The validity and reliability of the “Factors Related to Theft by Patients with Eating Disorders Questionnaire”

Abstract

Purpose

Although kleptomania is frequently observed in patients with eating disorders, research on its causes is scarce, and no scale exists to assess it. This pilot study was done to develop such a scale and to evaluate its validity and reliability.

Methods

Forty-six female patients with anorexia nervosa incarcerated at the Kitakyushu Medical Prison completed our newly developed questionnaire that is based on the DSM-5 diagnostic criteria for kleptomania. All participants had developed an eating disorder prior to their initial incarceration in a correctional facility, had no history of stimulant abuse, and were incarcerated for theft.

Results

Factor analysis (principal axis method, Promax rotation) identified three significant factors: “Elation/Pleasure”, “Material/Monetary Desire”, and “Impulsivity.” The test–retest correlation coefficients for each item were generally high, confirming test–retest reliability. While the Cronbach’s alpha coefficients for items with high factor loadings on each factor were moderate, the correlations between factors were low, suggesting that each factor represents an independent construct. The “Elation/Pleasure” factor aligned with DSM diagnostic criteria B and C, the “Impulsivity” factor aligned with Criterion A, and the “Material/Monetary Desire” factor contradicted Criterion A.

Conclusion

These preliminary findings suggest the validity and reliability of this questionnaire, which indicates that it may be useful as a foundational resource and contribute to the advancement of this field.

Level of evidence III: Retrospective observational study.

Background

Eating disorders (ED) [1] mainly affect women and typically begin during adolescence. They are generally difficult to treat, with the disease course often becoming chronic and severe without recovery [2, 3]. It is not uncommon for patients to repeatedly engage in theft. According to previous reports: (i) persons with ED engage in theft more frequently than the general population [4, 5]; (ii) there is a link between theft and binge eating [4, 6, 7]; (iii) theft can serve as an indicator of ED severity [6, 8]; and (iv) persons with ED face a higher risk than others of criminal conviction [4].

Although a relation between ED and theft is frequently noted, detailed reports are relatively scarce, and the pathophysiology remains largely unexplored [9, 10]. Theft by ED patients is often lumped into the broad category of kleptomania [1], as distinguished from ordinary theft. However, the strictness of the application of the diagnostic criteria for kleptomania (Table 1) appears to vary considerably between studies.

Table 1.

DSM-5 diagnostic criteria for kleptomania

A	Recurrent failure to resist impulses to steal objects that are not needed for personal use or for their monetary value
B	Increasing sense of tension immediately before committing the theft
C	Pleasure, gratification, or relief at the time of committing the theft
D	The stealing is not committed to express anger or vengeance and is not in response to a delusion or a hallucination
E	The stealing is not better explained by conduct disorder, a maniac episode, or antisocial personality disorder

In a review of the relation between theft and ED, Baum et al. [6] noted that kleptomania accounts for only a very small proportion of shoplifting. They stated, “Theft in various individuals with ED likely involves different combinations of factors. A framework of potential factors linked to and individualized for a specific person’s circumstances and characteristics is likely more accurate.” Furthermore, they proposed revising the diagnostic criteria for kleptomania, citing difficulties in determining whether stolen items are intended for personal use or monetary gain under the current definition.

In a paper that discusses challenges in the diagnosis of kleptomania, Sipowicz et al. [11] noted that while few patients meet all diagnostic criteria, a problematic subgroup exists. This group experiences compulsive stealing urges, pre-act tension, and post-act relief, yet does not meet the criterion of “stealing unnecessary items”. To define this, they proposed the concept of a “kleptomania spectrum”.

The number of female inmates with ED in Japanese prisons has increased in recent years [12–14]. The most common charge is theft, followed by violations of the Stimulant Control Act [15–18]. Even after incarceration, many continue with the behaviors associated with their ED: impaired eating habits and/or intentional vomiting. Significant weight loss often places them at risk of life-threatening conditions. Patients deemed difficult to manage in general prisons are transferred to medical prisons for specialized treatment [13–16]. The first author has treated numerous female patients with anorexia nervosa (AN) at the Kitakyushu Medical Prison. The clinical presentations of these patients show marked differences depending on their criminal charges and history of ED.

