranged from a low of 4.7% for item ‘burned yourself’, to a high of 37.6% for the item ‘abused alcohol’.”

Non-clinical Population: (Latimer et al. 2009; N = 423; 81% female; 53% between 17-19 years, 27% between 20-30 years; Australian undergraduate students)

• ≥ 5 is indicative of Borderline Personality Disorder (sensitivity 88.7%, specificity 82.1%)

Non-Clinical Population: (Latimer et al. 2009; N = 423; 81% female; 53% between 17-19 years, 27% between 20-30 years; Australian undergraduate students)

• Mean (SD) = 5.16 (3.6)
• Female participants endorsed more items than male participants.
  • Mean (SD) scores for female participants: 5.75 (3.75)
  • Mean (SD) scores for male participants: 4.31 (2.79)
• Older participants (20-30 years) endorsed more items than younger participants (17-19 years).
  • Mean (SD) scores for older participants: 6.08 (3.92)
  • Mean (SD) scores for younger participants: 4.39 (3.34)

General Population: (Müller et al., 2016; N = 2,507; 56% female; German participants)

• 48% of participants endorsed at least one SHI item.
  • Mean (SD) scores for men: 1.33 (2.03)
  • Mean (SD) scores for women: 1.20 (2.05)

General Population: (Müller et al., 2016; N = 2,507; 56% female; German participants)

• Adequate internal consistency (Cronbach’s alpha = 0.78)
• Adequate internal consistency among male participants (Cronbach’s alpha = 0.75)
• Excellent internal consistency among female participants (Cronbach’s alpha = 0.80)

Concurrent validity:

Non-Clinical Population: (Latimer et al. 2009; N = 423; 81% female; 53% between 17-19 years, 27% between 20-30 years; Australian undergraduate students)

• Excellent correlation with the DIB (r = 0.76)
• Excellent correlation with the PDQ-R (r = 0.73)

General Population: (Müller et al., 2016; N = 2,507; 56% female; German participants)

• Poor correlation between SHI and BIS-15 scores (Spearman correlation = 0.23)
  • Poor correlation for male participants (Spearman correlation = 0.26)
  • Poor correlation for female participants (Spearman correlation = 0.20)
• Adequate correlation between SHI and PHQ-4 scores (Spearman correlation = 0.37)
  • Adequate correlation for male participants (Spearman correlation = 0.36)
  • Adequate correlation for female participants (Spearman correlation = 0.40)

Convergent validity:

Non-Clinical Population: (Latimer et al. 2009; N = 423; 81% female; 53% between 17-19 years, 27% between 20-30 years; Australian undergraduate students)

• Relationship between SHI scores and depression, anxiety, and stress scores on the DASS-21 for participants who endorsed at least one item:

SHI scores

Mean (SD) depression scores

Mean (SD) anxiety scores

Mean (SD) stress scores

Low (1-4)

9.53 (8.15)

7.09 (6.53)

11.67 (6.40)

Medium (5-10)

11.58 (8.60)

11.41 (8.24)

16.52 (9.22)

High (11 or higher)

17.00 (10.51)

16.86 (11.39)

23.29 (9.53)

Meta-analysis: (Sansone et al., 2015)

• “the SHI demonstrated an accuracy in diagnosis of 84% at a cut-off score of 5, and 85% and 88% at cut-off scores of 6 and 7, respectively”

Meta-analysis: (Sansone et al., 2015)

• Adequate test-retest reliability of psychiatric inpatient samples (ICC = .61)
• Adequate test-retest reliability of internal medicine outpatient samples (ICC = .68)

Meta-analysis: (Sansone et al., 2015)

• Excellent internal consistency (Cronbach’s alpha = .85)