In our previous report [17], we classified 100 female AN patients admitted for treatment at Kitakyushu Medical Prison into four groups based on their history of ED and stimulant abuse prior to their first admission to a correctional facility. Table 2 shows profiles of the four groups and their comparisons. The largest group, comprising approximately three-quarters of the total, consisted of patients with a history of ED but no history of stimulant abuse prior to their first admission to a correctional facility [“ED ( +) Stimulant Abuse (−)” group]. This group exhibited the most typical and severe ED-like clinical presentation. They generally developed their ED at a young age (21.0 ± 8.7 years), had an extremely long duration of illness (20.9 ± 8.6 years), and were middle-aged or older at the time of admission to our facility (41.9 ± 9.2 years). They were severely underweight, with an average BMI at admission of 13.3 ± 1.6 kg/m². The binge-eating/purging type of AN was predominant at 79.7%. The criminal charge was theft for 97.3% of the patients. Psychological and behavioral deterioration occurred during the long illness duration, with theft often emerging after the onset of binge eating and purging.

Table 2.

Clinical characteristics of 100 female anorexia nervosa patients admitted for treatment at Kitakyusyu Medical Prison categorized into four groups according to their eating disorder and stimulant abuse history prior to their first admission to a correctional facility

	“ED ( +) stimulant abuse (−)” groupa n = 74 (74.0%)	“ED (−) stimulant abuse (−)” groupb n = 13 (13.0%)	“Stimulant abuse ( +) ED ( +)” groupc n = 4 (4.0%)	“Stimulant abuse ( +) ED (−)” groupd n = 9 (9.0%)	p	Multiple comparison (Tukey’s test)
Age at admission (years)	41.9 ± 9.2 (22–62)	52.0 ± 10.6 (39–73)	43.0 ± 9.4 (33–52)	37.8 ± 11.6 (24–54)	0.0029	a,d < b
Onset of eating disorder (years)	21.0 ± 8.7 (12–51)	48.3 ± 9.5 (38–67)	20.5 ± 4.4 (16–26)	32.8 ± 10.9 (19–48)	< 0.0001	a, c, d < b, a < d
Duration of eating disorder (years)	20.9 ± 8.6 (4–41)	2.0 ± 1.5 (0–5)	22.5 ± 10.4 (11–34)	5.0 ± 8.2 (0–24)	< 0.0001	a, c > b, d
BMI at admission (kg/m2)	13.3 ± 1.6 (10.0–17.4)	14.0 ± 1.4 (11.7–16.5)	13.4 ± 1.4 (12.2–15.4)	14.9 ± 2.0 (10.3–16.9)	0.0204	a < d
Body weight at admission (kg)	32.6 ± 3.9 (24.4–40.4)	35.1 ± 4.0(27.8–40.2)	31.6 ± 6.0 (23.6–38.0)	36.5 ± 4.4 (26.0–41.0)	0.0121	a < d
Type of anorexia nervosa					0.0011
Restricting type	13 (17.6%)	8 (61.5%)	0 (0.0%)	5 (55.6%)
Binge-eating/purging type	59 (79.7%)	5 (38.4%)	4 (100.0%)	4 (44.4%)
Unknown	2 (2.7%)	0 (0.0%)	0 (0.0%)	0 (0.0%)
Number of admissions to a correctional facility	1.8 ± 1.2 (1–7)	2.1 ± 1.5 (1–5)	3.3 ± 2.2 (1–6)	3.9 ± 3.0(1–9)	0.0009	d > a, b
Criminal charge of this time (multiple charges possible)
Theft	72 (97.3%)	11 (84.6%)	2 (50.0%)	2 (22.2%)
Violation of the stimulant control act	0	0	2 (50.0%)	8 (88.9%)
Fraud	0	3 (23.1%)	1 (25.0%)	0
Violation of the anti-prostitution act	1 (1.3%)	0	0	0
Murder crime	1 (1.3%)	0	0	0
Crime of injury	0	1 (7.7%)	0	0

^aGroup of patients with a history of eating disorder but no history of stimulant abuse prior to their first admission to a correctional facility

^bGroup of patients with no history of eating disorder and no history of stimulant abuse prior to their first admission to a correctional facility

^cGroup of patients with history of stimulant abuse and history of eating disorder prior to their first admission to a correctional facility

^dGroup of patients with history of stimulant abuse but no history of eating disorder prior to their first admission to a correctional facility

To deepen our understanding of the motives for theft by ED patients, we developed the “Factors Related to Theft by Patients with Eating Disorders Questionnaire”, which is based on the diagnostic criteria for kleptomania [1]. This study was done to examine the reliability and validity of the questionnaire by administering it to female AN patients treated at our facility who belonged to the aforementioned “ED ( +) Stimulant Abuse (−)” group and whose criminal charge was theft.

The hypotheses are as follows:

• (i)The questionnaire will demonstrate construct validity through factor analysis.
• (ii)The reliability of the questionnaire will be demonstrated through test–retest and internal consistency analysis.

Methods

Subjects

Of the 68 female patients with anorexia nervosa (AN) who received treatment at our medical facility between September 2018 and December 2025, 51 were classified into the “ED ( +) Stimulant Abuse (−)” group. Of the 50 in this group who were incarcerated for theft, the 46 who consented to participate were included in the study. All were Japanese.

Questionnaire development and administration

The “Factors Related to Theft by Patients with Eating Disorders Questionnaire” (Table 3) consists of eight items rated on a 5-point scale. It was designed to investigate factors that contribute to theft by ED patients. Most items are related to DSM-5 diagnostic criteria for kleptomania A–C (Table 1) (some align with them, while others contrast). This questionnaire was distributed to participants 1–2 months after admission to our facility, when they had stabilized physically and psychologically, and was collected 2–3 days later.

Table 3.

“Factors Related to Theft by Patients with Eating disorders questionnaire”

Factor analysis

Factor analysis (principal axis method with Promax rotation) was performed on the questionnaire responses.

Evaluation items

Age at admission, age at onset, duration of the eating disorder, BMI at admission, type of anorexia (restrictive or binge–purge), and number of admissions to correctional facilities were collected from the patients’ medical records and personal history records (forms created by correctional facilities containing chronological data on each inmate’s background, criminal history, and medical history). The frequency of theft, and theft history prior to the onset of the ED were investigated by self-reported questionnaire.

Verification of test–retest reliability

To assess test–retest reliability, the questionnaire was administered twice to 16 patients, at an interval of 2–3 weeks between December 2025 and March 2026.

Ethical considerations

After explaining the study conditions both verbally and in writing, we obtained consent to participate and to the publication of the results. This study was conducted with due consideration for the protection of personal information. Prior to the start of the study, approval was obtained from the Ethics Review Committees of the Kitakyushu Medical Prison and Okada Internal Medicine Clinic.

Results

Characteristics of the eating disorders and theft of all subjects

Table 4 shows the characteristics of the EDs and theft of all patients. The mean age at onset of the ED was 22.0 ± 9.8 (mean ± standard deviation) years, duration of ED was 23.6 ± 8.4 years, age at admission to our facility was 45.6 ± 8.3 years, BMI was 13.8 ± 1.3 kg/m², and 84.8% had the binge-eating/purging type of AN. The average number of incarcerations in correctional facilities was 2.6 ± 1.6, and the frequency of theft was 3.3 ± 2.4 days per week. Twenty-eight participants (60.9%) primarily stole food items, 13 (28.3%) stole both food and non-food items, 3 (6.5%) stole non-food items, and 2 (4.3%) were unknown. Eleven participants (23.9%) had begun habitual theft before developing an ED.

Table 4.

Profiles of the eating disorders and theft of all eligible patients

	n = 46
Age at admission (years)	45.6 ± 8.3
Age at onset of eating disorder (years)	22.0 ± 9.8
Duration of eating disorder (years)	23.6 ± 8.4
BMI (kg/m2)	13.8 ± 1.3
Type of anorexia nervosa
Binge-eating/purging type	39 (84.8%)
Restricting type	7 (15.2%)
Number of times admitted to a correctional facility	2.6 ± 1.6
Frequency of theft (days/week)	3.3 ± 2.4
Max frequency of theft in life time (days/week)	4.1 ± 2.5
Target of theft
Food-centered	28 (60.9%)
Food and Non-food	13(28.3%)
Non-food-centered	3 (6.5%)
Unknown	2 (4.3%)
Theft habit prior to onset of the eating disorder (once a month or more)
Yes	11 (23.9%)
No	34 (73.9%)
Unknown	1 (2.2%)

Item-level descriptive statistics

The mean scores and standard deviations for each item were as follows: Item 1: 2.89 ± 1.54; Item 2: 2.54 ± 1.41; Item 3: 3.35 ± 1.32; Item 4: 3.57 ± 1.26; Item 5: 3.02 ± 1.61; Item 6: 3.96 ± 1.37; Item 7: 2.52 ± 1.50; Item 8: 2.93 ± 1.20.

Results of factor analysis

The eigenvalues were 2.73 for Factor 1, 2.22 for Factor 2, and 1.03 for Factor 3. Because only these three factors had eigenvalues greater than 1, a three-factor analysis was conducted in accordance with the Kaiser criterion.

Table 5 shows the results of the factor analysis for each item in the questionnaire. For Factor 1, items 1, 2, and 3 showed high factor loadings. These items indicate a tendency to feel a sense of elation or pleasure immediately before and during the act, as well as heightened impulses to steal. We named this the “Elation/Pleasure” factor. For Factor 2, items 4, 5, and 6 showed high factor loadings. These items indicate theft motivated by a desire for possessions or money. We named this the “Material/ Monetary Desire” factor. For Factor 3, items 7 and 8 showed high factor loadings. These items indicate that high levels of impulsivity are involved in theft. We named this the “Impulsivity” factor.

Table 5.

Factor loading for each item of the “Factors Related to Theft by Patients with Eating Disorders Questionnaire”

Item	Factor 1 “Elation/Pleasure”	Factor 2 “Material/Monetary desire “	Factor 3 “Impulsivity”	Communalities
No1	0.89	− 0.03	0.01	0.79
No2	0.81	− 0.12	0.04	0.65
No3	0.71	0.14	− 0.07	0.54
No5	− 0.04	0.72	0.20	0.44
No4	0.01	0.65	− 0.37	0.73
No6	0.02	0.61	− 0.03	0.39
No7	0.24	0.15	0.65	0.55
No8	− 0.09	− 0.05	0.60	0.35
Variance	2.24	1.62	1.46
Contribution rate	28.01	20.23	18.29

Correlation between factors
	Factor 1	Factor 2
Factor 2	0.21
Factor 3	0.35	− 0.35

Factor loadings of 0.6 or higher are shown in bold

The total variance explained by this factor structure was 66.5%. Bartlett’s sphericity test yielded a p-value of < 0.0001, and the KMO index was 0.643. The correlation coefficients between factors were 0.21 between Factor 1 and Factor 2, 0.35 between Factor 1 and Factor 3, and −0.35 between Factor 2 and Factor 3.

Reliability

1. Internal consistency

Cronbach’s alpha, calculated for the groups of items with high factor loadings, was 0.85 for Factor 1, 0.72 for Factor 2, and 0.56 for Factor 3.

2. Test–retest reliability

The test–retest correlation coefficients were No. 1: 0.76, No. 2: 0.71, No. 3: 0.89, No. 4: 0.75, No. 5: 0.89, No. 6: 0.83, No. 7: 0.66, and No. 8: 0.54.

Discussion

Factor 1 (“Elation/Pleasure”)

Factor 1 corresponds to diagnostic criteria B and C for kleptomania. Items 1 and 2, which have high factor loadings, align with diagnostic criterion B (“increasing sense of tension immediately before committing the theft”) and criterion C (“pleasure, gratification, or relief at the time of committing the theft”). For example, the statements in Item 1, “Immediately before committing theft, I experience feeling an increasing sense of tension”, and Item 2, “During the act of theft, I experienced pleasure, gratification, or a sense of relief.” Item 3, which also has a high factor loading, expresses a “rising urge to steal”, it appears to share commonalities with heightened elation or pleasure. The factor scores for Factor 1 range widely from −1.30 to 1.90, suggesting considerable variation in the extent to which individual patients met diagnostic criteria B and C.

Factor 2 (“Material/Monetary Desire”)

Factor 2 is associated with the most critical aspect of Diagnostic Criterion A for kleptomania: “the impulse to steal objects that are not needed for personal use or for their monetary value.” However among the items with high factor loadings, Item 5 indicates a strong desire for money, as evidenced by the statement “I had money, but didn’t want to spend it.” Furthermore, Items 4 and 6 indicate a strong desire for materials, as evidenced by the statements “I stole to get what I wanted” and “I couldn’t feel safe unless I had a lot of things around me.” Therefore, Factor 2 indicates that theft is done for personal use or monetary value, in contradiction with Diagnostic Criterion A. While seeking the value of objects or money is typically the most common motive for theft, it is often argued that theft by persons with an ED is frequently driven by other factors. However, in actual clinical practice, it is often observed that the value of objects or money plays a significant role as a motivation for theft even by patients with an ED. Binge eating and purging, as well as hoarding food, require extremely large quantities of food. Because it is difficult for them to cover these costs on their own, they often resort to theft as a means to obtain the necessary funds. Under such circumstances, it is likely that their “Material/Monetary Desire” will intensify. In questionnaires regarding the motives for theft by patients with an ED, identifying factors related to “Material/Monetary Desire” is crucial for understanding the actual impact such desire has on theft. Factor 2 also exhibits a wide distribution of factor scores, ranging from − 2.14 to 1.04, suggesting that the extent to which theft is committed for the sake of material or monetary value varies among patients.

Factor 3 (“Impulsivity”)

Factor 3 corresponds to Diagnostic Criterion A for kleptomania, which describes the specific reasons and motives for theft in kleptomania. Whether or not a person meets this criterion can be considered a primary determinant of if they have kleptomania. Items 7 and 8, which have high factor loadings, align with the content of Diagnostic Criterion A: “recurrent failure to resist impulses to steal.” For example, the statement in Item 8, “I felt an urge to steal and ended up doing so”, corresponds to this. Furthermore, the statement in Item 7, “I didn’t really want the things I stole, so I often gave them away for free or threw them away”, suggests that there was an irresistible appeal in the act of stealing itself rather than in the value of the items, and that their theft was impulsive in nature. These items largely align with Diagnostic Criterion A, and persons with higher scores on this factor are thought to exhibit stronger kleptomaniac tendencies. Factor 3 scores also range widely from −1.63 to 1.66, indicating significant variation in the extent to which participants in this study meet Diagnostic Criterion A.

Reliability

The questionnaire demonstrated acceptable internal consistency and test–retest reliability. Although the Cronbach’s alpha for Factor 3 was only 0.56, this is likely due to the small number of items (only 2).

Validity

There are no existing questionnaires that we can use for comparison. Furthermore, because this is not a clinically established field, it is extremely difficult to obtain a sample of patients with clearly diagnosed kleptomania. Due to the absence of such external reference indicators, it was difficult to demonstrate criterion-related validity.

However, for each of the three factors, correlations were generally high among items with high factor loadings, while correlations between factors were generally low. These results suggest that each factor represents a distinct construct.

Furthermore, in terms of content, this questionnaire appears to comprehensively reflect the major items of the DSM-5 diagnostic criteria for kleptomania. Factor 1 aligns with diagnostic criteria B and C for kleptomania, Factor 3 aligns with Diagnostic Criterion A, while Factor 2 contradicts Diagnostic Criterion A. This questionnaire, which includes factors (Factors 1, 2, and 3) related to all the main diagnostic criteria (A, B, and C), also features a factor (Factor 2) that focuses on the most critical point in the diagnosis of kleptomania, “neither for personal use nor for monetary gain” and thus can be considered valid from a clinical perspective.

Strengths and limitations

Although habitual theft is frequently observed among patients with EDs, no objective assessment scale had been developed to identify its contributing factors. Therefore, we developed this questionnaire to identify factors that contribute to theft and administered it to female patients with AN. This study showed that the questionnaire has acceptable reliability and validity, which suggests that it could serve as a foundational tool for research and contribute to the advancement of this field.

This scale was developed based on the DSM-5 diagnostic criteria for kleptomania. If future research confirms these results, it would be useful as a tool for assessing the extent to which individual patients exhibit or deviate from the characteristics of kleptomania described in the diagnostic criteria, and for determining whether or not they meet the diagnostic criteria.

Recruiting patients for such studies is particularly difficult because they rarely visit medical institutions on a regular basis. A strength of this study lies in the fact that we were able to collect a relatively large amount of patient data from this particularly hard-to-reach population. This was made possible by the circumstances of Japan’s correctional facilities, where many patients with EDs are transferred from general prisons to medical prisons for treatment. Furthermore, the selection of a relatively homogeneous patient group exhibiting typical ED characteristics is believed to have contributed to the attainment of clear results.

A limitation of this study is that, due to the nature of the subject population and the insufficient number of cases, the statistical reliability and validity could not be fully demonstrated. This pilot study is exploratory in nature, and further research will be necessary, including analyses with a larger sample size. Furthermore, the subject population was limited to patients with AN incarcerated in medical prisons, meaning that all participants had the most severe form of ED. To obtain more general and comprehensive findings, future research must expand the scope to include a broader range of ED patients, such as those in general prisons, those in general medical institutions, patients with EDs other than AN, male patients, and patients of other nationalities.

Conclusion

We administered the newly developed “Factors Related to Theft by Patients with Eating Disorders Questionnaire” to female AN patients at the Kitakyushu Medical Prison. Factor analysis of the responses identified “Elation/Pleasure”, “Material/Monetary Desire”, and “impulsivity” as significant factors. Adequate reliability was established through acceptable internal consistency and test–retest reliability. While correlations were generally high among items with high factor loadings for each factor, correlations between factors were low, suggesting that each factor represents an independent construct. The questionnaire was developed based on the DSM-5 diagnostic criteria for kleptomania. Because each factor corresponds to a specific criterion, it can be expected to serve as a useful source of information for determining whether or not a patient meets the diagnostic criteria. The factor scores for each factor varied widely, indicating the extent to which individual patients meet or deviate from each aspect of the diagnostic criteria for kleptomania.

These preliminary results suggest that this questionnaire can serve as a foundational resource for future research and contribute to the advancement of this field.

What is already known on this subject?

Although theft among patients with EDs is often explained in terms of kleptomania, some reports indicate that many patients do not meet the diagnostic criteria for this disorder. To clarify this issue, it is necessary to develop an objective measure to assess factors related to theft among patients with eating disorders. To this end, we developed such a questionnaire and verified its validity and reliability.

What this study adds?

Factor analysis of responses to the questionnaire on theft by patients with EDs identified three distinct factors: “Elation/Pleasure”, “Material/Monetary Desire”, and “Impulsivity”. Furthermore, the questionnaire was found to have adequate validity and reliability.

Author contributions

MT designed, collected and analyzed the data. JK instructed statistical analysis. MK and YS conducted psychological tests. ST, NS, and TS provided advice on the results and the drafting of the manuscript.

Funding

Not applicable.

Data availability

The datasets analyzed in the current study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval consent to participate

This study was approved by the Ethics Review Committees of Kitakyushu Medical Prison and the Okada Internal Medicine Clinic. Informed consent was obtained from all individual participants included in the study.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.